Ethernet at 10 Gbps

An anonymous reader writes "This article talks about 10 Gigabit Ethernet and asks, 'But just how much data can a person consume?' Currently at work, we're working on a major project to re-architect our core application platform so that the different systems can be de-coupled and hosted separately. The legacy design implicitly relies on systems being in the same LAN due to bandwidth-expensive operations (e.g., database replication). Having this much bandwidth would change the way we design. What would you do with this much bandwidth?"

Comments

What would I do with this much bandwidth?

[RLiegh]

1)Get more porn
2)Download linux dvds
3)FINALLY get the coveted First Post!
4)Download even more porn

Re:What would I do with this much bandwidth?

[leapis]

This much bandwidth isn't going to help you do any of these things. I upgraded my network to gigabit ethernet about a year ago (from 100 mbit), and much to my surprise, the speed increase was only about 3 times when copying files from one machine to another. I did a little math, and found the answer. Your average ATA hard drive, even at max bus speed, only delivers 0.8 Gbps. And in the real world, you are lucky to get half that from a single drive. In my own test transfers from RAID1 and RAID5 arrays, my transfer rates never once exceeded 0.70 Gbps. Until there is a fundamental increase in the amount of data you can get off a spinning disc, its not likely that a home user is going to saturate a 1 Gbps line, much less a 10 Gbps line.

Re:What would I do with this much bandwidth?

[soimless]

keep that up and you'll have all the porn on the internet in only 3 years!

Re:What would I do with this much bandwidth?

[Gabrill]

maybe so, but you don't work with all your data on the hard drive. Working with data on another computer can really speed up with faster ethernet, especially databases that stay partially in RAM.

Re:What would I do with this much bandwidth?

[Anonymous Coward]

I upgraded my network to gigabit ethernet about a year ago (from 100 mbit), and much to my surprise, the speed increase was only about 3 times when copying files from one machine to another.

Latency and the "jumbo packet" is the problem.

With cheap hardware, I can push around 16-20MB/s across the network between two machines. At least with a gigabit switch, that means I'm not going to saturate the network and cause problems for anyone else.

SCSI 320 is 320MB/s (or 3.2Gbps)... which is 1/3 of the 10Gbps network. SATA/300 is supposed to come out in a year or two and is competitive.

(There's something faster then SCSI's Ultra320... but I forget what it's called. I do know that SATA is planned to go to 600MBps at some point, which is 6.0Gbps.)

Re:What would I do with this much bandwidth?

[Anne Thwacks]

Install it at work, and take the 1G Ethernet kit home to upgrade my 100M network.

For those of you who work for Xerosx, you could upgrade to 10G at home, and take the 1G to work to upgrade their system which is still on 10BaseT

Re:What would I do with this much bandwidth?

[Donny Smith]

> especially databases that stay partially in RAM.

Come on, 10GbE still suffers from latency the same way GbE does and unpredictable nature of database queries renders the idea of holding selected tables in RAM useless for most practical purposes.

The original article sez:

> bandwidth-expensive operations (e.g., database replication)

Bullshit.

They have to write data to something after it comes out on the other end - they can't write to the goddamn void - and as we all know, a server host can write about 150MB/sec which can be saturated with two GbE links - there's no need for 10GbE.

As the linked article concluded, it's very clear 10GbE is currently useful only for core networks and not for the server room (shiiit, big news...).

Yet another useless and speculative article....

Re:What would I do with this much bandwidth?

[deaddrunk]

Re-architect? What the heck does this mean. Redesign? If it does then why not say redesign? Stupid marketing newspeak.

Re:What would I do with this much bandwidth?

[TyrranzzX]

There's far more bottlenecks than that my friend; inter-southbridge comms speed, north/southbridge communication speed, the speed of a PCI Slot itself, harddrive data transfer speed, line transfer speed (harddrive to mobo), etc. If you used a striping raid array it'd go faster since you can double your throughput, but that's if you have a 64-bit 66mhz pci slot.

Once PCIexpress becomes standard, and 64bit procs get onto desktops in a few years, then you'll see a major improvement.

Re:What would I do with this much bandwidth?

[Alioth]

A speed increase of 3 times is worth it, don't you think? Also, when it comes to applications with lots of data in RAM (and wanting to transfer that data to another computer), disk speed is no longer a bottle neck. Not everyone's running a network of consumer grade PCs.

Re:What would I do with this much bandwidth?

[Onan]

...and as we all know, a server host can write about 150MB/sec...

Oh, do we all know that? That's funny, I think that these people seem to know something different.

Re:What would I do with this much bandwidth?

[hitmark]

pci-express anyone? wha as the amount of data a single 16x bus of that system could move again?

but yes this for the moment belongs to backbone connections and server hookups. nothing you will see on your avarage desktop just yet. it allso depends on the server your talking about. your avrage pc based web server will have a problem filling the connection but some of the ibm big irons have dedicated hd controllers that can handle that kind of traffic with ease...

Re:What would I do with this much bandwidth?

[jbplou]

Well if your running load balanced applications servers this can help still. Both Raid 5 and 10 can exceed .7 GBPS if configured correctly. Also most business servers use SCSI.

Re:What would I do with this much bandwidth?

[Donny Smith]

>Oh, do we all know that? That's funny, I think that these people seem to know something different.

No, _your_ comment is funny because you are ignorant about high end storage and you don't know it.

NetApp's high end filers write at about 250MB/sec - even in this scenario 4 GbE links are enough (stil, few companies can afford those models) - and SANs get up to 1.5 GB/sec per controller but any single server host connected to it cannot saturate that bandwidth, which means you usually get about 150MB/sec per host as I said in my earlier post.

Find a URL that shows a single commodity file server capable of over 200MB/sec in I/O (even if there are such servers, 200MB/sec does not justify GbE since it can be filled with three GbE links).

Re:What would I do with this much bandwidth?

[bwilliam13]

It's not bandwidth that's the problem...it's network saturation of many people connecting to many machines at high speed and saturating your switching backplane. Your port speed is simply your maximum potential...it doesn't imply you'll actually ever get that speed. If you use any kind of data encryption at all, that's obvious. I can see practical applications in some scenarios...but not any scenarios that involve any less than 50 users on a switch at one time.

Re:What would I do with this much bandwidth?

[Gabrill]

Here's another thought: MULTIPLE COMPUTERS OVER ONE LINK! All the computers in one office usually connect to one switch, which then has to pipe ALL the data to the next office or the server closet. This is where my particular workplace could really benifit by reducing the number of cables and opening the throughput.

Re:What would I do with this much bandwidth?

[Sepper]

Stupid marketing newspeak.

Marketing likes buzzword... it confuses the hell out of anyone... which may makes more sales...

something like:
Tactically Demand-Responsive Tele-Uncompromising Power-Contemporary Market-Connected Platform-Independant Process-Free Vendor Unimpaired Ressource Communicating Device.

Yeah, I just got this buzzphrase generator for palm... work like a charm :)

Re:What would I do with this much bandwidth?

[Donny Smith]

That's a valid idea - backbone networks use high(er) speed network to connect two or more "normal" speed networks.
But except big enterprises and telcos, very few users have that much data (or can afford the cost).

For example: the simplest 10GbE network needs two 10GbE switches with at least 2 ports, one cable and at least two 10GbE cards.

Now hold onto something and read this: "Starting at a base price of $5,495 and expandable to include up to two 10-GbE ports at $3,250 per port..."
Source:
http://www.gridtoday.com/04/032 2/102902.html

So you'd need some $7,000 dollars for a minimum configuration..

HDTV baby!

[Rectal Prolapse]

1920x1080p, minimal compression, streamed...

HDTV recording...

Porn. Lots of porn.

Obvious isn't it?

Re:HDTV baby!

[WiKKeSH]

HDTV is broadcast at 20mbit, give or take

HD Video for one

[SYFer]

This would be a boon for HD video workflows. I should think it would be attractive to companies like Pixar and the like.

Re:HD Video for one

[serutan]

Not to mention videophone apps. Every employee would have full-screen teleconferencing available any time. That's my bet for what will drive this into office networks bigtime.

Hmm...

[ravenspear]

What would you do with this much bandwidth?"

Check out more unusual positions.

A really big ...

[Compholio]

Doom III LAN party!

Hey, why not?

[AJYeary]

Build an entire slashdot-proof network!

Play original quake obviously

[DarthVeda]

I mean it's not worth playing if you don't have 1 or 2ms ping time...

Re:Play original quake obviously

[jrockway]

I'm under the impression that latency and bandwidth aren't really related. A 747 full of DVDs has a shitload of bandwidth, but damn bad latency. What's to say the same thing isn't happening here?

Also, I'll bet that latency is introduced at each hop by the router. The lookup in the routing table is what takes time, not the light travelling to the end of the fibre.

Re:Play original quake obviously

[hitmark]

what affects latency is qos and the packet queue(and that is again related to bandwidth).

the first give priority to time-dependent packets, like video and audio streams and interactive content like games. ip4 dont realy have a qos support. ip6 have it buildt in so that a packet can be taged with a priority flag if its effect is time related.

the last one is how long a packet will have to wait at a router until it can be moved on. lets say the router have a 10mb line in from the net and a 1mb line to you. this means that it can in theory get 10 times the data in then it can move on to you, therefor a queue builds up. its like a warehouse that have 10 trucks moveing stuff to it and 1 truck that moves stuff on. over time the warehouse will fill up and will have to start tossing out old stuff to make room for new stuff. when that happens you get latency as stuff have to be retransmitted.

and most routers are set so that they try to use the fastest (ie widest&shortest) route to the destination. but sometimes those connections are down and it will have to use lower priority ones. routing table lookups on the backbone of the net is damn fast as 99% of the net is buildt based on static routes.

and your jumbo comment is flawed in that all transfer rates are given in bits pr second. for a jumbo that would be more like terrabits pr day and would in calculation get a very low bandwidth in b/s terms. would you want to make a phone call where they had to record your speech onto some media and send it by physical transport to its destination and the play it back to the person your trying to talk to?

Re:Play original quake obviously

[fitten]

and your jumbo comment is flawed in that all transfer rates are given in bits pr second. for a jumbo that would be more like terrabits pr day and would in calculation get a very low bandwidth in b/s terms.

bits per second is an average transfer over time. That 747 full of DVDs has a huge bps transfer rate.

1TB/day * 1day/86400s * 8bits/byte * 1x10^^12 bytes/TB = 92.6 Mbps ~ 100Mbps (near 100Mbps Ethernet speeds =)

and a 747 can easily carry more than 1TB of data worth of DVDs (1TB of DVD is only ~100 DVDs, which will fit in a breadbox easily - I'd imagine a 747 could carry more on the order of a few PB at a time - which would mean a trasfer rate of over 1000 times the number above, which is faster than any generally available wide area network today - ~ 100 Gbps). However, the latency is high(1 day before the first bit arrives at the destination).

There were times (such as when the Pentium was being taped out) when companies shipped entire HDD rack cabinets to the destination because it was much faster (bandwidth) than transfering it over the network.

The age old saying is: Never underestimate the bandwidth of a station wagon hurtling down the interstate.

That being said, there are tasks that are latency sensitive, as your talking on the phone example shows.

Re:Play original quake obviously

[maharg]

umm. you are kind of correct. two scenarios:

1. A stream of data being pumped via UDP over a WAN which has a satellite link bang in the middle of it. Very high latency i.e. the bits take >600ms to get to the other end, but data can be sent at "wire speed" as there is no acknowledgement of each packet required == potentially massive bandwidth.

2. A large file being FTPed over the same WAN link. FTP typically runs over TCP/IP. TCP requires acknowledgement of each packet being sent. TCP (wrongly) interprets the long Round Trip Time i.e. >1200ms as link congestion and lowers the transmission rate. Oops !

Re:Play original quake obviously

[smallfeet]

You have to count in the time to burn and read all the DVDs. Unless you have thousands of writers and readers this could add considerable time to the latency.

Re:Play original quake obviously

[fitten]

Yeah, but if you took 9 days to write all the CDs, you'd still be at 10 Gbps (9 days of writing, 1 day of flying), which is still pretty high. If you took 99 days to write all the DVDs, you'd be down to 1 Gbps, which is *still* faster than you will transfer anything from LA to Sydney over the normal Internet, even though it took 100 days (1/3 of a year) to arrive =)

Re:Play original quake obviously

[Shakrai]

Yeah, but if you took 9 days to write all the CDs, you'd still be at 10 Gbps (9 days of writing, 1 day of flying), which is still pretty high. If you took 99 days to write all the DVDs, you'd be down to 1 Gbps, which is *still* faster than you will transfer anything from LA to Sydney over the normal Internet, even though it took 100 days (1/3 of a year) to arrive =)

Yeah but if the plane crashes that retransmit time is a real bitch :)

What would I do?

[Anonvmous Coward]

The company I used to work for was sending very high resolution images from multiple cameras uncompressed from one unit to another to perform analytical operations on them. I think they manged to work at a gigabit, but 10 would be much nicer for them.

What would Joe Sixpack do with it? I'm not sure at the moment. Thing is, since we're working within our limitations today it's hard to concieve of whta use it'd be. However, what happens when it becomes commonplace? It does open doors. Imagine if cable companies traded in coax for ethernet. They could easily send uncompressed HDTV. That'd be pretty slick.

Re:What would I do?

[Anonymous Coward]

. Imagine if cable companies traded in coax for ethernet. They could easily send uncompressed HDTV. That'd be pretty slick.

note: i'm pulling most of these numbers out of my ass, as i am too lazy to look up the real values. but they're close enough for our purposes here.

say hdtv is 1280x1024 (yes, i know it's not 4x3), at 24 bit color. that's 3,932,160 bytes per frame, uncompressed. at 60 frames per second, that's 235,929,600 bytes per second, or 1,887,436,800 bits per second, uncompressed. note that this is without sound. with ethernet overhead, that's about 2 gigabits per second PER CHANNEL. assuming that you could even approach a real throughput of 10 gbps on 10gbps ethernet, you'd have 5 channels (with no sound).

care to rethink your statement?

Re:What would I do?

[Alien Being]

"Imagine if cable companies traded in coax for ethernet."

I think you're confusing your layers. Besides, compared to coax, twisted pair is like dixie cups and fishing line.

Re:What would I do?

[nine-times]

Thing is, since we're working within our limitations today it's hard to concieve of whta use it'd be.

Isn't that always the way? I remember having a 20Mhz IBM PS/2 and wondering "How am I going to use all this power?" And the 30 MB hard drive- how would I ever use all that space?

It seems like when we have the capabilities, we find something to do with the extra. HDTV sounds probable, and more bandwidth can only help working over networks on a mass scale (remote home folders and roaming profiles, VNC/Citrix), but you never know. When processors were getting to the 1Ghz point, a bunch of industry analysts were predicting "Now that we have enough power to make working speech-recognition software, we can finally ditch those keyboards!" Yeah, right.

The big concern is, with the extra bandwidth, will Microsoft see this as an opportunity to release new, extra-inefficient network protocols?

Re:What would I do?

[Gilk180]

Well, I get 40, but whatever.

Then there is the other problem that many people seem to be ignoring that ethernet is by design limited to a pretty short distance (I'm too laxy to pull out the networking book). And the fact that because you might have a 10Gbps connection to one other computer, doesn't mean you are gonna have a 10 Gbps connection to anyone else. I know I have 100Mbps within my apartment, but then there is that darn internet connection that tops out at ~2Mbps.

In your world, 'hdtv' might be 1280x1024, but in the world at least a few of us live in, HDTV (notice the caps) isn't, it's 960x540(assuming square pixels) and I imagine less than 24 bit color (All this analog stuff!). I don't know, but I would also guess that it's the same frame rate as regular old TV (~30 fps).

906 x 540 x 16 x 30 = 235 Mb
add sound and round up give 250 Mb and forty channels, with sound. Still not stupendous, but even with some fast, lossless compression that would probably double to 80.

Re:What would I do?

[Osty]

assuming that you could even approach a real throughput of 10 gbps on 10gbps ethernet, you'd have 5 channels (with no sound).

And why would I need more than one channel at a time? Okay, I could see having two, possibly three channels at once (Tivo one, watch another, pic-in-pic a third?), but any more than that is pointless. If you were to provide such a feed, it would be trivial to provide a mechanism to choose what channel you want and only stream that channel.

Re:What would I do?

[rokzy]

>...will Microsoft see this as an opportunity to release new, extra-inefficient network protocols?

yes, every packet will contain an easter egg flight simulator.

Re:What would I do?

[perlchild]

Oddly enough, the article barely mentions 10G over fibre, which would be good(if a bit expensive to put in someone's home). It focuses... on 10GBase-CX4... to rehash a bit more the idea that existing equipment is reusable. And gets even more confusing when it speaks about the advantages of 10Gbase-CX4 in one paragraph, and quotes the sale of fiber equipment(FTTH to be specific) the next.

While I agree it's basically two paragraphs of the same standard, keeping the mediums seperate certainly makes sense if you want to talk about market penetration, mostly because the market penetration for both mediums so far, is radically different.

Coax is ok, TP is a nice hack, both do... what they were designed to do. Fiber is better, but it's not marketed the way that will encourage people to switch just yet(when a 1GE link is firmware upgradable to 10GE, we can talk).

The note at the bottom about desktop use did confuse me though, how would 1GE reduce latency in desktops?? Maybe it's just that I'm used to a different market, but I get the impression that this beneficial aspect of a bigger pipe is only visible in server-to-desktop large non-streaming tranfers? Like say database select queries or spreadsheets/word processing documents? Is that enough to consider it "lower latency", unqualified? When smaller packets, like AIM/mail checking/other regular, small transfers, can take considerably longer, simply because a lot of the larger bandwidth link is really an optimisation for larger packet sizes and such?

Re:What would I do?

[atrus]

HDTV uses MPEG-2 as the compression protocol. Henceforth, it will fit into the allocated airspace for OTA broadcasts (not as much density as analog TV)

Re:What would I do?

[Anonymous Coward]

i'm pulling most of these numbers out of my ass

As someone who actually does know what you're talking about, that's certainly apparent. About the only number you got even close to right was the 1280 horizontal resolution.
The spec for 1980i HDTV is 19.25 Mbps. Far, far lower than the 1887 Mbps you calculated.

Re:What would I do?

[bs_02_06_02]

traded in coax for ethernet You're confusing a physical medium (coax) with a transport protocol (ethernet). Most cable companies are fiber to the neighborhood, and coax to your doorstep.

If I were building a 10 gig-E system, I'd want it to run on fiber. Sorry, I won't trust 10 gig-E to run on twisted pair, coax, or anything less than fiber. It's not worth it.

I'm curious about 10 gig-E.
What's the maximum (practical) utilization going to be? Maybe a couple gig for a full duplex segment with multiple machines? Imagine the collisions! And, if you load the segment up with more machines, imagine the contention! I think it'd be nasty. I'd much rather see a pair of 1/2 duplex segments, one transmit, one receive between 2 boxes. If you want mission critical stuff with maximum bandwidth, that's the way to go. Reliable. More utilization, and no collisions or contention.

Re:What would I do?

[Anonymous Coward]

About the only number you got even close to right was the 1280 horizontal resolution.

Not really... 1280 is only the horizontal resolution for 720p (which is 720 * 16 / 9).

1080i is 1920x1080, either 60 interlaced frames per second or 24/30 progressive.

The parent poster is also talking about uncompressed HDTV, not the 19.25Mbps MPEG2 flavor.

Re:What would I do?

[Pieroxy]

Forget about HDTV dude! It's already taking so much freaking time when I try to save my 600 dpi US letter image that I just scanned.

Animated movie! That is a hog! Even with today's DVI codecs, I can't even play a video over the network. Save high resolution (Well, it's still 720x576, but still that's not that high)

Re:What would I do?

[Lennie]

First things, I'd think of is IPv6 and multicast.

(IPv6 because of the large number of subscribers, maybe ?)

Re:What would I do?

[Detritus]

Uncompressed HDTV is about 1.5 Gbps. Network HDTV feeds are about 45 Mbps (mild compression).

Re:What would I do?

[egomaniac]

In your world, 'hdtv' might be 1280x1024, but in the world at least a few of us live in, HDTV (notice the caps) isn't, it's 960x540(assuming square pixels) and I imagine less than 24 bit color (All this analog stuff!). I don't know, but I would also guess that it's the same frame rate as regular old TV (~30 fps).

There are two HDTV resolutions in current use, known as 720p and 1080i. 720p is 1280x720 60fps, and 1080i is 1920x1080 30fps (60 interlaced fields). Both of them are 24-bit truecolor.

I have no idea where you got 960x540 from, as it does not correspond to any HDTV resolution. I'm also not sure what the reference to "all this analog crap" is supposed to mean, as HDTV broadcasts are entirely digital.

Re:What would I do?

[Thing 1]

yes, every packet will contain an easter egg flight simulator.

Now that's a funny image... "Egg goes up, egg comes down... crack/splat!"

An easter egg in flight just doesn't seem like a fun game.

Re:What would I do?

[mpe]

Coax is ok, TP is a nice hack, both do... what they were designed to do. Fiber is better,

How is fibre better? It's more expensive and complex to install. Especially if the idea is to replace existing cable...

Re:What would I do?

[really?]

HDTV broadcasts are entirely digital

Err ... Japan has been broadcasting "HDTV" in analog for ten years or more; BS channel 9, if you want to know.

Re:What would I do?

[HiThere]

One really good use for this is inter-processor cabling in a cluster.

Even then you'd still want the components close together.

OTOH, this is ignoring the packet size. I don't know what the high speed block size is. (I tend to suspect that it's pretty large.) This might limit is't usefulness for this application, and restrict it to uses where streaming media are appropriate.

Re:What would I do?

[tijsvd]

Then there is the other problem that many people seem to be ignoring that ethernet is by design limited to a pretty short distance

Current networking equipment can transmit 10GE over copper (coaxial cable) only about 25m. However, with the right lasers (extended reach) and fibers (good quality single-mode), you can reach about 70km. The laser tranceivers are about $5K each though...

Re:What would I do?

[tijsvd]

The note at the bottom about desktop use did confuse me though, how would 1GE reduce latency in desktops??

Easy. 1GE operates at a higher frequency than 100Mb. It takes less time to transmit a packet. Cisco for example uses this argument to drive gigabit-to-the-desktop equipment sales.

Re:What would I do?

[Wrath0fb0b]

Did it occur to you that 1920x1080i is really the same amount of information as 960x540p.

540p and 1080i are basically equivalent (at least as far as joe sixpack is concerned).

Re:What would I do?

[cft_128]

Did it occur to you that 1920x1080i is really the same amount of information as 960x540p.
540p and 1080i are basically equivalent (at least as far as joe sixpack is concerned).

Umm, no they are not. Interlacing only affects the verticle resoultion. Assuming 60Hrz for both, 1920x1080i and 1920x540p would have the same datarate but would not look the same. 960x540p would be half the datarate. There have been some studies to what apaprent verticle resolution 1080i has and most point to something in the 600's, more when there is little motion, less with more motion; that is what the end user cares about.

As a side note, most 1080i HDTV cameras cannot get the full 1920 horizontal resution, most are around 1200-1300. My personal favortite is 720p, looks nice and hase no interlacing artifacts. ABC, ESPN and Fox (or will IIRC) use that, everyone else uses 1080i.

Re:What would I do?

[egomaniac]

Japan has different HDTV standards than the rest of the world. As most of Slashdot's readership is American/European, I didn't bother to mention that.

Re:What would I do?

[perlchild]

More resistant to interference, and properly laid pair sets can support greater throughput easier than coax or TP with lower reimplementation in a lot of cases(changing two transceivers is faster than retesting cat 5e cable for cat6).

Re:What would I do?

[julesh]

Both of them are 24-bit truecolor.

Err, I believe the ATSC standard, which is what most people mean by HDTV, uses 4:2:2 subsampled YUV, which is 16 bits per pixel.

Re:What would I do?

[CowboyNick]

I have a several DVDs that I have copie^H^H^H^H^H backed up on a server here that I can play at full res over the network just fine.

Re:What would I do?

[Pieroxy]

DVDs are not the same as DVI. DVI is the compression codec on most digital camcorder. IIRC, it is around 24MB/s, hence a little too high for a 100Mb/s network. Of course if you have gigabit network, you don't have these problems.

DVDs are much lower, around ~8MB/s.

Re:What would I do?

[CowboyNick]

Ah, thanks for clearing that up.

Re:What else?

[Mumbly_Joe6432]

Woah! don't jump to conclusions! Wait...never mind, you're right.

It's true though that if the technology is there, someone will find a use.

silly question

[00zero]

10 Gb is rediculous. 640K should be enough for anyone.

good political satire

Re:silly question

[00zero]

I now apparently advocate extremely dynamic spelling schemes.

good political satire

Re:silly question

[hunterx11]

In the future, spelling will not be a part of the post at all, but will be defined by external Cascading Spelling Sheets. Of course, in order to implement this we all have to learn Chinese.

Re:silly question

[00zero]

I knew I should have read the documentation.
Do jeh! Xie xie!

good political satire

Re:silly question

[Lennie]

The next question ofcourse would be, which chinese ? As I understand it, there are several.

I'm sure some1 here knows it.

Re:silly question

[HermesHuang]

The written language is universal. It's only the spoken dialects that differ.

Re:silly question

[TheLink]

No.

"In an effort to increase literacy, about 2,000 of the characters used in China have been simplified. These simplified characters are also used in Singapore, but in Taiwan, Hong Kong, Macau and Malaysia the traditional characters are still used."

And many of the simplified characters _really_ look different - no resemblance to the traditional versions.

Then there are the stylized versions of these (similar to cursive vs print).

true remote storage transparency

[The Clockwork Troll]

When your network pushes over 1 gigabyte/sec, diskless workstations become a much more interesting possibility.

Typical desktops of the past few years see roughly ~25 megabyte/sec sustained disk throughput (more for SCSI and more recent ATA models). A switched 1 gigabyte/sec network could easily and transparently support 25 remote drives virtually indistinguishable from local storage.

Re:true remote storage transparency

[Jarnis]

Diskless workstations aren't.

They are just (ab)using the disks of the servers. How Uber Are Your Servers(tm)? Show me a server that can sustain that 1 gigabyte a sec disk access to support those workstations... :p

Re:true remote storage transparency

[WindBourne]

Large Ram disk works wonders for a server that is handling 25 systems.

Re:true remote storage transparency

[tftp]

It is not a problem at all if most of your clients use the same, relatively small (and thus cacheable) set of files. For example, a herd of accounting workstations may need access to some s/w package. A diskless client should not need to run 300 different applications all the time; these are generally a single purpose boxes. The more RAM you throw in, the more universal they become. Besides, the disk I/O is not that frequent these days, once you have your app loaded.

Re:true remote storage transparency

[Onan]

Ah, network computers. Those are always a great plan, given that local storage is so expensive and small, and networks are so infinitely reliable.

Taking over the world

[keepr]

With that much bandwidth at my fingertips my dreams of taking over the world might become a reality.

Muahhha ahahahahahahahaha!

A lot more then 10

[mp3LM]

You can use it as a backbone in a building. And in a few years, it will be less then enough for the common user. Remember...over time...most programs get bigger.

Remember back in the day, when someone would have a 10Mb switch, and it was amazing? And 100Mb switches were over the top? Well, now 1Gb switches are amazing, and 10Gb switches are over the top.

To me, this means online games can run better(provided ISPs upgrade there backbones as well...past whatever they are). P.S.-Yes, I know ISPs backbones are more then enough now, but I was stating for future.

Re: A lot more then 10

[tylernt]

There are no collisions on switched Ethernet. These guys aren't using hubs any more. :)

Re: A lot more then 10

[mp3LM]

No one said it was using coper cable.

Re: A lot more then 10

[mp3LM]

I was making a point.

Repeat after me

[Lisandro]

You can never, EVER, download porn fast enough. There's no such thing as "enough bandwidth".

Re:What else?

[RLiegh]

It's true though that if the technology is there, someone will find a use.

And, if the internet and VHS are any indication; that "someone" will probably be the porn industry.

Reminds me of an old quote...

[Anonymous Coward]

Just get us the bandwidth - we'll manage to piss it all away. Easy!

1G should be enough for anyone.
-- Nicholas Cravotta, 2004

640K should be enough for anyone.
-- Bill Gates, 1981

Re:Reminds me of an old quote...

[WindBourne]

we'll manage to piss it all away. Easy!

WHile it will be pissed away at first, somebody will suddenly come up with an innovative idea that requires hire speed. It will go beyond simple transfer of a ripped(-off) movie or music. But it will happen.

Re:Reminds me of an old quote...

[ultranova]

WHile it will be pissed away at first, somebody will suddenly come up with an innovative idea that requires hire speed. It will go beyond simple transfer of a ripped(-off) movie or music. But it will happen.

Has already happened: clusters become much more effective as the speed of communication between individual nodes increases.

Imagine buying a future with nearly every computer running some kind of clustering software, forming a worldwide supercomputer with their spare resources. Imagine what you could do with such a thing.

We'd need 256 bytes just for process identifiers ;).

Re:Reminds me of an old quote...

[harkabeeparolyn]

Imagine buying a future with nearly every computer running some kind of clustering software, forming a worldwide supercomputer with their spare resources. Imagine what you could do with such a thing.

It's been imagined. The movies all seem to have the same answer: take over the world and destroy it.

Re:Reminds me of an old quote...

[PaulBu]

clusters become much more effective as the speed of communication between individual nodes increases

Clusters become much more efficient when LATENCY between nodes decreases, not when bandwidth ("speed") increases. Clusters want to exchange rather small packets and get the replies back as soon as possible, the actual time it takes to transmit a small packet does not depend that much on the link bandwidth, but on how fast you can get the packet to the link and pick it up on the other end.

Paul B.

Re:Reminds me of an old quote...

[ultranova]

Clusters become much more efficient when LATENCY between nodes decreases, not when bandwidth ("speed") increases. Clusters want to exchange rather small packets and get the replies back as soon as possible, the actual time it takes to transmit a small packet does not depend that much on the link bandwidth, but on how fast you can get the packet to the link and pick it up on the other end.

The faster the link is, the less time it takes to transfer that single packet -> less latency. Also, at least my Internet connection starts getting much higher ping times when it's loaded past a certain point, presumably because there's packets waiting in line. A high-bandwith line, one with so much bandwith that all the computers connected to it simply couldn't saturate it, would not develop this problem.

Also, with enough bandwith, it becomes possible to run programs that share memory in different machines. With low bandwith, the speed which changes in the shared memory area can be communicated from machine to machine becomes the bottleneck, but with ultra-high bandwith, it is no big problem.

And yes, you need low latency for that too, to make acquiring locks an operation that concludes in a reasonable time.

Worked for a medical company

[RoundTop-VJAS]

I worked for a medical imaging company and they would use it.

they are using gigabit already and you can see slowdown...simply put, a couple hundred 100MB+ x-rays to a single box.... multiply that by however many boxes the hospital has..and 10 gigabit is nice.

The problem hits in not having enough RAM..and with a 4GB limitation on workstation OS's for the most part this amount of bandwidth could get funky.

Re:Worked for a medical company

[phsdv]

4GB limitation on workstations? What brand of station are you using? OK, I know many are limited at 4G (or 8G). But we are daily dealing with files/product databases that are over 4GByte in size. And we have workstations that can handle these files. They have 32 or 64GByte of memory on board!

And yes I would love to have 10Gbps network on those machines. Now when I load such a file I can go drink a cup of coffee. ( No, I can not go surf porn, we have limited internet access :-( )

Re:Worked for a medical company

[name773]

i'll bet you the parent poster ment 32 bit workstations...
pity, 64bit r0cks :)

Cant use the bandwidth anyway

[PimpbotChris]

From TFA: " The Windows operating system imposes one of the primary bottlenecks at this speed. As one network administrator says, "When we want to stress test our network, we use Linux, not Microsoft." His experience is that contention and file-system-overhead issues within Windows limits 1 GE desktop links to 1.25 Mbps. Even a quad-processor server peaks at 250 to 300 Mbps, with the processor at 100% usage."

So if we cant use the extra bandwidth then whats the point of having it. You can already get 100 Mbps dedicated from a switch to the desktop with old technology

Re:Cant use the bandwidth anyway

[HermanAB]

Uhm, well, you gave the answer yourself - use Linux.

Re:Cant use the bandwidth anyway

[rokzy]

if you're going to have a network like this of course you aren't going to use Windows.

unless you're the kind of wanker^H^H^H^H^H^H person that likes to buy a $300 car and put $3000 rims on it.

Remote Virtual Immersion

[weston]

But just how much data can a person consume?

If I was going under the knife remotely, I'd want the surgeon to have as much bandwidth as possible (and very, very, very low latency).

Re:Remote Virtual Immersion

[gnuman99]

If I was going under the knife remotely, I'd want the surgeon to have as much bandwidth as possible (and very, very, very low latency).

Instead of very low latency, I would prefer no lost packets and *smooth* motion, and not that jagged back and forth you sometimes get! Ouch!

Re:Remote Virtual Immersion

[MP3Chuck]

Imagine a malpractice lawsuit!

Exhibit A: Surgery Log
[DR]Surgeon opened Xx[Patient]xX's abdomen with a scalpel.
[DR]Surgeon punctures Xx[Patient]xX's stomach with forceps.
Xx[Patient]xX: OMGWTF??!!
[DR]Assistant: ROFL PWNED!!1
[DR]Surgeon: STFU N00B i ping 350
Xx[Patient]xX: w/e

Re:Remote Virtual Immersion

[hurricane_sh]

Simply the more, the better, it would be cool to download a big game, software, even movie in minutes

WWID?

[z0ink]

How about setup a digital broadcast network in my house?

10 gigabit is kinda much

[Jarnis]

With 10 gigabit LAN, the bottleneck won't be the LAN. It will be your servers. Their I/O busses, disk systems etc.

Even at 1 gigabit, usually the bottleneck is elsewhere.

10 Gigabits = roughly 1 gigabytes/sec. Considering that PCI bus is 33MB/sec, and even PCI-X is 66MB/sec... Heck the memory bus of my brand new system is only about 1 gigabytes a second.

Re:10 gigabit is kinda much

[cynical kane]

PCI is 33Mhz, not 33 MB/sec. 33 X 32-bit-bus = 133 MB/sec. PCI-X goes up to 133 Mhz and 64-bit, so that's 800 MB/sec.

Re:10 gigabit is kinda much

[dewpac]

Actually it's 1064MB/sec...

Double the bus width from 32 to 64 bits and you double from 133MB/sec to 266MB/sec

Now 4x the Mhz from 33 to 133.. 266 * 4 = 1064MB/sec.

Re:10 gigabit is kinda much

[gunpowder]

Almost correct.

Most Mainboards have only 33MHz/32bit PCI slots.
However "Conventional PCI" actually goes up to 66MHz/64bit = (533 MB/s), although you'll only find that in older workstation/server mainboards (newer ones often use PCI-X instead).

Then there is PCI-X. The latest PCI-X standard specifies PCI-X 533 (64bit), giving you a theoretical I/O limit of 4.3 GB/s.

And of course there is this new PCI-Express, with a transfer rate of 250 MB/s for a single lane (and you can bundle 32 of them = 8 GB/s). However most new mainboards that offer PCI-Express come with only one 16X-PCI-Express slot (as replacement for AGP) and a couple of 1X-PCI-Express slots (still faster than 'normal' PCI).

What would we do with it?

[JoeLinux]

"porn. Lots of porn" /Neo

Video, lots-o-data

[1337 Twinkie]

I could definitely use this for transferring large amounts of video between my PC and my other PC that has a DVD burner. I suppose that companies like Pixar would be interested in this regard.

There are also companies (finacial, law, etc...) that could need to transfer huge documents (1000+ pages) between people. Huge databases could also be accesed better, I suppose.

To be politically correct on /.

[deunan_k]

Of course, Pr0n, pr0n and even more pr0n! (notice that I use pr0n instead of porn!)

Well, being an anime otaku, instead it'll be hentai, hentai and even more hentai!

Uuhhh, just imagine those tentacles! :-P

Seriously, with that much of bandwidth available, the world keep getting smaller.. Sooner or later, it'll get crowded! Then'll we be in trouble.. Oh man!

It's either survival of the fittest, or... We go to 'Space.. The Final Frontier.. These are the voyages of...' you get my drift...

But when?

Re:To be politically correct on /.

[PingPongBoy]

with that much of bandwidth available, the world keep getting smaller.. Sooner or later, it'll get crowded!

I second that. You'll have to admit that this is the kind of bandwidth that will make us involuntarily stuff the entire Internet into the same office tower.

Come to think of it, the sun outputs so much power why don't we build a Dyson sphere?

FTP Installs of linux/BSD

[Crackez]

On my 8 different machines, at the same time...

Of course, each one is less than 1GHz, so being able to handle 10Gbit is not very likely. Damn, most of em' can't even max out 100Mbit, stupid slow harddrives...

Re:FTP Installs of linux/BSD

[Crackez]

However, I just thought, this would sure help out with all those Beowulf clusters people like to imagine around here...

Holo-Porn

[Hallowed]

That might be just enough bandwidth to get a life-like signal to the holographic projector!

it goes without saying

[B1ood]

i'd setup a usenet server.

Re:it goes without saying

[MavEtJu]

-1 Redundant: see all previour porn-postings :-)

Funny little sidebar...

[lofi-rev]

The Windows operating system imposes one of the primary bottlenecks at this speed. As one network administrator says, "When we want to stress test our network, we use Linux, not Microsoft." His experience is that contention and file-system-overhead issues within Windows limits 1 GE desktop links to 1.25 Mbps. Even a quad-processor server peaks at 250 to 300 Mbps, with the processor at 100% usage.

Yeah I'm a snob, but at least my OS works.

Re:Funny little sidebar...

[JamesKPolk]

OK, you linked to your kernel (for now... those guys are giving up on making stable releases for users as of this week). But what OS are you running?

Re:Funny little sidebar...

[lofi-rev]

I certainly concede that my post was unclear, but it's really the kernel that matters here - the userspace stuff has little impact. For the record, I use Debian, Gentoo, and a little Slack when I have to.

What would I do?-Horror-vision.

[Anonymous Coward]

"What would Joe Sixpack do with it? I'm not sure at the moment. Thing is, since we're working within our limitations today it's hard to concieve of whta use it'd be."

The Goatse.cx experience in holographic, 5.1 surround-sound, smello-tactile-vision.

The Network Is The Computer (tm)

[gusnz]

OK, so for stuff like streaming MP3s and so forth, this is a little overkill for the current style of usage. However, where I think this will come in useful is for stuff like remote disk and memory access over IP.

With a 10G LAN, you'd be able to come up with a great distributed computer system (e.g. for compiling software). IIRC protocols are in the works now for native-ish memory access over networks, turning a network into one huge computer, and you can already access remote disks with the right software. Imagine the simultaneous distributed encoding of several HDTV streams to redundant archives on several different computers, and you'll probably find that more bandwidth = better.

So yeah, there'll definitely be possibilities for this sort of stuff, even if it is only as a base requirement for the post-Longhorn Windows version :).

Re:The Network Is The Computer (tm)

[TubeSteak]

We're going to run out of storage space long before we can really use that much bandwidth to do the things you're describing. P2P networks will essentially become the defacto method of archiving things like HDTV streams or anything else that is large & not confidential.

I also imagine that we'll discover just how stable our computers, NIC's & their drivers are. My win2k box (which is way past due for a format/reinstall) tends to bluescreen when pulling anything around 400K~500K for any significant period of time from the internet. It has zero problems over the lan, it just doesn't like the intarweb.

Re:The Network Is The Computer (tm)

[IceFox]

Well in all my tests of compiling software using multiple computers the network did have a difference, but only in two cases. The first case was if it was slower then 100 (say 10 or 2). If you were using 100 or 1000 you were fine. The second cases was if you only had one nic in the server and had some large number of nodes (50 was the most I tested with and it was still ok). As more computers were connected you get more and more collisions, but even this can be resolved just by putting the computers on different networks and adding more nics to the host computer.

So having 10000 speed network would be cool network compiling is here and network speed improvments will be little in comparision to moving to better build systems that can build in parallel (vs autoconf which does it directory based)

You can check out my full article here:
http://www.csh.rit.edu/~benjamin/articles/o ptimizi ng_distcc.php

-Benjamin Meyer

In short

[krray]

In short I'd have to replace all my NIC's and switches. That sucks. Considering that the move _recently_ made from 100Mbit to 1Gbit (LAN) with the steady 10Mbit uplink to the Internet.

The major problem (today) with 10Gbit? None of the sub-systems could handle the bandwidth. The absolutely rockin' stations with SCSI Ultra-320 (like my Mac @ home for example :) simply couldn't handle the bandwidth. At the bus level or at the hard drive level. So in addition to replacing NIC's and switches we'd be completely replacing computers too. Not going to happen any time soon for the majority of the systems...

Considering that I easily handle multiple VoIP lines and general Internet traffic along with video with the 10Mbit uplink to the Internet and constantly move data around which completely saturates the hard drive and bus bandwidth of the majority of local system ... I have no use for 10Gbit. Anywhere.

Yet. :)

Imagine a beowulf cluster...

[3) profit!!!]

Seems like this would be useful for people trying to build clusters with commodity hardware.

Re:Imagine a beowulf cluster...

[rokzy]

as opposed to beowulfs built with non-commodity hardware?

beowulfs does seem like one of the best uses, either 10 gigabit or at least push the cost of 1 gigabit LANs right down.

Re:Imagine a beowulf cluster...

[jmv]

Not that useful actually. Right now, the main problem for clusters is not the network bandwidth, but the network latency. AFAIK, that's why Virginia Tech used Infiniband and not Ethernet.

Distributed computing?

[erice]

For distributing intermediate results, I don't imagine there is such a thing as too fast.

While there are certainly applications that don't need to communicate that fast, more bandwidth means more alogrithms can become practical.

It's not like you can use it download porn, unless the action is happening in the next room. This is not a WAN technology.

Re:Distributed computing?

[Shuasha]

This is not a WAN technology Actually, yes it is. I currently sell a lot of 1 GB/sec pipes that can go up to 60 miles, and we will be selling 10 GB very soon. I know that some colleges (University of Illinois) are currently experimenting with multiple 10 GB links across 120 miles or more.

Beats me...

[dopefish3]

On a simular note, what would you do with 640k?

NC-PC-NC

[basking2]

So, we used to have little dumb terminals that talked to the big smart backend. Then computer became cheaper and we had Personal Computers, but we have to manage and distribute all these updates and it's a real pain and it sometimes destroys your computer during the upgrade/install process. Now we can swing the pendulum back towards the Network Computer a little more.

This isn't a new idea. Software companies like MS would love to sell you a subscription to MS Office which you renew and they in turn patch and maintain the software on your company's server or on the MS servers. It's a neat idea for sure. Companies like Novel have made some interesting claims about Network Computers.

There is also the whole Plan9 type of mentality too.

Re:NC-PC-NC

[MoNsTeR]

That's what they said ~8 years ago when everyone was pitching "thin clients".

I don't think the rekindling of the NC concept died on the vine because of lack of bandwidth. I think the issue is that today, people have PCs at home, and there's a lot of value in having the same kind of computer at home as at work. Migrating to some kind of NC-based computing in the enterprise would inevitably break some of that "skills compatibility".

Of course, part of the reason for that is that it wasn't Microsoft pushing the NC's; it was more of a Sun/Java thing. If MS tried it, there's a good chance they could make Windows NC Edition and what-have-you look and feel enough like the home version to make the changeover relatively painless.

I'm still skeptical though.

Posts mentioning Porn

[munch0wnsy0u]

I honestly cannot understand the Slashdot crowds (or maybe a selects few's) obsession with pornography. I read the title to this article and was reminded that the ensuing discussion would be centered around the act of "downloading more and more porn." Seriously gentleman, move beyond the teenage years and enter into a discussion that isn't focused around the act of parading women. Sorry, it isn't funny, intelligent, and I think most readers would say that you are simply embarassing yourself.

Re:Posts mentioning Porn

[lachlan76]

Uhhhhh....no

Us geeks without girlfriends, our pr0n usage scales O(n^2) with bandwidth. So if there's one thing we want, it's more bandwidth, and better compression.

And some of us are teenagers.

Re:Posts mentioning Porn

[Anonymous Coward]

move beyond the teenage years and enter into a discussion that isn't focused around the act of parading women

No, thanks :-)

Re:Posts mentioning Porn

[Anonymous Coward]

"Seriously gentleman, move beyond the teenage years"
I think it's you who is the naive child.

Porn continues to be one of the leading drivers of technology (war being the other one) having "made" the VCR, VideoCD, color-printing, Video Streaming, and many other industries.

Porn also continues to be a serious business, with the New York Times (may 18 cover story) claiming Pornography has $10 - $14 billion in annual sales - bigger than any major sports league.

The porn industry employes 12,000 people in California alone, contributing $36 million in taxes to the state. Comcast makes $50 million per year on cable programming. DirecTV is estimated (by CBS) to make $500 million/year.

Recent estimates of Internet porn are that it's about a 1 billion dollar industry.

CBS reports that "Consumer demand is so strong that it has seduced some of America's biggest brand names, and companies like General Motors, Marriott and Time Warner are now making millions selling erotica to America. Correspondent Steve Kroft reports.".

The CBS article claims that Hilton, Marriot, Hyatt, Sheraton and Holiday Inn all offer pay-per-view porn and such programming is purchased by 50 percent(!!!) of guests accounting for 70% of in-room-profits.

" Sorry, it isn't funny, intelligent, and I think most readers would say that you are simply embarassing yourself"
We're not laughing it it -- we're respecting it as an important technology driver, and serious business that is very important to both broadband communication (this article) and other technologies including digital video, alternative payment mechanisms (phone-sex is the biggest driver for phone-calling-cards), etc.

Perhaps one day when you reach puberty you'll understand too.

Re:Posts mentioning Porn

[bigdreamer]

Seriously gentleman, move beyond the teenage years and enter into a discussion that isn't focused around the act of parading women. Sorry, it isn't funny, intelligent, and I think most readers would say that you are simply embarassing yourself.

You must be new here.

Speaking as a heterosexual female in a committed relationship, even I enjoy watching pornography every once in a while. It's not a terrible thing. Besides, after you hang around male geeks for a while, you'll realize that many of them are seriously sexually deprived. Porn is all they have, which is really sad, actually. Makes me want to clone myself to satisfy the male geek population. ;)

It's not the bandwidth

[overshoot]

It's the latency. No matter what your bandwidth may be, some tasks (e.g. file servers) need to be "close" to keep latency from being nasty.

"Close" applies both in physical distance (I have to count picoseconds for the kind of stuff I do) and in network distance, since every router adds considerably.

For some jobs (backup is a classic) latency is relatively tolerable. However, even for those you have to watch out because one error can cause the whole process to back up for retries. Ten to the minus fifteen BER sounds good until you look at what it can do to your throughput in a long-latency environment.

Re:It's not the bandwidth

[kasperd]

I have to count picoseconds for the kind of stuff I do

Unless you are working with individual gates inside a chip, I doubt picoseconds really matters. On ethernet we are certainly not talking picoseconds. We are still limited by the speed of light, so it would take the signal 100 picoseconds just to get through the RJ45 connector. With a 1.5m ethernet cable there will be at least 10 nanoseconds of roundtrip time, because that is the time it takes light to travel 3m.

Re:It's not the bandwidth

[OeLeWaPpErKe]

If you want real throughput at very high latency, you should not use tcp. IP/UDP with the right protocol over it (say ... something zmodem alike) will use all available bandwidth regardless of latency.

Re:It's not the bandwidth

[Some Dumbass...]

I have to count picoseconds for the kind of stuff I do

Unless you are working with individual gates inside a chip, I doubt picoseconds really matters.

I think you're missing something. If the cabling adds a constant delay to any times this guy's measuring, then he can still measure times in picoseconds (assuming his timer is accurate enough, of course). The fact that network cabling would add nanoseconds to a recorded time is irrelevant. Just as long as it doesn't add a variable delay (I wouldn't recommend doing this timing through any sort of switch or router, for example).

Not that this guy is necessarily using ethernet for what he's doing. Note that he didn't actually say that -- he just said that you had to be close for the kind of stuff he does.

One possibility is that the guy's a physicist working with a particle detector. He's could be talking about detecting the exact timing of the decay of various particles. If these decays occur on the order of picoseconds, and his equipment can accurately keep time in picoseconds, then the fact that the cabling adds, say, 5ns to all of the measured times is no big deal. Just subtract 5ns from everything. That's good enough to get the relative times of all the measured events, e.g. the amount of time between the detection of emissions created by the initial collision (and thus presumably particle creation) and the decay of the various particles.

Re:It's not the bandwidth

[overshoot]

Unless you are working with individual gates inside a chip, I doubt picoseconds really matters.

More like individual transistors, but it's all oriented towards getting traffic between chips. For some of the stuff that I'm doing the total chip-to-chip budget is less than a nanosecond (and yes, that means that they have to be really friendly.)

Re:It's not the bandwidth

[maunleon]

I think that the limit you discuss can be bypassed by compression and parallelism. ;) Sure, there may be 10 ns on one wire, but 1 ns on 10 wires.

Also if someone gave up this binary system and figured out how to represent a decimal number (or higher) in one unit of information over the wire, that would speed up things.

That's exactly the quote I remembered

[empaler]

when I saw the news item.

I mean... there's no such thing as too much of anything in computers. When's the last time you said "Accursed be this high transfer rate" or "I wish the computer had less RAM so it would swap more!".

Come on.

Re:That's exactly the quote I remembered

[SillyNickName4me]

Well..
I wish my compu had less ram... so that a system dump takes a bit shorter.. ;)

Re:That's exactly the quote I remembered

[TheToon]

I haven't said that, but a columnist in Byte Magazine in the mid-80s had a rant about this.

He programmed on a Mac, and the compilation took typically 5 to 10 minutes. Enough to get a cup of coffee, check the newspaper and have a quick chat with a cow-orker. Then he got a new Mac, and it compiled the program in a minute or so. No time for coffee, no time for news, no time for smalltalk.

So the new, faster computer was too fast... he had to wait at his desk more with the new computer.

Re:That's exactly the quote I remembered

[DMUTPeregrine]

SATA or SCSI hard drive, sir, SATA or SCSI. SATA is nice and fast, SCSI can be faster, but the huge cables restrict airflow a lot. ugh. (For the record, I've overclocked. I have 9 fans. And yes, SATA.

Re:That's exactly the quote I remembered

[SillyNickName4me]

And I have scsi. No matter how fast your disks, more memory will still make a dump take longer.

backups

[Squeezer]

we have several servers at work that tar.bz2 themselves up nightly and then scp the large .tar.bz2 file over to another server connected to an 8 tape autoloader. all the servers go through a copper cisco gigabit switch and even at gigabut speeds you are topping out at about 23megabytes/sec and transfering a 20GB+ file at this speed still takes several minutes. The servers all use raid 0 and can read faster then 23megabytes/sec so the quicker it can be transfered to the tape connected backup server the better.

Re:backups

[Squeezer]

tar doesn't have an incremental option i'm aware of.

Re:backups

[Richard_at_work]

I had problems with this form of backup once the tarballs got over the max file size limit of the filesystem, so I moved to rsync over ssh, all remotely done. It now takes about 3 hours to fully back up everything, and when I want to restore a file I dont have to untar the backup.

Re:backups

[TheLink]

Why aren't you getting more than 23MB/sec? Running out of CPU with your scp?

Have you tried a nonencrypted network copy of similar sized dummy data? How fast is it?

If the data doesn't change that much every night and you have enough disk space on the tape server you could try using cvsup. But you need enough CPU to keep up.

Lots

[BrookHarty]

We use lots of shared drives, remote desktop applications, X traffic, moving core files, database dumps, email with very large attachments (exchange to boot).

We migrated to 100meg, it was like night and day, and we still need more. We finally got 1gig to IT's network, and still to slow to push files with lots of users.

We have a burstable OC192 to our 2nd remote datacenter, OC48/12's to the smaller datacenters. But this is for production networks that need bandwidth, not desktop usage.

Also, my buddy in Japan just told me he got 100Meg DSL, the stuff you can do when bandwidth isn't a concern. Already Internet TV stations popping up there, amazing. Can't wait for this to catch on in the US. I just upgraded to 6M DSL from speakeasy, and its too fast for fileplanet.

Speed kills :)

Re:Lots

[throbber]

Congratulations! You'll be able to do the same stuff as you do now but faster. That's not exactly anything earth-shatteringly revolutionary. And guess what ... you'll still be waiting the same time for "very large" attachments to arrive in your email.
It's just that you'll be able to zoom in see the cells in Cameron Diaz's nipple and the phishing mails will include the entire website.

In threads from this article I've seen all sorts of stuff about HDTV over the Internet, Internet TV stations and all sorts of video related technology.
Quite frankly, a crap story presented in high resolution video, even over the Internet, is still a crap story; and if Internet Video is anything like the Internet Radio I've experienced then I'll take current version where I don't have to pay for the bandwidth, thank-you very much.

I guess what I'm saying is that most of the ideas I've seen presented here are just more of what we've got now, either on the current Internet or other media.

Are games these days really that much better than 10-15 years ago for all the extra computing power that we have now?

Oh yeah ... I'ts acceleration that Kills BTW ;-)

Re:Lots

[thinkninja]

Meanwhile in Broadband Britain[sic(k)]...

I believe the fastest DSL one can get is 4096/512 (in inner London only, iirc). :/

What's your buddy's upload, 25Mbits? Can he bond several lines to give himself 400Mbits?(!) Is 100Mbits only available in Tokyo?

Re:Lots

[Lord_Dweomer]

"Also, my buddy in Japan just told me he got 100Meg DSL, the stuff you can do when bandwidth isn't a concern. Already Internet TV stations popping up there, amazing."

For those of us who speak english but would be interested in these tv stations, could you possibly post some URLs for us to check them out at?

Yes, because as we all know,

[empaler]

women find sex repulsive, and so do anyone beyond the age of nineteen.

Except one of my friends. She just borrowed a porn VHS from her teenage brother. Then again, she doesn't read /.

play an entire orc army

[Machine9]

on world of warcraft, using a single router to hook several dozens of PCs to the net.

You won't actually have to control the orcs, the mere sight of them on your screen will initiate instant lag-death for people with lesser video cards.

Number four,

[empaler]

Crunchy frog, the crunchy frog.

What is that in MegaBytes per Hour?

[wadiwood]

why is it that network and modem speeds are measured in bits per second but hard disk space and ISP download limits are in BYTES?

I want to know at 10Gbps, how many Meg per hour is that? How long would it take to blow my ISP download limit (4GB) or fill up the hard disk (120GB)?

If I tune into an online radio at 20bps - how many MB is that per hour? Even worse are the audio or video files available for download that say they are "five minutes long" but don't bother mentioning how many Bytes or bits at all.

It's not quite as bad as trying to convert from Metric to American Imperial measurements which are not the same as English Imperial measurements but not far off.

Re:What is that in MegaBytes per Hour?

[imemyself]

A byte is 8 bits. So, 20 bits/sec = 2.5 bytes/sec. 10 Gbs = 1,310,720 bytes per second = 1.25 Gigabytes per second. In other word...unbelievable friggin fast. What makes it really fun is that, your 120 GB hard drive isn't really 120 GB. Its really about 114 gigs becuase the hard drive manufactures use 1 KB = 1000 bytes, instead of 1 KB = 1024 bytes.

Re:What is that in MegaBytes per Hour?

[mindstrm]

Answer: because laypeople insist on talking in imprecise terms like kilobytes and whatnot. Even the byte, historically, could be of varied size depending on the architecture.

When talking about bandwidth, always use bits, and always use k=1000.

Further, how much useful data transfer you get out of the system is not an accurate number.. it fluctuates based on a number of factors, including the network itself, quality of equipment, protocol stack and version, stack settings, local hardware speeds, etc.

However, what we DO know is that the medium transfers *exactly* 10 billion bits per second, no more, no less.

Transmission speeds are measured atthe base rate they transmit data at, without taking into account the protocol in use generally. It has to be this way, because everything else varies upon use. 11Mbps wifi in no way lets you transmit 11mbps of useful data, or anywhere near that, between two hosts.. but the data rate on air is precisely 11Mbps.

A fully utilized 100Mbps ethernet hub will have exactly 100 million bits per second going through it.. yet it is impossible for a single host to transmit at 100mbps continuously.. there are mandatory pauses in between frames, and stuff like that.

Further to that.. to add any kind of meaning whatsoever to download limits, if it's a service you pay for, you need to inquire to precisely how such things are calculated.

Re:What is that in MegaBytes per Hour?

[PeeCee]

why is it that network and modem speeds are measured in bits per second but hard disk space and ISP download limits are in BYTES?

I would guess historical reasons. The amounts of data transferred over networks have always been several orders of magnitude less than the data you could store. So a 300bps modem made a lot more sense than 37.5 bytes/sec. Also, when you think of network links, you usually think a lot closer to the physical layer, in terms of how data is signalled, and stuff there is always considered at the bit level. For storage systems, which deal with much larger amounts, this doesn't really matter all that much.

I want to know at 10Gbps, how many Meg per hour is that?

Wadiwood, meet Calculator. 1 Byte (B) = 8 bits (b), so 10Gbps = 10/8 GBps = 1.250GBps = 1250MBps. An hour has 60 minutes, each with 60 seconds, so 1 hour = 60*60 = 3600 seconds. We just calculated how much we could transfer in a second, so in an hour we could transfer 1250*3600 = 4500000 MB = 4500 GB (note that this is a rough approximation, since it is not exactly true that 1GB = 1000MB, but it'll do for now). So that would be maybe about 1000 (single-sided, single-layered) DVDs per hour.

One think you really have to take into account is that these are maximum theoretical speeds, which are hardly ever 100% achievable (depends on the medium, distance, and devices involved, to name a few). Also important is that this doesn't consider protocol overhead; it is the raw number of bits that would go through the pipe, but there is Ethernet and (probably) TCP/IP control information to consider there. There are some figures on how much is lost on average to these (can't remember; maybe 20%?)

How long would it take to blow my ISP download limit (4GB) or fill up the hard disk (120GB)?

Well, since we just got 1.25GBps (still in the theoretical 100%-performance realm), that would be 4/1.25 = 3.2 seconds for your ISP's download limit (you really have to lose that ISP man), and 120/1.25 = 96 seconds, or a bit over one and a half minutes to fill up your hard disk. Another hint: it ain't gonna happen; all the other components in your computer involved in the filling of your hard disk (speed of the drive itself, the bus, the CPU, the filesystem overhead) will probably become a bottleneck much much sooner.

Well, I think that's all the calculating I'll do for tonight, I'm sure you can continue yourself. BTW I'm somewhat tired right now so feel free to correct me if my numbers are off (hopefully not by too much).

- PeeCee

Re:What is that in MegaBytes per Hour?

[jerde]

why is it that network and modem speeds are measured in bits per second but hard disk space and ISP download limits are in BYTES?

Well, it's not always easy to compute. Dividing by 8 is the first step, but you also have to count all of the overhead used by the various protocols at play.

PPPoA DSL in particular has a very high overhead due to the ATM layer. I calculated the theoretical overhead for a TCP data stream to be 13.9%, and in the real world I found closer to 18%.

Each technology is different, of course. A PtP T1 has 0.5% overhead of its own; so you get just the 4-5% of TCP/IP.

One more "gotcha": All telecommunications units are true metric units. A 10Gbps link is exactly 10,000,000,000 bits per second. To turn that into "binary" GBps, you'd have to divide by 8, and then by 1024*1024*1024.

(BTW, I reject this notion of gibibytes, abbreviated GiB, just because it sounds so stupid.)

- Peter

Re:What is that in MegaBytes per Hour?

[Detritus]

A byte is usually 8-bits, but not always. That's why standards documents talk about octets instead of bytes for 8-bit quantities.

In reference to the grandparent post, communications engineering uses the bit as the fundamental unit. Everything is measured in bits per second. A kilobit is 1000, a megabit is 1000000, etc. A byte is a higher-level abstraction that is irrelevant in the world of information theory and communications channels.

Presidential Bioinformatics

[buckhead_buddy]

But just how much data can a person consume?'

100 Megabytes per chromosome
x 23 chromosomes per gamete
x 20 million gametes per ejaculation

Therefore Ms. Lewinsky can consume roughly 46,000,000,000 megabytes
(assuming that there is no overflow to a dress)

How much can you consume?

Re:Presidential Bioinformatics

[Anonymous Coward]

Close, but not entirely accurate.

A single gamete has 1.5 billion individual base pairs. Of course, that's base-4, since DNA doesn't work off of binary. ACGT is what you're made of. :) In other words, you have 3 billion bits per DNA strand. The average male ejaculation contains around 150 million sperm. This means that there is a total of 450,000,000,000,000,000 bits of information, which turns into 56.25 petabytes of information. That number is close to yours, but your information is wrong.

The fact that I just corrected you is pretty sad as well.

Re:Presidential Bioinformatics

[buckhead_buddy]

I knew someone would correct my figures, but I don't mind at all.

A Lewinsky joke just seemed more appealing to the slashdot audience to express the idea of DNA size being an enormous data handling problem (especially if I wasn't sure of my numbers as you've pointed out).

My sister is the real genetic statistician, but I have a mild interest in her field just because of the massive size of data sets involved in those calculations.

I did expect the first reply to my post to be one comparing Monica's data capacity for DNA to be morally and quantitatively dwarfed by the DNA tracking ideas that John Ashcroft appears to have a thirst for.

Re:Presidential Bioinformatics

[macdaddy]

You were on such a roll I doubt anyone wanted to interupt you. ;-)

Re:Presidential Bioinformatics

[Pieroxy]

Actually, not quite true. One sperm contains only half your DNA. Combined with half the DNA of your sexual partner, that's what makes everyone unique.

Two sperms (although unlikely) could contain your entire DNA. But not one.

Re:Presidential Bioinformatics

[haunebu]

Bill Gates could buy each one of those bytes for a buck a piece and still have money to burn. Pretty staggering that.

Way overkill

[JRHelgeson]

As a CCIE, I have been designing networks for years. I have analyzed traffic to/from desktops and watched traffic to the average desktop never even get above 27mbps. This is due to the average file size of the transfer which is rarely above 10 megabytes. At 10 megs, it only takes a few seconds to get it transfered and it only has a few seconds to get up to speed, by the time it gets all revved up, the file transfer is complete.

High-end workstations such as CAD with gigabit connections, working with 500 mb files, or multi-gigabit video files will occasionally reach 500 to 600 mbps, and even then only for a couple of seconds. At these speeds, power users can use that network connection as if it were a local drive, because at those speeds you are matching the speed at which you're reading/writing data to your local hard drive.

The only time I've ever seen near gigabit traffic at a steady pace was at network servers, where traffic can reach a steady 600mbps on a single gig link - which is maxing out the speed at which the server drive can read/write data to its hard drive. Think of it this way, a 1 gigaBIT link can transfer a 1 gigaBYTE file in about 10 seconds, that's FAST! Conversely, it takes nearly 20-30 seconds just to write that large a file to the hard drive.

Now, on a Cisco 6500 core switch, or a Cisco GSR 12000 where traffic is aggregated, these are the only places where I've actually seen multi-gigabit traffic rates, and that was across the switch fabric - not all directed to a single interface.

The 12000 GSR already has a 10gb interface, it is a single line card that takes up a full slot. It sells for about $60,000 and is used to move data from the switch fabric of one GSR to another GSR, which means you need to put in 2 of them at a mere $120,000 to get the two connected.

Moving to optical links, you can get up to 36Gbps using Dense Wavelength Division Multiplexing on multimode fiber. This uses several colors of laser light to transmit multiple 'channels' across a single fiber link.

Even at these tremendous speeds, they are only used at traffic aggregation points, again because any network device, even a turbocharged SAN couldn't handle reading/writing at those speeds for anything longer than a quick burst.

I say this: If you think that 10gig/sec is your answer, you're looking at the wrong problem. You can get the performance you need at gigabit rates.

I'm not saying that we'll never need 10gigabit to the desktop, just not until we solve the hard drive bottleneck. Solid state storage could solve the problem, but we'd need to have solid state drives that store 100gb of data in order to match the throughput of the network.

Re:Way overkill

[dbarclay10]

Most of your argument rests on people not being able to read/write data from hard drives fast enough to use the network bandwidth. Some examples:

The only time I've ever seen near gigabit traffic at a steady pace was at network servers, where traffic can reach a steady 600mbps on a single gig link - which is maxing out the speed at which the server drive can read/write data to its hard drive. Think of it this way, a 1 gigaBIT link can transfer a 1 gigaBYTE file in about 10 seconds, that's FAST! Conversely, it takes nearly 20-30 seconds just to write that large a file to the hard drive.

More:

Even at these tremendous speeds, they are only used at traffic aggregation points, again because any network device, even a turbocharged SAN couldn't handle reading/writing at those speeds for anything longer than a quick burst.

And lastly, your conclusion:

I say this: If you think that 10gig/sec is your answer, you're looking at the wrong problem. You can get the performance you need at gigabit rates.

Given your premise, you argue for your conclusion quite well. I don't, however, think your premise is accurate. Or perhaps better, I don't think it's relevant. First and foremost, there's all sorts of storage mechanisms which can transfer data as fast or faster than 10Gbps. Think solid-state drives and some decent-sized drive arrays (they don't need to be *that* large, we're talking roughly 1 gigabyte per second; that can be done with 5-10 consumer-grade drives, let alone the arrays of hundreds of high-end 15kRPM SCSI drives and the like). So on the basis of storage speed alone, your argument fails.

Second, what does storage speed have anything to do with it? You mention servers not needing this - a *huge* number of servers never touch their drives to read the data they're serving. Drive access == death in most Internet services, and people invest thousands of dollars in huge RAM pools to cache all the data (they used to invest tens of thousands, but now RAM is cheap :). So for a huge number of servers, drive speed is simply irrelevant; it's all served from RAM and generated by the CPU, so unless you're trying to say that CPUs can't deal with 10Gbps (which you aren't, and quite rightly), the conclusion falls down again.

Do desktops need this? No, of course not. If that's what you're really trying to say, then all fine and dandy, just say it. Acceptable reasons would be "people don't need to be able to transfer their 640MB files in less than 10 seconds" and "their Internet connections aren't even at 10Mbps yet, they certainly don't need 10Gbps!" However, you'll find that this technology quickly percolates downwards, so at some point in the future people will be able to transfer their 4GB (not 640MB at this point) files in a few seconds, and their "little" 640MB files will transfer near-instantaneously.

Re:Way overkill

[Acidangl]

WS-X6704-10GE Cat6500 4-port 10 Gigabit Ethernet Module (req. XENPAKs) $20,000

Re:Way overkill

[toastyman]

Nah, it's not overkill.

We've got web servers that routinely push >700mbps of traffic. We keep them load balanced to never go above 800mbps, but in the "lab" I've had max output as high as 950mbps.

Granted, with current CPU speeds, we're getting near 100% cpu utilization at those rates anyway. But, I haven't tried this on any of our new dual 3.2Ghz systems, which I'm guessing could handle full wire speed without choking the CPU at the same time.

"But wait," I hear you saying, "drives can't keep up with that!" Well, no, one drive can't. But we have RAID, and a very high cache hit rate. When a new video is posted to the front page(not safe for work at all), the same handful of videos account for 99% of the bandwidth from our video download servers. If we've balanced things correctly, the disks hardly get touched. The only limitation on download speed is the most congested link between you and us.

I'm not saying right now that we'd use 10Gb, but I can see it coming in handy soon. The day we start having to buy extra servers just because we're hitting the limit of gigabit ethernet is the day we'd probably invest in 10Gb gear.

Re:Way overkill

[gsasha]

Well, there is a use for this - shared memory.
What's the bandwidth of the memory subsystem of your computer? Chances are, well over 10 Gbps (giga-bits).

Re:Way overkill

[JRHelgeson]

Quite right, thanks for the reply.

I think it's fairly obvious by now that my experience lies primarily in the corporate environment with database servers and the like.

I do have experience in internet convergence points, but not as much with ISP's serving up video files, or rather the same video again and again. When I think of data transfers, I think of hauling bits from server to workstations, or servers to servers where sustained transfer rates would kill a server - much as you stated; drive access == death.

On servers that can handle 1 gig throughput, as you stated, the CPU is at or near 100%, ad a second CPU, plug in another gigabit uplink and team them. Even still, disk access==death. Its kinda like the indy 500: at 200+ MPH, touch wall==race over.

I'm just saying that even at today's CPU speeds, with huge chunks of ram, trying to handle speeds like this at the server level creates serious scalability problems.

Re:Way overkill

[mnmn]

Most of IP traffic is very bursty, and when they do burst, the users want the minimum latency. An IDE interface cant handle anything to saturate a 100Gbps, (well maybe IDE interface can, but the disks cant).

Note also that the average memory will soon be 512MB, so if 3 other users are connected to this machine, plus the user of this machine is playing counterstrike with voice/video chat, the requirements cant be handled with 100mbps unless the priorities of the various types of traffic are carefully managed.

We dont have to use 10Gbps constantly to justify its necessity.

Re:Way overkill

[Anne Thwacks]

Yeah, well just wait till you have to carry country to country VoIP carrier traffic - then you will see data flowing!

Re:Way overkill

[JRHelgeson]

You're supporting my point, where no single server can handle 10Gbps, you only see that type of traffic at network aggregation points.

See, routers only look at layers 3&4 information when making a routing decision. The computing end-point, be it a server or a workstation needs to analyze each of the seven OSI layers once it receives the packet. That is far more work than a router has to do, work that on a computer is done by the CPU, controlled by the operating system.

If you want to offload the processing power to a high end network card in an effort to attain 10gb on the server side, then that network card would have to have its own CPU, much like the high-end graphics cards of today - except that graphics cards are much, much simpler. Video cards process exactly what they receive from the machine, then puke it out to the monitor. They do not necessarily send data to the CPU for processing, unless you're doing video capture.

The network card is bi-directional. That CPU on the network card must be able to communicate with the OS as it is fetching and processing information independant of the Core CPU. On top of all this, the PCI bus speeds cannnot even support 10gpbs...

I'm not saying its not gonna happen, I'm just saying that even with todays network connection speeds 99% of people are rolling a marble down a sewer pipe as it is. People that think the answer is 10gbps to the desktop is the answer don't know what the hell they're talking about.

Re:Way overkill

[A_Non_Moose]

While I will profess my knowledge of fiber doesn't go much beyond "more better, more faster", you stated yourself that it depends on your "average user".

Plain users on wordprocessors, meh, they'll never be able to tell.

CAD/CAM: Helps, they might notice.

GIS: Oh, god yes, they can and do notice immediatly.

On topic tangent: Why ram isn't being used on devices is beyond me.
Think about it when it comes to video cards, thier memory, GPUs and AGP speeds. The limiting factor is usually memory amount and speed (oversimp'd, I know).
Why some of these fiber cards don't have DDR/SDR slots like some of the IDE RAID capable cards to help speed up when multiple copies are needed.
Even pulling off a 16 channel SATA RAID5 seems to fall over after about the 3rd copy.

Back to my point:
My predacessor for some reason never used the fiber link on my lab's powervault, just the 10/100 link. So I plugged in the fiber, did a few trials to compare (there is lots of 1.2G+ files available for me to choose) and when I switched over someone logged in 5 minutes later and said "Damn, what'd you do this is fast!".

It has been steps up since then, because my network (inherited) was a mish-mashed mess of:
Gb Fiber into the switch, 10/100 out.
Cat5 from the switch to patch panel and from the wall to the desktops.
Cat5e in the ceilings.
Powervault 740 would top out at 7MB/s over fiber.
Snapserver topped out at 3MB/s (ugh) (10/100)
Eventually got a pogolinux box ~10MB/s (Gb copper on 100Mbit {sigh})

Added a powerconnect switch (5224?) and while not the best, added an undeniable boost, and a strange loss:
Cat6 on the Gb copper cabable machine's patch panel and wall connection (cat5e still in place).
Added a fiber link from stacked powervault's switch to powerconnect switch.
Gb fiber on pogo box to 5224(?)
Snapservers = same (no suprise there)
740 dropped to 5MB/s
Pogolinux box = 38MB/s average, 48MB/s sustained.
(samba 3.02, 32Kb buffers seems to be the sweet spot with 2k/xp desktops)

Overkill?

Depends on you "average users" and infrastructure doesn't it?

And there's always going to be a bottleneck usually in the order of: user, desktop, sysadmin (naaah), network, server.

I agree with you, 10Gb is a solution, to a problem I don't really have, yet. The problem in need of a solution is SMRF (Storage, More, Redundant, Faster).

Sigh.

I think bandwidth might change everything

[Omega1045]

Even if we do run into a situation where the network is faster than the server I think we will then see the true power of P2P, distributed computing, server farms and the like.

Imagine a corporate environment where every machine is the file server, storing some pieces to the puzzle? We don't have to rely on backups as often since data is redundantly stored all over the network. Huge servers (big proc, big disk, big ram) may not be needed as much since the big pipe and a lot of small workstation servers are taking the load. I mean really, most users do not use the power of the new desktops coming out.

Even processing tasks can be distributed out to other machines once the network is fast enough. Just in my own group our project takes a minute or so to compile clean. I imagine being able to spread that task across the desktops of the other 5 developers? Or across the desktops of the other 500 employees of the company?

If are law makers don't fark things up, this "future" thing people talk about could be really cool!

Re:I think bandwidth might change everything

[Omega1045]

I know the trolls are going to point out the wrong "are" instead of "our". That will teach me to hit the "Submit" instead of preview on a Saturday night when I have had a few beers....

Perfect for...

[CoolMoDee]

getting rid of that annoying coworker...well atleast temporarly. ping -l 102400000000 [insert annoying coworker's ip here]

Argh! No more!

[namespan]

Dude, what *is* it with the porn meme? There's like 15 comments to that effect already in this thread, before it broke 50 comments. If every slashdotter was as dedicated/addicted to downloading pr0n as is stereotype goes, the whole freakin' internet would have been DDOSd long ago.

Yeah, I know it's popular, but geez. Not all of us are spending our time gazing and wanking. Some of us actually code (and even talk to women!)

I hereby banish this to the Beowulf cluster of memes, along with Soviet Russia/Hot Grits/Profit!

Re:Argh! No more!

[Commander Trollco]

<voice "nerd">And some of us code women! Girl robot, here I come (cum?)!</voice>

And don't bash hot grits. Hot grits was cool.

Re:Argh! No more!

[Anonymous Coward]

I hereby banish this to the Beowulf cluster of memes

Funny, the "Wow! Imagine a Beowulf cluster of these!" was my initial response.

Re:Argh! No more!

[optikSmoke]

I hereby banish this to the Beowulf cluster of memes, along with Soviet Russia/Hot Grits/Profit!

Umm, ya. Well done. The, um, banishing of things into..... popularity. That'll be effective. We all know how unused each of those oft-repeating jokes are. Oh, wait......

damn.

All your base are belong to porn?

Re:Argh! No more!

[TubeSteak]

Ha! You're a funny guy.
Lemme give you some advice, from one /.'er to another:
Don't trust anyone who downloads porn but won't admit it.

If we don't get massive bandwidth sooner rather than later,
half the internet traffic will be spam
and the other half will be porn

Re:Argh! No more!

[Anonymous Coward]

People are interested in the business and technology-driving aspects of porn.

Porn continues to be one of the leading drivers of technology (war being the other one) having "made" the VCR, VideoCD, color-printing, Video Streaming, and many other industries.

Porn also continues to be a serious business, with the New York Times (may 18 cover story) claiming Pornography has $10 - $14 billion in annual sales - bigger than any major sports league.

The porn industry employes 12,000 people in California alone, contributing $36 million in taxes to the state. Comcast makes $50 million per year on adult cable programming. DirecTV is estimated (by CBS) to make $500 million/year on adult content.

Recent estimates of Internet porn are that it's about a 1 billion dollar industry. CBS reports that "Consumer demand is so strong that it has seduced some of America's biggest brand names, and companies like General Motors, Marriott and Time Warner are now making millions selling erotica to America. Correspondent Steve Kroft reports.".

The CBS article claims that Hilton, Marriot, Hyatt, Sheraton and Holiday Inn all offer pay-per-view porn and such programming is purchased by 50 percent(!!!) of guests accounting for 70% of in-room-profits.

Bottom line is merely that we respect porn as an important technology driver, and serious business that is very important to both broadband communication (this article) and other technologies including digital video, alternative payment mechanisms (phone-sex is the biggest driver for phone-calling-cards), etc.

Perhaps one day when you reach puberty you'll understand too.

Re:Argh! No more!

[killjoe]

I don't see what you are so upset about. I for one welcome our porn overlords.

Re:Argh! No more!

[o0zi]

The phrase, I believe, is:

PWNed! :)

Re:Argh! No more!

[UserGoogol]

Yeah, that meme is banished... IN JAPAN.

Re:Argh! No more!

[mog007]

Actully porn is impossible to DDOS on the wide scale. Every person has certain tastes when it comes to porn, and the internet is no less than 90% porn, anyway. By the time IPv6 becomes vital to continued internet usage, then there might be enough people around to DDOS the porn sites that're floating around today.

Now then... where are my "educational" bookmarks...

Is that really so much ?

[Tobias Luetke]

10 GBps doesn't strike me as all that much. For example its still puny compared to the speed the data is beeing passed around on the main board.

At the current rate I would predict that our 100mbit switches/routers will be the bottle necks of our internet connections within the next 2-3 years. So an upgrade to 1 gbps switching gear / nics is forseeable but compared to the jump from isdn to broadband, IPv4 to IPv6 or 32 bit address space to 64 bit address space its rather a step forward then a new era.

Re:Is that really so much ?

[gl4ss]

**At the current rate I would predict that our 100mbit switches/routers will be the bottle necks of our internet connections within the next 2-3 years.**

being a bit overconfident about the building of the backbone networks to actually provide such connections to people(if you're talking about cheap consumer connections)? maybe you're thinking that starting to host slashdot size sites is going to fall to 20$ per month too?

oh and 10gbps is much compared even to the data passed around the motherboard, that's why it's unlikely for _one_ current home computer at a time to saturate 10gbps link.

what would i do with a large 10gbps network? start selling it..

oooh

[iamdrscience]

IDE over IP. Yes, it does exist.

Telecommunications

[nasteric]

Better VOIP would be nice. Maybe even for the home depending on upload rates.

anecedote

[qqqqarl]

'But just how much data can a person consume?'

10 years ago, our school was trying to decide whether or not to upgrade its bank of 2400 baud modems. the guy leading the student advisory board replied "no - spend it somewhere else. just how fast can one person read?"

i really hate people like that.

K.

10Gbps Wow!

[pauldy]

I would spend the first few hours probably doing nothing more than taking cd's and dvds converting them into iso's and trying to clog up the network throuwing them around. After amazing myself with the speed and finding myself content I would probably have to find a few of the fastest machines ont he network and ping flood the slowest one i could find then run over to it and see if it effects how fast notepad opens. Then I would setup a vnc conection to see if 10Gbps would allow me to use it form machine to machine without lag. After all of theat I would get a couple of my friends together for a nice lan game. This process would be repeated for the first week or so until the newness wore off. Then I would take the time to learn first hand if the increase in wire speed actualy effected the performance of most machines seeing as how it is difficulat for some memory to keep up with those speeds the only obvious application for such bandwidth is for backbone services.

Re:10Gbps Wow!

[name773]

the only obvious application for such bandwidth is for backbone services.

but the things you mentioned would still be oh so much phun :)

good post

Re:Porn!

[rokzy]

It had to be said.

It was.

Ah, but that is because

[empaler]

your harddrive is too slow ;p

How about a 270GB 3-D image dataset?

[mixwhit]

We'll soon be getting an x-ray microtomography instrument. This 3-D CAT scanner with a 4k x 4k CCD produces 3-D image datasets of small objects down to a 5 um resolution. A full dataset can run 270GB! We'll be purchasing a RAID and a dedicated tape library to handle the data.

However, we'll also be doing lots of analysis of that data on multiple machines throughout the lab...and you have to move that data somehow!

backbone.

[transiit]

10GigE is what I'd use as a backbone between buildings, metro area networks, etc.

Come in real handy if GigE rollouts to the desktop start happening.

And before anyone starts spouting off about maximum 100m spans, I'm talking 10GigE over fiber

-transiit

The Copenhagen Metro

[empaler]

Had some pretty slick security cams installed in them from the beginning (~3-4 years ago) - but they couldn't use them. Why? Not enough bandwidth to send the images uncompressed. Which was what they had set them up to do. Solution? Turn off cameras. Wait a few years for more funding.

Re:What would I do with this much bandwidth?-Music

[Lord Prox]

Ya know, so far everyone seems to think of this as a long distance pipe. It's not, it ethernet. RTFA useful distance is in meters *NOT* kilometers. This is an intraoffice connection not a WAN pipe.

Patches.

[empaler]

Several hundred megabyte patches.

Oh.

kick your ass at Counter-Strike

[_Shorty-dammit]

cuz, like, my ping would rule

Frag

[zerogeewhiz]

Frag lots and frag often.

A simple answer

[kko]

How much data can a person consume?

As much data as can be provided.

Remember the 640Kb barrier? As soon as it was broken, we found ways to use the larger amounts of memory available to us. Same thing happens to bandwidth.

Remember people used think twice before downloading a small sound clip on a 28.8 modem, but think nothing of torrenting a movie on a 768 Kbps DSL link.

... riiiiiiiight.

[empaler]

Its so sexy, but all they will let me do is run the fiber for it.

I hope they don't let your near any of the equipment unsupervised...

Why ask...?

[verbatim_verbose]

I'm sure everyone wanted to know what we would do with 10Mbps too...

What can I say...

[empaler]

Screens and cams in every room, long business trip, miss S.O. at home... Video calls...

Re:What would I do with this much bandwidth?-Music

[Chmcginn]

Didja ever notice how long-distance transfer rate is a few years behind short-distance transfer rate... and it is pretty consistent?

(In other words... true, 10Gb per second isn't available from New York to Hong Kong today... but in 2014, that'll be standard... if not so-three-years-ago.)

PCI Express....

[Desmoden]

While I can't say where, I have seen 10Gbps ethernet running at full line rate on a 16lane PCI Express card.

All I can say is WOW. It was quite amazing. However, we are a long way off from seeing normal usage.

10Gbps is really only needed for areas where data merges. For example, would you rathat have 10 interfances bounded together? Or just have one? 10Gbps will take off from a ease of management and port desity.

So for the most part we are just talking about telcom, banks, government, and brocast groups.

It will be a long time before you will see/need this in your PC or laptop.

On the telcome side, I can tell you that I know of several companies working very hard to start shipping these products in volume due to pressure from the telcoms and their customers.

There is demand for this, and it's growing.

Easy

[SteamyMobile]

Clearly, this isn't something that you plug into your desktop computer so that Slashdot loads faster. But, if that's all you think of using the net for, here are some other ideas:

• It allows distributed computing. 10gb/sec bandwidth is higher than the bandwidth between the CPU and memory in most PCs. With "Ethernet" like that, you could have a group of PCs all sharing a large chunk of memory, or extra CPUs, or other resources. Of course, the bandwidth is only one part of the infrastructure needed to make that happen, but it's part of it.
• Continuing the topic above, it would be great for parallel/cluster computing. Some problems parallelize easily and don't require lots of interaction among nodes. Looking for hash collisions, for example, needs only low bandwidth among nodes. But other parallel tasks, like fluid modeling, do require high bandwidth. 10gbps "Ethernet" would make that easier, and if it's a commodity networking system, it will be the cheapest way to build super computers.
• Of course, it's great for backbones. Why bother with Sonet when you can use "commodity" "Ethernet" instead?

Applications involving shared computing resources will definitely require plenty of other infrastructure, including (especially) software infrastructure. I'm using the word "Ethernet" in quotes above because this protocol doesn't really have much in common with Ethernet other than the name and the fact that it uses frames.

What would i do?

[Ynazar1]

Can you say targeted DDoS attack?
Slashdot the internet?
Download more porn and music, slashdot RIAA website and then DDoS their servers so they will never find out?

My currest boss used to be one of the chief IT people at Hostway (webhosting company) according to him he was in an optimal position, and had enough power and bandwithd, to DDoS anything he wanted to (provided it was connected to the net of course). I believe 10 GBit was about the size of Hostway backbone at the time. According to him he also worked at University of Illinois at Urbana-Champaine and wrote parts of the original Mosaic (web browser), but that's just selfish bragging.

Grab the net?

[Maljin Jolt]

wget -m http://www.google.com

That should keep such line very busy for some time...

Re:In short-Amish computing.

[kasperd]

Screaming video cards.
I think you should replace the fan.

Invent ways of consuming it.

[ozmanjusri]

Things like bandwidth are enabling technologies. In themselves they do little, but just by existing, they allow a all of the inventive minds out there the freedom to create and to progress.

To Get a Better Human Interface

[suchire]

Ideally, you'd need even more than 10 Gbps for the perfect user interface.

windows updates

[ehkz]

Finally would be able to keep up with all the windows security fixes.

Dumb terminals? Cluster computing?

[Entropius]

Let's see. There are about a million pixels on my screen (1280 x 800). Assume 24 bit color, so that's 24 megabits per frame.

This at 60 fps will be 1.44 Gbps.

So 10-Gbps ethernet is enough to stream the output of a monitor, *uncompressed*, at full framerate, to either a dumb terminal or another computer. Even the most elementary compression (only reporting changed pixels, or PNG/jpeg techniques) could cut this to a fraction of 1.44Gbps.

More generally, it could allow more of the things that are currently on the PCI/USB bus to become external, and could become a more flexible replacement for USB. Scanners, cd writers, audio devices, you name it ... lots of things could be externalized and generalized. This would also allow more devices to be shared across networks more easily, since they're *on* the network in any case. With the Internet, nobody cares about the physical location of the machines they access; likewise, with this system peripherals aren't associated as strongly with one specific computer.

This sort of thing might also have applications for cluster computing, allowing more sorts of things to be done with clusters since you have higher inter-node bandwidth.

Re:Dumb terminals? Cluster computing?

[Jeff DeMaagd]

Clusters more often need high bandwidth, low latency. I don't know if 10Gb fixes the latency concerns of gigabit.

The typical trend is that the fastest links are sparingly used as interconnects in the server rooms, and those links gradually filter out to the rest of the servers, then to the desktop as the need arises. It was that way with ethernet, 10Mb, 100Mb, gigabit and eventually with 10Gb.

How much can someone consume? Wrong question..

[sllim]

How about how much information can someone serve?

Off the very top of my head I can think of a couple of ways for me to digest huge amounts of data over the internet. For one, how about a noncompressed HDTV stream?
What about video games that don't require a hard drive (and are then more secure)?
Hell, how about loosing the hard drive altogether and just having a dumb terminal?

Nah, asking how much data can one person consume is a lot like saying that building a hard drive over 20 gigs is stupid cause it will never be filled.

The real question is how much data can one person (or a company) serve?

Back to the idea of uncompressed HDTV content. If HBO decided to put such a thing on the web for subscribers how many unique subscribers could it possibly hope to serve at a time?
Seems to me that asking a server to supply many streeams of these uber bandwith apps at once would be problematic.

Actually,

[empaler]

my biggest chunk of porn has been copied from my flatmate ^_^

[OT] America West Reply

[weston]

I understand that they're struggling somewhat and they've taken bailout loan guarantees from the feds and there was talk of Chapter 11 filing. And yes, from their sec filings you can see they're putting greater pressure on their fleet/staff. But I figure they are the second largest low-fare carrier, next to Southwest, which is really the only airline kicking butt and taking names at this point, and they don't appear to be hiring for the kind of job I want (pricing/yield/revenue management), while America West does. And also, lean times at a company can present a real opportunity to distinguish yourself by merit.

But feel free to email me further if you've got more inside scoop. If I'd be walking into not just a challenge but a really adverse environment, I'd definitely be interested to hear about it (I would have emailed you already, but you don't have an address listed, so...)

Simple answer(s)

[TubeSteak]

• Things we'd do with massive bandwidth:
• • Send more spam
  • Download dvd's of porn
  • Video/Tele-conferencing with family
  • VOIP
  • More Porn
  • Uncompressed CD's ---} Cd Burner
  • More crummy swf & flash sites
  • Multimedia rich advertisements
  • Did I mention the large quantities of porn?
  • Realtime distributed processing of huge files
  • Actually back up all your data
  • Wonder why HD's are so small
  • Pirate any/all of the above
  • Stream the TV shows we want w/o commercials from China (see above)

  I'm sure I missed a few things, but I hope i hit all the important ones.

The same old question

[Neckel]

I remember times sitting behind a PC and wondering how the hell I am going to fill up the 20Mb of my smart card harddisk just plugged into my XT. Hell, I got them filled up eventually ;-) - so don't you worry.

There is already a thriving market for this

[j. andrew rogers]

Requests for multi-gigabit circuits across Ethernet fabrics are already moderately common, particularly as more longhaul and backbone infrastructure moves to pure Ethernet. You can do this today by channel-bonding lambdas on a fiber with Gigabit ethernet (the silicon will support it), but 10G would make it that much easier.

The most common uses for this: long-haul video transfer, massive network file systems, and extremely high-volume internet servers. Massive distributed private file systems and applications are hot, and will be a major driver of high-performance fiber ethernet fabrics.

For most purposes, a properly architected network will add almost no real latency beyond what is required due to physical limits i.e. the speed of light. Regional ethernet fabrics should be easily sub-millisecond between any two points in the region, which is quite adequate for most applications that might be amenable to being networked to begin with. Most people still think of networks in terms of routing bottlenecks and other topology biasing issues like that, but large-scale ethernet fabrics, an increasingly popular choice for network providers big and small, make most of these issues go away for a lot of purposes and mitigate the performance hit for most others. 10G will just make ubiquitous ethernet fabrics that much more scalable.

Interactive porn? No?

[quarkscat]

I'm still waiting for decent DSL, since I'm
on 30 year old buried POTS wiring that's 5
(plus) miles away. Fiber to terminal point
will not happen here before hell freezes over,
since the Baby Bells are not spending that
kind of money.

However, with that kind of bandwidth to the
internet, I could set up some homebrew web
sites, and telecommute to work, and go back
to (online) school all at the same time.

I hate to be repetitious, but that kind of
infrastructure would allow some really great
collaborative (beowolf?) computing.

Re:Interactive porn? No?

[mattyrobinson69]

i agree with the clustering idea, not beowolf though, something like openmosix, so all applications are transparently clustered.

Stop asking

[Lord_Dweomer]

You know, whenever a story about bandwidth increases comes out, and there is the inevitable question of "but what will people do with it" I always find myself answering:

"Don't worry about it, just provide us with the bandwidth and we'll figure out a way to use it."

Seriously, there's really no telling WHAT will take off until people get their hands on it, start tinkering, and start doing things.

For starters, how about upping the quality of the media we transfer? Storage space is increasing and becoming cheaper, coupled with this there really will be no reason to keep cruddy 128kbps mp3s and low rez divx files around. People will start switching to lossless quality formats once the storage and bandwidth increase to usable amounts and lower respectively in price.

So, aside from all the potential new uses that people will invent for this bandwidth, you can ALWAYS assume that people will just take what they currently do with their current bandwidth, and just upgrade it in size till it works well with the improved bandwidth. Its practically a given.

Well...

[sw155kn1f3]

I'll be building DRBD clusters in a blink of an eye.
Actually I already do on 1gbps :)
Redundancy is good.

10Gbps? How about 10Mbps?

[AnotherBlackHat]

Video downloading is possible at 1 Mbps--although it takes longer to download than to view or consume--and more than feasible at wired 100 Mbps.

4.7 Gig for a 2 hour DVD is under 6Mbps.
The average consumer probably won't buy more than 10Mbps.
Sure, we'll all want 10Gbps, but not many would be willing to pay extra for it (unless someone comes up with something even more bandwidth intensive than video).

A publisher might need more overall, but they can probably get by just fine with 100Mbps and a contract that requires subscribers to run something like bit torrent.

-- less is better.

New Protocols

[Percy_Blakeney]

Everyone seems to be saying either porn or movies (or a combination thereof), but I wonder if large amounts of bandwidth will enable more "expensive" protocols that allow for new ways of communication. Look at P2P -- it uses scads of bandwidth, but allows you to mostly take central servers out of the picture, which in turn enables a more decentralized and democratic way of communicating. Or look at things like XML, which don't try to squeeze every last byte of space out of the data but instead focus on being human-readable. These kinds of things are the result of certain resources (such as bandwidth and CPU power) becoming more and more of a commodity.

Perhaps it is like water. A person that grew up in the desert where water was scarce and critical to life could hardly imagine the fundamental changes that an abundance of water would cause. If asked what their life would be like with more water, they would probably think, "Wow, I could drink a billion gallons a day if I wanted to," when the real value comes instead from things like sanitation and the use of water in the creation of other goods.

So, while I agree that we'll see a lot of high-bandwidth media files use the new bandwidth, I also believe that we'll start rethinking the fundamental bandwidth scarcity assumptions that we've made and begin to design new protocols and data formats accordingly. There could be cascading results from it, as well, reaching into politics, education, and economics. I doubt the Earth will suddenly become a paradise, but it has the potential to improve things on a lot of different levels.

Decoupled disk storage, dumb terminals, HDTV

[stuartkahler]

Instead of your site having 1-2 underutilized hard drives in every computer waiting to break down, you can consolidate everything into a single server. Less noise, heat and power usage. Much easier to deploy extra storage capacity for multiple users and back it all up.

You can set up dumb terminals capable of running any game that the server can handle by streaming the video and sound over the lan. Who cares how noisy the PC runs when it's down in the basement running your game of Far Cry that you're playing upstairs. Probably on par with watercooling your CPU, GPU and mobo chipset.

I would really like it if I could pick up a few diskless laptops with USB, a 15 inch screen, 64 MB ram and 400mhz(?) processor for $200 that would connect to a master desktop server. Drop them around the house to use as picture frames, TVs, music players, game machines, or web access. The only limit is the power of the master server that they all connect to.

Distribute HDTV video around the house.

Re:Decoupled disk storage, dumb terminals, HDTV

[stuartkahler]

Heat generated by a typical desktop:
CPU: 60-110 satts
Memory: 5 watts/256MB
each HD: 10-20 watts
MOBO chipsets: 30+ watts
graphics: 20 watts (40-70 watts for 3d gaming)
audio: 5-10 watts
usb: up to 5 watts per port being used
+ about 50% for PSU inefficiency

Single point of failure means that there is only one machine to back up. Each terminal is identical, so it's easy to determine if the terminal is faulty by checking another terminal. Otherwise it's the server's problem. To fix a problem, you either debug the server from your own office, or go drop off a replacement terminal. Once you have a terminal that is capable of handling a screen at 1600x1200, you never need to upgrade that box. Just transfer the existing server to a new machine.

There are already free software solutions for this type of thing in small environments(VNC for example). More would quickly come out as the technology matures. Part of the reason that current thin clients are expensive is because they currently have to run applications locally. I'm talking more about dumb terminals, which are more like a networked data interface. Thin clients are also much smaller than standard desktops(slightly larger than my cable router), making their cost higher. The fact that they are produced in small quantities doesn't help any either. I would bet that if they were being produced at more than 10million per year, the cost would easily drop below $100 each (with mouse and keyboard), plus whatever the display costs. I got my 802.11b cable router (which runs linux and a web server) for $30. A dumb terminal isn't that much more complex.

As computers become more integrated into our daily lives, people are wanting more convienent access to their machines, rather than more of them. I may want to use my computer at work, in the office, den, kitchen, referigerator, bedroom, garage and back patio without having to drag the entire machine with me. Laptops don't really suit my needs very well because I like to upgrade frequently. With desktops, I have to install software on every machine, and they take up too much space.
Also bear in mind, that most people only use 5-20% of their computer's cpu power most of the time with spikes during things like loading a new application, opening a file, or 3d gaming. You can limit any single application to 70% of the cpu's time, and still have a very responsive experience for a few other users doing things like web surfing, word processing and listening to music. Dumb clients could be a very good alternative to individual PCs in a family with a few kids that are sick of having to wait their turn to use the computer.

More deterministic ethernet

[cosyne]

Ethernet isn't really deterministic, but it gets more predictable as the traffic drops. So with a really fat pipe, you can send your pr0n or whatever and be reasonably confident they'll be finished by the time something else needs to be transmitted. Which simplifies things like machine control.

I think we all know the answer already.

[j0se_p0inter0]

Because this has been asked before. Just like "Dude, Gigahertz? Why would we even need that?", and "50 gig hard drive? What the hell would you fill that up with?". The answer is: We'll know what to do with it when we get there.

What do you mean...

[Limax Maximus]

...We've already got 10Gigabits. What we do with it? I've not seen it go above 15 Megabits yet and we've had it in for a year.

Re:What do you mean...

[NerveGas]

I was in a building just a couple of hours ago with two OC192s. That totals up to 20 gigabits per second.

The two OC192s easily handle all of the company's voice and data communication in the city, which is not at all a small amount.

Very few people realize just how much data 10 gigabits per second really is, and fewer still realize how much hardware it takes to be able to send/receive/route/process data at that rate without even looking at the network card.

steve

+1 Insightful.

[weston]

That's an excellent point -- no point in opening the throttle outside of the motherboard if there's still a pipe into the machine that can't handle the flood.

Re:+1 Insightful.

[mattyrobinson69]

what about pci-x?

also, if a few computers were using it at the same time, it would make up for it.

Ohh I know I know!

[Gldm]

How about cluster the video output buffers of all my machines in like a giant beowulf SLI so I can run doom3 faster than slideshow?

acronym

[itzdandy]

ONE ACRONYM PEOPLE "SAN"

storage area networks. at 10Gb/s, SAN's are not a slow data storage medium, but an ultra fast storage medium, that may ever be faster that your local hard drive.

OR, how about full diskless clients, with BIOS manufactures being able to map specific locations on the network to a local harddrive via something like NBD(network block device) or some next-gen NFS. at 10Gb/s, your datatransfer is 1.164GBytes/s, minus overhead, lets say %25 to be safe, or just short of a GIGABYTE/s. how fast is your local hard drive? with a SAN, your could run a very fast RAID5 setup, with a rediculously large cache and have nearly instantanios access to 1,000s of large files. transfer a full DVD in 5-10 seconds, boot your machines faster because of increased datarates.

-

CLUSTERING software. imagin a beow...., how about openmosix!, with 30 desktops running linux, all with an openmosix kernel and a very high speed interconnect, each user would feel like they are on a supercomputer when rendering digital media, encoding video or audio, compiling programs etc. You would basically be using CPU's like ISPs use bandwidth, each user feels like they have a faster connection even thought they are being oversold bandwidth, idle users give up spare cycles for greedy users to prosper.

-

also, ethernet attatched expansion devices. Need a new sound card? need one to run the loby music system and want your server to handle it? Ethernet Attatched Expansion Port(EAEP), plug the device in, connect to it accross the network, and hit play on xmms!

EAEP scanners and video cameras = awesome. 1 640x480 camera compressing plain mpeg1 video can saturate a 100Mbps network pretty quickly, 3 or 4 can bog down a 1000Mbps, with 10Gbps, you can get 25-30 without boging down your network.

-

HDTV delivery, speeds like this would allow High Definition media delivery without bandwidth pains

Re:acronym

[NerveGas]

storage area networks. at 10Gb/s, SAN's are not a slow data storage medium, but an ultra fast storage medium, that may ever be faster that your local hard drive.

You still think you're going to get 10 gigabits to/from that thing? Think again. That's over a gigabyte per second. It would take at least twenty disks in RAID 0 just to barely get up to that speed. However, putting twenty disks in a RAID 0 array means that your realistic MTBKYBG (Mean Time Between Kissing Your Butt Goodbye) would be something like a month.

Want 1+0? Well, then you better have at least FORTY disks in the SAN. Want RAID 5? Well... tough. It's going to be quite a while before you find any RAID 5 solution that's going to write at anything near a gigabyte per second. In fact, even READING at a gigabyte per second is going to be very expensive with RAID 5.

Sure, there are SANs that can dish out over one gigabit. But hitting ten gigabits is going to be tough, and not just because Ethernet has so many limitations.

OR, how about full diskless clients, with BIOS manufactures being able to map specific locations on the network to a local harddrive

Sorry. Virtually no desktop-level computers can even realistically take full advantage of even a one gigabit card at the moment. Slapping a bigger card in the machine isn't going to overcome architectural limitations.

CLUSTERING software. imagin a beow...., how about openmosix!

What? Serious clusters don't even use gigabit ethernet, it's too much of a hacked-together kludge. Bumping it up another order of magnitude will probably make things worse, rather than better.

EAEP scanners and video cameras = awesome. 1 640x480 camera compressing plain mpeg1 video can saturate a 100Mbps network pretty quickly

So, use MPEG2, or an even better CODEC. It'll be cheaper to implement that than to implement something that could handle a 10gigE network.

HDTV delivery, speeds like this would allow High Definition media delivery without bandwidth pains

Amazingly, there's already plenty of bandwidth to deliver HDTV.

Don't get me wrong, there are places in networks that really need 10 gigabits. There are just extremely few, and it's extremely expensive to be able to transmit/receive/process 10 gigabits even with a more advanced network than Ethernet.

steve

Re:acronym

[pe1chl]

>OR, how about full diskless clients, with BIOS manufactures being able to map specific locations on the network to a local harddrive via something like NBD(network block device) or some next-gen NFS. at 10Gb/s, your datatransfer is 1.164GBytes/s, minus overhead, lets say %25 to be safe, or just short of a GIGABYTE/s. how fast is your local hard drive? with a SAN, your could run a very fast RAID5 setup, with a rediculously large cache and have nearly instantanios access to 1,000s of large files. transfer a full DVD in 5-10 seconds, boot your machines faster because of increased datarates.

Todays disks run at about 60 MB/s. That is about 500 Mbit/s.
Moving all your local disks to a central location will mean the LAN is again the bottleneck when you have over 20 workstations. And of course then there would have to be at least 20 disks in the server.

Considering the prices of workstation disks, the prices of server disks, the costs of LAN technology, and the limited advantages (remember you are just moving physical disks, not making things easier to manage or whatever), I don't think this is going to be a good idea.

Full diskless clients have been possible for a long time. But there are limitations, especially to the usefullness. And this remains the same.

Parent not informed unfortunately

[anti-NAT]

Today's "ethernet" doesn't have limitations - it is really only referring to a frame format.

The distance limitations were initially related to running ethernet in half duplex mode, due to the requirement for all devices to be able to detect a collision.

Now that ethernet is run in full duplex the distance limitations due to collision detection have gone.

Distance limitations in "ethernet" are now related to physical media the ethernet frame format is carried over at the specified clock rate. In most cases, cost is providing a constraint, in the sense that longer distances can sometimes be achieved over the same media, however the costs to do so rise dramatically, such that the technology might be priced out of the market it is intended for.

For example, from memory, Cisco have been selling a variant of 1Gbps Ethernet for at least four years now called "1000BaseZX". It would reach around 90 000 metres over single mode fiber. From memory though, the GBICs (Gigabit Interface Converters) were $12 000 US each or something like that, and you needed one per end of the link. And that would be really, really, really cheap when compared with the cost of the 90 000 metres of single mode fibre.

I don't know if the article mentions any distances for 10Gbps, at the moment it has been slashdotted to death.

Is That A Challenge... Knave?

[Shadow of Eternity]

i feel as though this article were challenging my skills...ph34r. Generally i will match my net usage to my conenction, but given a decent broadband connection i will quite effectively Rape the Internet. Flashget makes mass-downloadings so much easier.

Videophones, duh

[Anonymous Coward]

We're still stuck with voice-only phones. One day we'll look back at these times and wonder how in the world we survived without video-phones.

Re:Videophones, duh

[SteamyMobile]

This is pointless. You don't need anything like gigabit for video phone. Gigabit lets you do full-motion uncompressed HDTV signals. Just guessing, 10gigabit would probably let you do digital cinema quality images very easily. 10gigabit has nothing at all to do with videophones. You can easily get a perfectly acceptable MPEG4 video stream into a DSL line. The hard part with videophones is having the right hardware and software to do that compression, plus having compatible software everywhere.

Re:Videophones, duh

[Junta]

Uhh, I doubt it. People rarely want a videophone. They want to answer the phone fresh out of bed, or naked, or something.
Also, those work conference calls where you flip off the unwitting participants are less workable...

10GE core

[swmike]

Ok, let's say we start giving people 10 or 100 meg internet access (this is already happening all across the world). When you start to aggregate thousands of these users, you really need 10Gig at least. How much do these people want to pay for their bandwidth? As little as possible.

Therefore the problem currently is not access speed or even intra-metro speed, it's the core long haul speed that is being crunched, or at least the cost of it. Another poster referred to $120k as being a "mere" cost for OC192 interfaces. This cost is WAY WAY too high (now, the list price of OC192 cards for the Cisco GSR is $225k per card, but that's another matter). Transporting bits long distances right now is cheaper than production cost because of the 2000 bubble buildout still hasn't run out of capacity, but we're getting there now. The cost of bw right now (can be as low as $15-25 per month per megabit of capacity) doesn't pay for the manpower and interfaces to move the bits long distances.

10Gig is not enough for the core, 40gig for SONET is getting close but won't be enough either, soon.

Already at 10gig you're running into all kinds of optical problems such as dispersion and so on, which needs to be handled. At 100Gig it's going to be quite a lot more problematic, if possible at all (you need to go 60-80km without repetition).

100meg internet access is enough for users, 10gig is enough for intra-metro, it's the inter-metro and international bw I'm worried about, and access speed isn't of as much use if you can only get good speeds within the same city as you're in.

Why even ask?!

[zmooc]

What would you do with this much bandwidth?

How about moving all harddisks to somewhere I can't hear them?

Re:Why even ask?!

[owlstead]

Yup, that was my idea too. Make a nice raid-5 disk storage somewhere (s-ata + command tagging would do nicely thank you) and connect the computers too it. Your data is more safe, the irritating hard drive is somewhere else. Do one install for all your computers on it using the old boot from network option and there you go.

Unfortunately my fans are making way more noise than my hard drives together though, so you might want to use special computer for this as well.

Fast Networking and NAS at home

[MikShapi]

I saw some comments here saying Gigabit Ethernet is enough. There's nothing we can do with 10GbE.

I beg to differ. Sit tight.

Here's an idea for you geeks that for some reason nobody is busy doing yet.

Quite a few IT people I know run some form of Linux or BSD server at home, doing a variety of stuff from fileserver to firewall to mail/DNS server etc., though on their desktops they run 2K or XP for reasons such as gaming, simplicity, wife, and so forth.

Here's the idea. Pool all your harddrives at home on the Linux/BSD box, configure a software RAID-5, share it using samba and network-boot all the 2K/XP machines at home from this network-attached storage. Using Gig ethernet of course.

What do you get? Every box gets a system drive "Drive C" that can go at 100MBytes/sec. RAID-5 redundancy for all your machines at home. Harddrives, which generate heat and noise are no longer in your computers.

The benefits are enormous.

There's a small con though - you won't be able to drag your computer to a LAN-party (unless you drag the server too ;-)

Currently there is a shortage of one element though: Software that can boot Win2K/XP using PXE from a fileserver. Such software exists in the commercial world and is made by a french company called Qualystem, which doesn't sell it in less than 1-server+25-client licenses, which costs a whopping 2750Euro. They show zero interest in smaller clients. A second product, Venturcom BXP, does the same but falls short as it has a dedicated server that only runs on 2K/XP/2K3 - no BSD/Linux with SAMBA for you.

If someone in the open-source community were to pick this glove up and write a small driver that emulates a harddisk for 2K/XP on one side (the kind you throw in for a RAID controller by pressing F6 when installing windows), and uses SMB /whatever to access a UNIX fileserver on the other, we'd all be able to rig up a very nifty setup, and use the combined speed of all our harddrives at home.

We'd also realize that Gigabit Ethernet is not enough, as a cheap 4-modern-ATA-drive RAID5 setup (which effectively streams enough data to store on 3 of them, one of the four being used to store parity info at any given moment) writes at 40MByte/Sec x 3 = 120MByte/Sec, and reads at 60MByte/Sec x 3 = 180MByte/Sec.

The Gigabit Ethernet _will_ pose a bottleneck.
If we add more drives, the bandwidth requirement broadens.

There's also the small issue of the PCI bus, your server must have its ethernet off the PCI bus, like in Intel's 875 chipset, nVidia's nForce 250 or on a PCI-Express card. Otherwise the IDE and GB will choke each other on the too-narrow PCI bus.

Anyway. once people start doing this, 1000BaseT is back to where 100BaseTX has been for 5 years - choking. I say - Bring on 10GbE!

Data storage implications

[MarkTina]

I have enough trouble keeping enough storage space on our SAN and NAS boxes as it is, now if things move to 10Gbps I'll never get to sleep for the pager going off to tell me that "volume is 95% full" ... it's far too easy to overwhelm the infrastructure as it is without making it 10 times easier! :-(

Clusters

[noselasd]

For one, fast ethernet cards are nice for clusters, not only because
they can transfer lots of data between the cluster nodes, but just as much
because of latency. Think e.g. MPI programs passing lots of small
messages between processes; the faster the better.

Person?

[Bobke]

What a person consumes is irrelevant, at what speeds a file can be copied is irrelevant. What is relevant is how fast processes can migrate to other nodes in my openmosix cluster.

What would I do with that much bandwidth?

[LordP]

Depending on the cost (and advertising capability), I'd think about running a Slashdot Cache...

Somewhere that you could prefetch a web site before an article gets posted so you could take some or all of the brunt of the Slashdot Effect (tm).

Or perhaps hook it into the Akamai network to distribute the load even further. Give each site a 24 hour TTL, should be enough for most topics.

Huh?

[Anonymous Coward]

I don't understand why everyone is so impressed with that post. Insightful? I'd say thats just common knowledge.

How fast can your processor communicate with your RAM?

How fast can that ram communicate with the HDD?

How fast can your computer connect to the LAN?

How fast can your LAN communicate with the internet?

See what I mean? It's pretty common understanding that the farther you wanna go, the slower things get.

Well, I guess its not that common since you all modded that +5. I am disappointed in all of you. You are not geeks. Go away.

10G not for desktop but for core network

[tijsvd]

10G Ethernet is currently being sold (a lot), but not to connect computers. There are currently three major drivers for 10G Ethernet:

• Connect backbone LAN switches, e.g. two Cisco 6500 machines, each one full of Gigabit links to access/distribution switches.
• High-speed links for ISPs or research networks. 10GE can reach about 70 km without repeaters and is significantly cheaper than OC-192.
• Link between access and core switches. Since more and more offices are (for some reason) switching to Gigabit to the desktop, the access switches need much bandwidth to deal with the (possible) traffic. A good example is the new Cisco 3750 stackable with 16 GE and one 10GE links.

10GE to the desktop is at this time ridiculous, but don't think that means that 10GE technology is not used.

Did you RTFS?

[Anonymous Coward]

I think you are totally missing the point.

OK, transfering between two computers won't beable to consume all that bandwidth.

HOWEVER, when TENS or HUNDREDS of machines are using a network, then you will really see a difference. And I belive that the SUBJECT OF THIS STORY IS AT WORK, WHERE THERE IS PROBABLY MORE THAN TWO MACHINES.

In WORK networks, not your pansy home setup, the LAN or whatever can often be running at capacity. When many machines are busy transfering large loads, you will easily get a bottleneck at the 'network' level. If you increase this 10 fold, you will likely get much more than just a 3x increase in performance, as a result of widening that bottleneck. Especially with MANY nodes using a shared network, you will find these gains to be substantial.

Re:Did you RTFS?

[tomhudson]

In WORK networks, not your pansy home setup, the LAN or whatever can often be running at capacity. When many machines are busy transfering large loads, you will easily get a bottleneck at the 'network' level.

... so replace your antique ehternet hubs with cheapie ethernet switches

I could not believe my eyes earlier this year when I saw a busines that had just had their system redone - a 24-port 10mbps hub. Expensive and f$cking lame.

Re:Did you RTFS?

[tomhudson]

The 10gbs stuff quoted in the story is only suitable for machines phyically VERY close together - like in the same room, connected by cat7.

Not something you're going to see on the local lan. Too expensive to recable.

Even 1gbsp won't work in a lot of existing lans, because of bad cabling, kinks, routing near flourescents, etc.

Ethernet is not Ethernet

[Detritus]

From looking at some of the specifications for 1000Base-T, it may be called Ethernet, it may have an Ethernet compatible software interface, but the hardware implementation looks nothing like traditional 10-Mbit Ethernet. Forget about collisions and shared media. They no longer exist. A 1000Base-T NIC contains four 250 Mbps full-duplex modems, one for each pair of wires in an 8-wire networking cable.

Never enough

[sanermind]

'But just how much data can a person consume?'

...just wait until pr0n goes high-definition!

Yes that's why you do NOT need 10GbE

[Donny Smith]

>It is not a problem at all if most of your clients use the same, relatively small (and thus cacheable) set of files.

Then you don't need 10GbE.

>Besides, the disk I/O is not that frequent these days, once you have your app loaded.

Exactly - more RAM and you don't need 10GbE.

And finally it is cheaper and more reliable to have the server connected to four GbE switches (and 1/4 of the clients connected to each of those switches) than implementing a redundant 10GbE.

First of all...

[illumin8]

Let's see, what would I do with all that bandwidth:

Try to find a host OS with a TCP/IP stack that can properly utilize 1 gigabit ethernet, let alone 10 gigabits. Hint: It ain't Linux...

Try to find a storage solution that can read or write that fast. I'm thinking something like EMC with about 6-8 2 gigabit HBAs using Veritas DMP (dynamic multi-pathing).

Try to get all of the above, along with a 133 mhz. 64-bit PCI-X bus that still can't actually keep up with 10 gigabits of data. (133 mhz. 64-bit PCI-X is only about 1024 megabytes per second, not counting overhead).

The problem is, right now, the rest of the parts of a system just can't keep up with 10 gigabit ethernet. The only box that I would use that can handle that many I/O paths to storage (we're talking six to eight 64-bit 66 mhz. 2 gigabit FC host adapters) is a Sun Fire 6800 or something larger. The problem is, Sun doesn't yet support PCI-X, so now your 10 gig ethernet card is going to be limited to a 66 mhz. 64-bit PCI version, which will only transfer a maximum of 512 MB per second, not counting overhead. That is less than half of the available bandwidth of 10 Gig Ethernet.

You can forget about putting it in any Intel based system. There are not enough I/O busses and I/O controllers in even the beefiest Xeons or Opterons that can handle this much bandwidth (to disk).

Also, if your application doesn't need to write all of that data to disk, then how large is this dataset in memory that needs to be transferred at 10 gigabit speeds? If you had a server with 64 GB of memory, it could transfer it's entire memory set over 10 gigabit ethernet in less than 60 seconds.

A far better, and more economical solution, if you really need 10 gigabits of data throughput to the network, would be to use the same Sun server, and a product called Sun Trunking, which allows you to bond multiple gigabit ethernet interfaces together. You get all of the throughput you want, plus more fault tolerance. I've set it up before, and you can have a continuous ping going, across 4 connections, and pull 3 of those 4 connections and the ping keeps going, without even a dropped packet. It's really fault tolerant, and uses your existing switches, NICs, and hardware, without forcing you to upgrade your entire core switch architecture.

Re:First of all...

[dpokorny]

Hang on there a second -- I think you're being a little too loose with your definitions.

You most certainly can do real-time 10Gbps with an Intel based system.

I'll conceed that it won't be iA32 based. However, any of Intel's high-end iXP28xx series processors can do full-speed 10Gbps packet processing.

Take a look at: http://www.intel.com/design/network/products/npfam ily/ixp2850.htm

holographic video

[Killshot]

I read an article a while back talking about some of the various holographic video technologies that exist.
In some of them, you are required to render the video from all possible viewing angles, creating terrabytes upon terrabytes of data for just a short video clip.
So, I am guessing that Three dimensional video will have no problems using up the bandwidth in the future.

Never underestimate what people can come up with to use available bandwidth, processing, and storage.

i could have told you that category

[yoha]

Currently at work, we're working

nice, it all makes sense now

I'd wait less.

[NotZed]

What for? Shit knows.

10 Gb/s thats it.

[jellomizer]

Remember this is 10 Gigabits per second that is only 1.25 Gigabytes per second (Assuming 100% speed which I never seen happen). Although Right now that is faster then most computers can handle data internally but there are its uses.

1. System to System Backups. If you have seen the prices of memory and Hard drives space $/GB dropping at a fast rate while magnetic tape remains near constant. Soon the price/GB of memory and Hard drives will be lower then tape so it would be cheaper to backup your data on other Systems and Removable hard drive. So if you have a 3 terrabytes of data it can still take up to 40 minutes.
2. Imagine a Beowolf Cluster connected at 10Gbs still right now the main slow point with a Beowolf cluster is the network bandwidth. at 10Gbs you are getting closer to the speeds of a supercomputer bus.
3. Uncompress Video and Sound. No more lossy compression needed with a bunch of people fighting over compression standards. And we always get high quality Audio and Video realtime off the network
4. 3D Now with 3D displays starting to become available there now will be more data needed to send 3D information. Over the network.

And thats just the tip of the iceberg. Back when the 300bps modem came out they figured the speed was as fast as anyone needed because it was near impossible for anyone to type more then 30 characters per second. Then the 1200 and 2400 bps modem cam out and they though those were as fast as anyone needed because almost no one can read at that rate. Then the 9200 and 14.4k because it takes almost no time to go to the next page 80x25 of colored text. then the 33.6k and 56k modems (Still the fastest modems for 1 normal telephone line) you can now download a 300x200x256 colors picture in no time. As bandwidth increases we find new ways to max it out and also with increased bandwidth we come with new methods of using the computer because it can now do it.

someone's trying to sound important

[dekeji]

"Cardinality" is the number of elements in a given mathematical set. When modems ran at 300 baud, you could forget about sending large data sets, such as images, because text and voice data took up all the available bandwidth. As connection rates increased, so did the cardinality of data that users could send. [...] Video currently represents the highest cardinality data

The term "cardinality" is wrong for several reasons. First, image data isn't represented as sets, it's represented as ordered sequences, and when talking about ordered sequences, both computer scientists and mathematicians talk about their "length", not their "cardinality".

Furthermore, what matters is not the size of what you want to transmit, but the rate at which you need to transmit it. We call that the "data rate" or (somewhat sloppily) the "required bandwidth".

So, the overall point of the article, that there is no single media stream that requires 10 Gbit bandwidth, is correct. However, that's pretty much irrelevant: file servers, video servers, and aggregate usage still require that kind of bandwidth. A family of four might require that bandwidth. You might want that bandwidth to have your backup happen in 1 minute instead of 10 minutes. So, there are lots of reasons to want 10 Gbit Ethernet, provided the price is right.

As for his use of the term "cardinality", the author apparently doesn't quite know the terminology of the field.

Saturated

[Dolda2000]

I don't get it - So you have 10Gbps LAN, and what now? Your PCI-X system only provides 4 Gbps. DDR400 RAM (which is the fastest right now AFAIK) is only 3.2 Gbps. You'd need like 4-band PCI Express just to get the required bus bandwidth, and even then, my 2200+ Athlon XP CPU can still just run arount 3500 instructions per second to process this amount of data, provided it can even get into RAM fast enough.

Seriously, when the LAN bandwidth closes in on that of the L1 cache bandwidth of even the fastest CPUs in existence, what the f* kind of hardware do you need to be able to actually do something with the amount of data you get?

10Gbps LAN has to be something for only the highest-performance servers in existence today, and even they should be more or less saturated by that kind of bandwidth. How long will it take before the hardware on the desktop will even stand a chance to process that much information?

Re:Saturated

[aXis100]

What about when 100 users want 4Gbps each over a comms link?

High bandwidth LAN has it's place, because in many situation the uplinks are bottlenecks.

Getting the job done

[SQL Error]

On Friday I was faced with having half a terabyte of data here which really needed to be over there. In a hurry. Unfortunately, the link between here and there was only 100Mbit, so people simply had to wait.

With 10GE it would have taken a few minutes rather than overnight. Or alternately, no time at all, because no-one would care anymore where the data was.

How long...?

[kimota]

>Having this much bandwidth would change the way we design. >What would you do with this much bandwidth?"

How long before we have viruses that target gigabit NICs on PCI-X machines for DDoSing?

--Kimota!

Comment removed

[account_deleted]

Comment removed based on user account deletion

What's the "good enough" point for bandwidth?

[kimota]

It seems to me that that at least with certain technologies, there's a "good enough" point, such as

a. billions of colors at 80+ Hz on a 21" monitor. There are special needs (or desires) for larger monitors or faster refresh rates, but while we might need better color accuracy, we don't need to worry about going for trillions of colors.

b. CD quality audio (or whatever makes the audiophiles happy. No, wait, *can* you make audiophiles happy?)

So is there a good enough point for bandwidth, a point at which pretty much all imaginable (or practical) datasets are transferred nigh-instantaneously?

I have a feeling the answer will involve the words "holodeck" or "matrix...."

--Kimota!

Framerate (Re:Dumb terminals? Cluster computing?)

[anynameleft]

Your 1.44 gbps is way overkill. LCD panels are the future, they are 60 Hz. But your eyes don't need those 60 hz for a smooth experience, 30 Hz is enough - AFAIK a cinema movie is only 25 pictures/second. That would mean 600 MBit were enough.
And you don't even need that bandwidth: when you do desktop work, most of the screen remains the same for most of the time, and when you are watching a movie you don't need such a high resolution nor 24-bit color. Now lets say that you want to watch TV and want to scroll 5 pages per second, then you would need:
1280x800x24x5 = 123 MBit
833x625x16x25 = 208 MBit
So a gigabit ethernet link should be enough to watch different channels on two TV's, to record two programmes on your harddisk and do two remote desktop sessions at the same time. Isn't that enough for most home users?

Re:Framerate (Re:Dumb terminals? Cluster computing

[swilver]

A cinema movie is 24 fps, and it generally looks smooth because it has motion blurring. Freezing a frame will show you that it isn't a perfect sharp picture but the moving parts are actually blurry, which helps give the impression of smooth movement (but only when played at the recorded rate of 24 fps).

If you however render perfect pictures (like your desktop does, but also games depicting a 3d world), then you need to go up to 70 Hz or more to make the animation look perfectly smooth.

On your desktop this is especially easy to see when dragging windows or moving your mouse pointer; if you ever thought your mouse seemed to stutter a bit instead of appearing perfectly smooth, then that's because in early windows versions your mouse update rate was fixed at 30 Hz. In newer versions you can actually change how often it is updated, making it appear smoother.

In games (and some movies) it is also easier to see stuttering when the camera moves from side to side or up and down (try strafing in a 3d game which is running at say 40-50 fps, and you'll notice it stutters, even though you would never notice it when you simply run forward).

Our eyes can easily see stuff up to 100-200 Hz, depending on how it is displayed. For example, the human eye can easily see a bright white flash lasting only 1/200th of second in an otherwise dark room...

Bit Torrents of course!

[ylikone]

and after a hefty session of downloading movies, mp3's, games you need to distribute these to other computers in your local network.

Medicine.

[Promethyl]

I work with PACS - A high quality raw picture server dedicated to medical imaging - X Rays and Heart Caths. I'd put the 10 gig server and 10 gig NICs in place so my doctors would quit b!tching about how slow the network is. Incidentally, for anyone else in my shoes, use Network Card Teaming with the EA. The ROC says it can't be done, but it can.

Bleh

[upsidedown_duck]

...we're working on a major project to re-architect our core application platform so that the different systems can be de-coupled and hosted separately.

Whenever I hear sentences like this, my mind's application platform fault-tolerant BS sensornet activates and plays soothing music to save me from becoming de-coupled myself.

thinking of bandwidth the wrong way

[GC]

you shouldn't consider bandwidth as how fast you can transfer data from A to B, but rather as the capacity of the network as a whole... 10Gbps networking is need more in backbones connecting 1Gbps or 100Mbps hosts that communicate amongst each other at their respective wire speeds.

I doubt you will ever get 10Gbps per second from a fileserver or web host, unless the data required is constantly in cache or high speed memory.

175Gbps ought to be enough for anyone.

[Doc Ruby]

5.1/16b/96KHz surround sound is 8.8Mbps and 8Kx6Kx32bit HDTV @60fps is 85.8Gbps, so A/V is 86Gbps, uncompressed. Two way communication means 172Gbps (1Mx: a million times CD playback rates). Symbolic application data might be another Gbps, and smell/flavor/tactile data is left as an exercise to the reader; consider the nominal 40Hz signal rate over the remaining 10 cranial nerves.

Re:Framerate (Re:Dumb terminals? Cluster computing

[mskfisher]

A cinema movie is 24 fps, and it generally looks smooth because it has motion blurring.

Yeah - I hate large landscape pans and quick motions in movies for that very fact - they aren't smooth. I can always see the framerate, and that bugs me. Now, when we get ~60 Hz framerate for movies (should be easier when we go all-digital), that'll come closer to full immersion for me.

Not just Sun anymore

[Namarrgon]

Try to find a host OS with a TCP/IP stack that can properly utilize 1 gigabit ethernet, let alone 10 gigabits. Hint: It ain't Linux...

Define "properly". If you mean efficiency, that's desirable but not critical. If an Intel/Linux server is 75% the efficiency of a Sun server, yet costs 30% the price, you can install two or three for the same bucks. That's efficiency of a sort too, yes?

Try to find a storage solution that can read or write that fast.

Well, in terms of raw sustained bandwidth, this doesn't seem all that difficult. A single Ultra320 SCSI HBA manages about 2.5 Gb/s, and 4-6 of those should meet requirements. Modern drives can sustain 400-650 Mb/s easily enough, 32 or even 24 of them would give plenty of headroom. 4 per HBA would be ideal. Even consumer 4-way SATA RAIDs would likely do the trick - being point-to-point they have more headroom than a shared SCSI bus (though less transfer efficiency).

Try to get all of the above, along with a 133 mhz. 64-bit PCI-X bus

Thanks, I'd rather use PCI Express. A 4x PCIe slot easily matches PCI-X, but it's point-to-point rather than shared, so I get that much bandwidth for each HBA. Motherboards are in production now with 4x, 8x and 16x slots, chipsets with 32 available PCIe lanes - that's around 80 Gb/s total bandwidth. A dual Opteron system today also has around 80 Gb/s memory bandwidth, and quad- and 8-way systems have much more.

Sun systems have traditionally been right up there with sgi for high-bandwidth servers while humble x86 consumer systems haven't held a candle to them. But that ole' world, it just keeps on changing...

Re:Not just Sun anymore

[Anonymous Coward]

Motherboards are in production now with 4x, 8x and 16x slots, chipsets with 32 available PCIe lanes - that's around 80 Gb/s total bandwidth

Theoretical bandwith. I doubt you'll get very close the that maximum. FYI no-one in the server industry considers that a 4x Gen1 slot is enough to feed 10Gb/s NICs (except Intel but they still don't understand servers). Protocol overhead kills you.

10Gb/s NICs is an important part of the debate to decide the signaling rate for PCI Express Gen 2. Intel wants 5Gb/s (2x Gen 1) but server companies want 6.25Gb/s so that they's be able to feed a 10G/s NIC with 2 lanes (instead of 4 5Gb/s lanes).

The reason for 10Gbps ethernet

[dayton967]

The reasons from my point of view, for 10G ethernet is not for end users, as I see rare requirements for the average user to require such bandwidth, in the near future. But with such devices, as IP KVM's, NAS, iSCSI, among many other devices that require high speeds, just think of a network, 24 devices of 1Gbps connected devices, needing to talk to a NAS at 1Gbps, all transfering a 1TB worth of data at the same time. And what if 2 devices that are talking to each other decide to change their priority or TOS so that they take precedence over all other traffic, that could be problems.

Application Server for terminals running X

[keeboo]

10 Gbps would be very useful for remote X.

I work in a Brazilian university and we have application servers and terminals running X.
Currently the application servers are local (the machine is physically in the local rack of its building). We cannot put them in a central place due to bandwidth limitations, what is not convenient since we have to care about the physical access security of each server.

Also, more bandwidth is nice for playing videos remotely. We cannot allow this currently due to bandwidth (ok, there's the server's CPU to consider aswell).

How much data can one person consume?

[SnprBoB86]

How much bandwidth will that person have to consume? ;-)

Thinking about it. There are hundreds of varying ways to reduce bandwidth overhead, but there are two primary reasons why people could always (and I mean ALWAYS) use more bandwidth.

1) Reducing bandwidth requires thought and time from developers as well as end users. Developers need to keep the size of the data sent down and end users need to use web friendly graphics or compress files. All that requires additional development time as well as end user training.

2) The more bandwidth the more global things can become. I would love to be able to go to any computer in the world and instantly load a full size movie. A full high definition quality, full length movie is a pretty heafty file and even streaming such a video needs a lot of bandwidth. What if I wanted to jump to any point in the video and fast forward and do all sorts of crazy stuff? I need the thickest water pipes possible :-)

Joe Sixpack CAN use gigabit Ethernet and faster

[orionware]

I am not an A/V professional but I sure benefited from upgrading the internal network of my house to Gigabit.

After dumping the hour of DV from the camera or capturing the latest episode of Deadwood (14 gig per hour of video) it's nice to have the bandwidth to transfer it to the Network Attached Storage (Go Linux!).

Having the extra bandwidth makes the NAS seem and act like a local drive. After editing the video I process it right to the NAS.

Having the extra bandwitch is noticable and I wouldn't want to go back to 100mb.

Beowulf baby or yea., or Dude Wheres my bus?

[tempest69]

This would rock for some supercomputing apps. If they manage to crank the latency down to something reasonable I'd be trying to steal a few hundred from the future.

Free the Bandwith..

Honestly some simulations really need the bandwith. One of the big supercomputers uses 6 gig-e lines in a torus formation. to get better bandwith.

Heres the other side.. 10 Gigs per second will take 1.25 gigabytes/sec per direction.. PCI64 is almost maxed ~70% by a full gig-E.

(350 MB/Sec *8bits /2direction= 1.4 gigabits max for channel) PCI X is only twice as fast.

(800 MB *8bits / 2 directions=3.2 gigabits max per channel)

So someone needs a be building a faster/fatter bus

In America we ask for faster, but settle for fatter :)

Re:FlashMob

[PingPongBoy]

Mod parent +1 funny

What will happen

[mattyrobinson69]

is coders will start writing programs to not consider bandwidth consumption, which will lead to a replica situation of what is now 'fasterprocesser == slowercode == fasterproceessor == slowercode etc'

Re:What would I do with this much bandwidth?-Music

[Lord Kano]

I can see from that high UID, you're relatively new here. This is Slashdot. You'd have better luck getting Planned Parenthood to recommend that you adopt out your baby than finding people here who actually RTFA.

LK

Re:What would I do with this much bandwidth?-Music

[Anonymous Coward]

Ya know, so far everyone seems to think of this as a long distance pipe. It's not, it ethernet. RTFA useful distance is in meters *NOT* kilometers. This is an intraoffice connection not a WAN pipe.

Single mode fiber gives you upwards of 20 kilometers.

Re:What would I do with this much bandwidth?-Music

[dglaude]

It can be from your office to your provider. Then encapsulated or transported accross very long distance with things like cisco ONS.

It is not unsusual to have 1G or ATM OC3 or ATM OC12 accross long distance (in my case 500 Km), except if I work on another planet.

No such thing as overkill

[Lewis Daggart]

Overkill only exists when the rest of technology hasn't caught up yet. Sure, you wont be able to fully take advantage of the 10 gigabits. There will ALWAYS be a bottleneck. This is good though because, for a while atleast, the connection speed wont be it.

what else but...

[chrismtb]

pr0n!

10GbE for SuperComputing

[nboscia]

Our plan for 10GbE is to support researchers with huge datasets (in the terabytes) who use our supercomputing facility. We currently use GbE, which is not sufficient for transfering such large amounts of data. So we are upgrading to 10GbE and also getting WAN connectivity at that rate (not sure if this is going to be 10-GigE-WAN-PHY or not) so that researchers across the country can transfer their data in a matter of minutes or hours, as opposed to days or weeks.

For those posters who are complaining about not getting near GbE performance, you are not properly tuning your system and network. You need think big - large frame sizes (network, 9k - 64k), large TCP windows (system buffers - think MB for GbE and GB for 10GbE), large I/O read/write(system disk), and account for latency (calculate your bandwidth*delay product). I've gotten constant ~980 Mbps throughput on a GbE network that was tuned.

you know what i'd do?

[pizza_milkshake]

two chicks at the same time.

10Gb filber != 10Gb Ethernet

[erice]

Sure, how many are Ethernet? Probably none. The "10Gb" telco standard is OC192.

Moo

[Chacham]

It's not how much they can send, it's how quickly it can be sent.

Re:What would I do with this much bandwidth?-Music

[Zaak]

useful distance is in meters *NOT* kilometers

The useful distance of 10GBase-CX4 is 10-20m. However, 10G ethernet is not limited to CX4. It has a fiber-optic PHY and can interoperate with OC192, both of which have a useful range of kilometers.

TTFN

I would...

[ZeroVerteX]

... stream multicast video and pop-up ads of natalie portman pouring hot grits down her pants to all my co-workers!

I aspire to be a professional troll one day.

Death to American dogs!

[Moqawama]

You have invaded our lands and raped our women and killed our holy fighters, and for this we will burn your cities to the ground! 11 September will come to America once again!