Where's Our Terabit Ethernet?

carusoj writes "Five years ago, we were talking about using Terabit Ethernet in 2008. Those plans have been pushed back a bit, but Ethernet inventor Bob Metcalfe this week is starting to throw around a new date for Terabit Ethernet: 2015. He's also suggesting that this be done in a non-standard way, at least at first, saying it's an opportunity to "break loose from the stranglehold of standards and move into some fun new technologies.""

Comments

2015

[DarkTitan_X]

Yes, and I suppose we'll have flying cars by then too?

Re:2015

[milsoRgen]

Yes, and I suppose we'll have flying cars by then too? And Windows will be smaller, faster and more stable!

Re:2015

[framauro13]

I don't know about that, but I'm looking forward to the retro '80's cafes. Plus we'll no longer be dependent on oil after Mr. Fusion is invented.

I doubt (read: hope) the Cubs will be World Series Champions though.

Re:2015

[Cctoide]

And it will have been rechristened "The Six Million Dollar OS", in homage to its price tag.

Re:2015

[Anonymous Cowpat]

and duke nukem forever will run on it...

Re:2015

[DarkTitan_X]

And maybe they'll abolish all lawyers.

Re:2015

[theun4gven]

Considering it would be pretty tough to sweep Miami.

Re:2015

[tenton]

Considering it would be pretty tough to sweep Miami.

Well, it's not out of the realm of impossibility. Since there's a precedent for a team to switch leagues (i.e. Brewers from the American League to the National), maybe the Marlins (technically play in Miami) could be forced to jump leagues. Seeing as the Marlins have 2 World Series already (they have to wait for that perfect storm of good young players and cheap salaries), the hard part appears to really be the Cubs making it to the World Series. :P

Re:2015

[darthfracas]

definitely won't be against Miami though.

Re:2015

[Anonymous Coward]

And Windows will be smaller, faster and more stable! Windows CE is typically stored on flash. Flash is smaller, faster and more stable than a mechanical drive.
Unfortunately, in CE most built-in program have no means to quit or exit them. So it's about as convient as a garbage can that has to be emptied one piece of trash at a time (force quitting) or buying a new garbage can when the old one gets full (hard restart).

Stranglehold?

[Brian Gordon]

I'd like to see the internet held together by his fun new technologies. See how well machines communicate without basic protocols.

Re:Stranglehold?

[KublaiKhan]

I see it as an opportunity for a new standard to evolve in a more natural fashion. Consider HD-DVD v. Blu-Ray--you have two competing formats come out, neither of which is compatible with the other's standard, but after a while it becomes apparent which one is going to be used.

Besides, it's not like this is going to affect TCP or IP or whatnot--this is way down at the bottom of the OSI model at level 1.

Re:Stranglehold?

[Jarjarthejedi]

Right, because corporate competitions in which two big companies do their best to ensure that their format wins the battle, with the individuals being frightened that their purchases will become obsolete is soo much fun.

Standards should be decided on BEFORE the material comes out. In this case it's not such a big deal, as the only people who are going to want terabit ethernet are huge enough geeks (or companies) to support whatever standard they choose but for the most part a lack of standards hurts everyone (just look at IE/Office, those are 'competing' standards...would you call them a good thing?)

Re:Stranglehold?

[Anonymous Coward]

Yes, but waiting for competing standards to shake out can be a huge waste of time and money.

Doesn't anyone remember the bad old days before TCP/IP over Ethernet became standard?

How many organizations are still laboring to expunge the last remaining vestiges of Token Ring, IPX, Netware, etc.?

Re:Stranglehold?

[geekoid]

I think you don't understand where he is coming from.

You would need to use the existing protocols on some level, but the protaocols to hit terabyte might need to be different. So he is saying Think about how to get reach the goal firsts, then delve into the protocol arena. If it is superior then eventually we would discard the older protocols and only use the new one.

Re:Stranglehold?

[houstonbofh]

I agree. After all, this worked so well for the American cell phone network.

Re:Stranglehold?

[moderatorrater]

In this case it's not such a big deal, as the only people who are going to want [HD players] are huge enough geeks (or companies) to support whatever standard they choose Your quote applies equally well to his example as to what you were saying.

just look at IE/Office, those are 'competing' standards...would you call them a good thing? They're not standards at all, that's the problem. IE's supposed to be compatible with the standard and it's not, so your example seems moot. Office has no standard at all, which would seem to be compatible with the discussion, but the big difference is that it's gone well beyond the point where there should have been a standard.

However, I don't think any products should make it to the market before there's a standard developed. Computer equipment has a way of going outdated very quickly when there's no standard attached, and I sure as hell wouldn't want to buy $1000 worth of equipment only to have everyone standardize to a different technology and leave me in the cold. At least your HD-DVD's will still play, if everyone switches to one type when you bought into the other type, your equipment becomes worthless.

Re:Stranglehold?

[KublaiKhan]

Not entirely.

There are still a few token rings and other such mesozoic cruft wandering around in the wild out there, but they still work--because some clever folks invented a way to get from one kind of network to another.

Keep in mind, also, that it's really only the early adopters--those who are willing to buy 1st-generation equipment--who would get 'screwed over', and they have, by definition (as the first generation of a given kind of thing is always several times more expensive than the 'production' generations), the money to waste on this sort of thing.

Re:Stranglehold?

[Penguinisto]

Besides, it's not like this is going to affect TCP or IP or whatnot--this is way down at the bottom of the OSI model at level 1.

Therein lies the rub, I think.

In spite of the OSI model (which TCP/IP doesn't map to very neatly BTW) - if one layer sneezes, they all catch a cold (with severity decreasing by distance). You screw with one layer, odds are good that you're gonna screw with its neighbors.

A good parallel of standards and what happens to them when folks try to create new paradigms? It can be found as close as your nearest fiber-based SAN installation (in its early days, anyhow)... "doesn't play well with others" is the most polite description I can find for it.

/P

Re:Stranglehold?

[Scootin159]

that should be an quick & easy transition, just like IPv6, right?

Re:Stranglehold?

[SuiteSisterMary]

X2 vs Flex56.

Re:Stranglehold?

[waterlogged]

"Besides, it's not like this is going to affect TCP or IP or whatnot--this is way down at the bottom of the OSI model at level 1"

Media Access Control and Logical Link are Layer 2
IP is Layer 3
TCP is Layer 4

Geek card....give it here.

Re:Stranglehold?

[KublaiKhan]

Physical layer is level 1. This covers the hardware--the cables and interface cards that would need to be modified to operate at the Terabit rates, and which is the primary concern in this case.

Layer 2 will be involved, of course, but the primary difficulties in this endeavor is going to be layer 1.

Re:Stranglehold?

[harry666t]

> your equipment becomes worthless.

Hey, you can always turn these HD-DVD readers into lasers!...

http://slashdot.org/article.pl?sid=07/08/09/0412215 ...but dunno how to get a HD-shark...

Re:Stranglehold?

[pragma_x]

X2 vs Flex56.

I think that's more of an exception than a solid example of what the GP was referring to.

Both sides were rolling out firmware based modems near the end of the "high speed" modem wars. In the end, a standard protocol was agreed upon (V.90 or just "V.Everything") and firware updates for the new standard was provided to everyone; so nobody but the very earliest adopters (no flash) had to replace their hardware. It was also resolved *very* fast - it took about a year after Flex was introduced, if I remember correctly.

Also 100% of that equipment could simply connect to your ISP of choice at 28.8kbps which, while not all that fast, is a rather nice failure mode for an "incompatible" piece of hardware. This is in sharp contrast to most hardware/format/protocol wars where any competing technologies are mutually incompatible with the competition's implementation, with option to run at reduced capacity.

IMO, a real consumer tech "war" is marked by the conflict ending with everone on the loosing side having to "upgrade" by buying a replacement for their purchased-too-soon piece of tech.

Re:Stranglehold?

[Anonymous Coward]

Blu-net, HD TCP/IP or Betacom?

I know which one I'm rooting for!

Re:Stranglehold?

[Anonymous Coward]

i kinda like the way dvd - vs + went... instead of one format winning every drive is now a + and - rw supporting drive. Blue-ray and HD DVD took different approaches, bd-rom with a larger 25-50 gb media using a smaller 'blue' laser, while hd-dvd was originally going with using 'mpeg-4' compression to get hd on regular dvds, (called 3x dvd in the current hd-dvd standard) but wound up using a 15gb blue laser media, because even with mpeg4 they couldn't get 2 hours on a 3x dvd.

since both hd dvd and blue ray used blue lasers dual format drives could have come out, but Sony spent a significant amount of cash for that not to happen. that, and the ps3 being a 'no brainer' for supporting blue-ray playback really made a difference. so many people who buy high def tvs buy a ps3 that best buy has a pile of ps3's right in the high def tv section.

Re:Stranglehold?

[GooberToo]

Besides, it's not like this is going to affect TCP or IP or whatnot--this is way down at the bottom of the OSI model at level 1.

Early research indicates IP protocols will not scale well with high speed links. CPU load goes through the roof and because of limited buffer sizes relative to line speeds, retries and fallbacks plague applications. The end result is a slow, high speed link.

In a nut shell, for high speed links to become useful to a large category of users, IP, and especially TCP must be revamped. Some research has already progressed down this road but late I heard, much more is required.

Wednesday?

[suso]

Um, they just made an announcement that they reached 16Tbits/sec on Wednesday, sheesh. Use the bandwidth you have for something useful.

Re:Wednesday?

[sm62704]

Um, they just made an announcement that they reached 16Tbits/sec on Wednesday, sheesh. Use the bandwidth you have for something useful.

Like porn?

Re:Wednesday?

[milsoRgen]

Like porn? Pfff, 16Tbits/s of porn? Hardly. Only LOLcat pictures need that kind of bandwidth.

Re:Wednesday?

[Anonymous Coward]

I CAN HAS TERABIT?

but but but

[grasshoppa]

we LOVE our standards. Without standards, where would we be?

K, just RTFA, and let me save the rest of you folks the suspense: There isn't one. It's a blurb about breaking standards and terabit ethernet. The slashdot summary just about nailed it.

Re:but but but

[alen]

slashdot's precious standards are like trekkie's and star wars fans precious canon

Re:but but but

[plague3106]

The slashdot summary just about nailed it.

So, are we at the start of the end times now?

Re:but but but

[milsoRgen]

Metcalfe says that the current approach being taken in the standards bodies won't get us to terabit rates. So, without going into too much detail, he said he expects a technology revolution, during which proprietary and innovative approaches to Terabit Ethernet will rule, at least at first. He said he sees it as an opportunity to "break loose from the stranglehold of standards and move into some fun new technologies." Ahhh, the struggle to stay relevant I suppose. Especially considering this guy has one awards from IEEE, a standards body. It almost feels he has an axe to grind from that short statement, at least in regards to the process perhaps. But then again he is a venture capitalist, perhaps he is laying down some good press for some startups he might have dumped some cash into? Also he has had some incorrect predictions before, my favorite being Windows 2000 would crush Linux.

Re:but but but

[TwilightSentry]

Depends on which armageddon standard you use.

• Christianity's rapture
• Supernova
• Microsoft Doom 2.0-as-a-Service
• etc.

Re:but but but

[masdog]

Wait? Star Trek has canon?

Re:but but but

[glittalogik]

Without standards, where would we be?

I don't know, but we probably wouldn't be getting laid there.

Re:but but but

[dbIII]

The prediction of linux getting crushed was just wishful thinking. The only thing that makes sense with his attitude in the past is that a linux user bit him and gave him rabies :) The reason he called linux users "communists" back then is that in some circles it meant the same thing that "terrorist" means now.

One of the inventors of ethernet and the guy that got the communist tag to stick to linux - he really can shape opinions.

Re:but but but

[cthulhu11]

We'd be in Redmond.

Re:but but but

[master_p]

No, we are at the end of the start times. Computer/IT technology is just leaving the state of puberty.

Four stories down, you whiner.

[binaryspiral]

Tag this story with "Scrolldownyouwhiner" ...

Long Time

[ilovegeorgebush]

7 years is a long time. Wouldn't it make more sense to work towards a new ethernet technology that has larger capacity? Think of the amount of data we currently send over the web etc. That's only going to increase. Those using ethernet on their networks I'm sure would prefer something that could deal with their daughter watching You Tube while their son is playing his friends on Duke Nukem Forever (haha!) on the LAN. Petabit Ethernet sounds more useful.
Meh, it's a shitload of data either way...

Re:Long Time

[sm62704]

7 years is a long time.

Seven years is the blink of an eye, kid.

Re:Long Time

[Durinthal]

Seven years is the blink of an eye, kid.

You might want to get that checked out.

Re:Long Time

[sm62704]

Seven years is the blink of an eye, kid.

You might want to get that checked out.

I did. I go back to see Dr. Odin about my eye on the fifth.

Thanks for asking.

Re:Long Time

[hraefn]

Seven years is closer to 25 million blinks. Which is about 115 days with our eyes in the process of blinking.

Does this mean that four percent of our lives pass in the blink of an eye?

Re:Long Time

[VENONA]

I wonder what you'd need for a bus? PCI, AGP 1, 2 or 4 won't do it. We'll need a standard there before we can a have a NIC to plug into it.

Who needs it?

[mollymoo]

One terrabit per second is roughly:

6 x as fast as 32-bit 2.8GHz HyperTransport
16 x as fast as x16 PCIe 2.0
60 x as fast as 20GFC fibre channel
400 x as fast as SATA-300
700 uncompressed 1080p HDTV streams (24bpp, 30fps)
15 million telephone calls

Other than the LHC, who the hells needs that kind of bandwidth?

Re:Who needs it?

[KublaiKhan]

Do you have any idea the volume of pr0n you could serve on that pipe?

Re:Who needs it?

[Alioth]

640k is roughly:

10 Commodore 64s
20 BBC Micros
640 ZX-81s
6 times a SDSS floppy disc

Who needs that kind of memory?

We might not need terabit ethernet *now*, but in 25 years time, it may be the basic expectation of your LAN's speed.

Re:Who needs it?

[Anonymous Coward]

I suppose you also think 640k of RAM is enough for anyone?

Re:Who needs it?

[L4t3r4lu5]

Every ISP in the world, to meet the bandwidth allocations they've sold fraudulently.

Re:Who needs it?

[archen]

Switches and machines that aggregate multiple saturated gigabit connections?

Re:Who needs it?

[marzipanic]

I was informed by a gamer that they do!

Other than that people who need everything yesterday I guess, well them and Virgin

Re:Who needs it?

[EriDay]

Regional ISPs. This is not a consumer product. Running ethernet on the backbone allows a homogeneous stack on all hosts from end to end.

Re:Who needs it?

[Anonymous Coward]

Other than the LHC, who the hells needs that kind of bandwidth? You just set yourself up for "640K should be enough for anyone" comparisons. Unless you are aware of all the innovations that will take place over the next decade, I think your response is short-sighted.

Better to have it an not need it, than to need it and not have it.

Re:Who needs it?

[kvezach]

Other than the LHC, who the hells needs that kind of bandwidth?

Evil overlords who want to build their own Borg collective? If 10^10 bits per second bandwidth is required (comparable to the bandwidth of the bundle connecting the brain hemispheres), then you get 100 drones per wire. (On the other hand, wired Borg would be really limited -- for obvious reasons.)

Re:Who needs it?

[Shados]

Yeah yeah, who cares about all that abstract stuff. How many LIBRAIRIES OF CONGRESS is it?!

Re:Who needs it?

[Anonymous Coward]

640k is roughly:


You miss the point.

There is nothing you can do with a big-ass pipe except move bits.

Plugging your firehose into the neighborhood drip irrigation system isn't going to get your lawn watered any faster. In situations where insane bandwidth can be installed end to end and there are insane amounts of data to move, this would be a great thing. However, the GP's point was that this really isn't the most common situation.

Most LANs have TONS of bandwidth to spare today. Work on an Internet (both pipes and servers) that can keep up with my cheap commodity HW LAN. Now, THAT would be useful.

This is not to say that there is no one on earth that needs to move insane amount of data in a LAN. Good on them. However just that increasing LAN speed won't help most folks while they wait for "the network"

Re:Who needs it?

[TeknoHog]

One terrabit per second is roughly:

81 lunabits per second.

Re:Who needs it?

[altoz]

Other than the LHC, who the hells needs that kind of bandwidth? 640k ought to be enough for everybody, too, right?

if it's there, it'll get used, probably for a purpose that's just the twinkle in some person's eye right now. think innovation, man!

Re:Who needs it?

[jandrese]

Maybe. One of the things that I've noticed is that as the bandwidth increases it becomes harder and harder to fill it up. Back in the Commodore 64 days it was not hard at all to run your machine out of memory by just typing a paper that was too long, and that's without graphics/charts/etc... These days there is no way a person would be able to type enough text to even make a noticeable dent in the main memory of any commodity machine. When everybody used 56k modems and serial lines it was trivially easy to fill up the link. However, when they moved to 10Mb Ethernet it got harder, but not impossible. Suddenly compressed music files were not a problem, although compressed video (DivX) still was. Then we went to 100Mb Ethernet and compressed video is no longer much of a bottleneck. Even now most modern machines come with Gigabit Ethernet ports that your average person can't fill with anything. Without new and bandwith intensive applications people won't be inclined to improve their bandwidth.

That's not to say someone won't come up with some application that requires a ton of bandwidth (distributed neural nets?), but none of our current applications would even really scale up to requiring 10GbE. The only realistic thing that comes to mind is some sort of Super HD video format, but anything like that is at least a decade away.

Re:Who needs it?

[Surt]

In 25 years we better have 10 terabit wired speeds. I would be surprised if 100 terabit wasn't becoming common.

Re:Who needs it?

[phorm]

In 25 years will be fine. But 2015 is well within that deadline.

I think the grandparent meant "who needs it *now* or in the near future"?

Re:Who needs it?

[Anonymous Coward]

must leach more porn!

Re:Who needs it?

[Kjella]

Haven't got a clue, but *my* Internet isn't fast enough so faster equipment all around means it'll trickle down. Do I need 4GB RAM in my box? No, but it's very nice to have when it's cheap. If they can get terabit connections for cheap, I'll happily settle for a GigE connection...

Re:Who needs it?

[Anonymous Coward]

Do you have any idea the volume of pr0n you could serve on that pipe?

That's exactly what we want. Volume-metric porn. No more of this 2D stuff. Truly 3D porn would take a huge amount of bandwidth.

Re:Who needs it?

[Anonymous Coward]

Its getting harder and harder to reach capacity because designers of said items are realizing the need to scale.

Which is exactly what we're doing here. Design it knowing that you'll reach a limit, and have a plan to move to the next scale.

Re:Who needs it?

[Anonymous Coward]

Jakob Nielsen once famously stated that the "perfect user interface" will require 1Tbps.

Of course I live in America so if I got 2Mbps through my broadband connection I'd be happy.

Re:Who needs it?

[cheesethegreat]

If bandwidth speed (along with encode/decode time on your local machine) were trivial, we could be doing things like centralized video processing. Instead of buying video cards for each computer, we buy a central video processor, and have everything routed through that. 100 computers on a LAN running graphic design programs rendering at full-tilt could chew up that bandwidth really quickly. And would save the company money on having to buy good graphics processors for each system.

Re:Who needs it?

[MortenMW]

Well, you have a (unintentional) point there. Currently the bottleneck is the hardware used in our computers. It will not help to have a 100 Tbps network if our hdd's can not store that much data at the same speed (or faster). Theres no real point in terrabit ethernet until we have clients with fast enough hardware.

Re:Who needs it?

[rrohbeck]

You're talking about data rate.

So, make that LibrariesOfCongress/nanofortnight for example.

Re:Who needs it?

[Anonymous Coward]

It seems like you are assuming that ethernet is only used in the comsumer world... I just finished helping wire up three cisco routers that we had to upgrade to because they have 10GbE ports on them and our pervious core routers with GbE were not able to handle the traffic any more. If cisco has to make a router with a total capacity of 92 Tbps, im sure TbE could be put to use.
http://www.cisco.com/en/US/products/ps5842/index.html

Re:Who needs it?

[failedlogic]

Side comment to your post. I thought it interesting that you noted the problem with typing papers on the C64. Its interesting now that computers are considerably faster that too much data, graphics and charts are added to papers that don't need to be there. I think the C64 would be a welcome regression for some professions!

Re:Who needs it?

[Shados]

Indeed. My short term memory had the last few lines of the parent post in mind, which were data quantities instead of rates, thus the mistake :)

LoC per nanofortnight is full of win though.

Re:Who needs it?

[xoundmind]

How many LIBRAIRIES OF CONGRESS is it?

Can't tell here, as I don't have Silverlight installed.

Re:Who needs it?

[FrzrBrn]

It's not a question of your end points needing that kind of bandwidth, it's a question of the links between ISPs and such. Look at any of the large router manufacturers today and see what kind of interfaces they're putting on their high-end gear: multiple 10Gbps ports. You can safely believe that if 100Gbps were available now that people would be using it. The step to 1Tbps is a large one, but there's no such things as "too much" bandwidth.

Re:Who needs it?

[Anonymous Coward]

There is nothing you can do with a big-ass pipe except move bits.

Free clue: 10Gb ethernet is currently used mostly in clusters and as backbones for large network installations to move lots of data around very fast. It's a long way off being a LAN technology. In seven years time, Terabit ethernet will be used mostly in clusters and as backbones for large network installations and 10Gb ethernet will be a LAN technology.

Re:Who needs it?

[ortholattice]

640k is roughly:....

- over 10 DEC PDP-11/45s running the RSTS time-sharing system

The maximum memory on these things was 28K words (16-bit) without memory extension hardware. In the 70s we had 8 users on a system with 28K of memory sorting lists, printing reports, data entry, editing with TECO, batch runs in the background at low priority, with relatively few swap thrashing problems. I implemented an ultra-low priority batch mode that waited until there was nothing else running for 5 minutes before activating a u.l.p job, then swapped it out instantly as soon as someone pressed a key - amazing what you could do with its sys calls. I did all sorts of hobby computing in this mode (with employer's permission) - looking for Fermat number factors, analyzing stock market timing data for patterns, you name it - I was like a kid in a candy store.

In retrospect, it was simply amazing how they shrunk it into that amount of memory and amazing how much I could do with it. 640K was simply unimaginable at that time.

Re:Who needs it?

[Anonymous Coward]

You mean go back to the way things first started ?? Or move to the windows "terminal server" or citrix model ?

Might be a good idea to look up X as in X11

Re:Who needs it?

[sloth jr]

It's needed as long-haul, not end-to-end. I would hope to see this implemented on the backbone. Then, standardize on last-mile internet connectivity of 100Mbit or more. I'd be just fine if YouTube supported decent video quality.

sloth jr

Re:Who needs it?

[Jeff DeMaagd]

The thing is, this terabit standard was "supposed" to be available now. I think the point was that the standard was way too far ahead of its time. Gigabit is just now becoming a baseline standard any day now. 10Gbit is barely on the horizon, 100Gbit still just a fantasy. Terabit will probably happen eventually, though I don't know if it would ever be used in a home because of the pressure to stick to wireless.

Re:Who needs it?

[LWATCDR]

I think you will see the migration to this the same way that a lot of people migrated to 1000Base-T.
They will use it for connecting switches and for SANs.
You may also see it replacing SATA or even for other internal links.
I don't think that the limits are the applications. The limit to how useful this is will the machines hooked to it. Right now I wonder how many machines could saturate a 10GbE link?

Re:Who needs it?

[Anonymous Coward]

I need it for my pr0n.

Re:Who needs it?

[commodoresloat]

you clearly aren't downloading nearly enough pr0n.

Re:Who needs it?

[PONA-Boy]

It's not a question of your end points needing that kind of bandwidth, it's a question of the links between ISPs and such. I, for one, am in no wise confident in ISP's and major backbone carriers using (of ALL things) ETHERNET to transfer data between one another. Ethernet was never intended nor designed to function as a safe, fault-tolerant, platform for Wide Area Networks. It is probably the most inefficient way of utilizing expensive long-distance hauls of any medium. For local area networks, which are by their very nature bursty and tolerant of retransmission, Ethernet in all of its forms is perfectly acceptable. Ethernet LAN's are cheap to build and easy to maintain. In the WAN, equipment and reliability become a much more expensive and tenuous affair.

Re:Who needs it?

[zeet]

Ever been to a colo 'meet me' room? That's a perfect use for Ethernet between ISPs. Right now, Gigabit links are the norm, and 10G links are starting to become reasonably common.

Re:Who needs it?

[Radhruin]

In addition, even though we HAVE gigabit ethernet today, all but the most expensive networking equipment (read: most consumer equipment) can't push that much data through. And even then, it's sufficient for most purposes. Most home built/lower end NAS devices don't even need gigabit as, even with gigabit equipment, they'll push 30MB/s or so on a good day before the CPU, NIC, or router/switch chokes.

Re:Who needs it?

[ClamIAm]

Well, it looks like ATSC TV channels take up somewhere around 20 megabits/sec, so would be one answer.

Re:Who needs it?

[PieceofLavalamp]

Open 30 youtube videos. Where would that happen? Go to a highschool and i bet you could see double that. It became such a problem in my old one that streaming video became strictly prohibited. Of course im assuming the problem was lack of availiable bandwidth and not a desire to not have to buy more bandwidth

Re:Who needs it?

[Anonymous Coward]

Regional ISPs. This is not a consumer product. Running ethernet on the backbone allows a homogeneous stack on all hosts from end to end.

Someone please inform Comcast.

Re:Who needs it?

[Anonymous Coward]

Wouldn't it be great to just *throw out* HyperTransport, PCI, Fibre Channel, SATA, IDE, HDMI, FireWire, USB, etc etc and replace them *all* with one kind of cable and one kind of plug? Imagine the ease of building a computer. Imagine the possibilities of building an incredibly scalable multi-processor machine where adding more processors is as easy as plugging them into the hub. Imagine the economies of scale we would reach in building Ethernet controllers and cables. Imagine ditching the limitations of all the specialized interconnects we have, like the impracticality of routing and splitting HDMI, the inability to plug two USB devices into each other directly, the inability to provide PCI expansion slots with easy access, and the inability to add more RAM or processor slots. Finally, (since terabit Ethernet will be fiber-optic) imagine Verizon installing it in their FiOS boxes, eliminating the last-mile bottleneck and allowing them to provide 1000 1Gbps Internet connections per box without any overselling.

Someday, when terabit Ethernet controllers are trivially cheap, all this can happen.

Re:Who needs it?

[Jarik_Tentsu]

And I'm sure there's more things that might bottleneck too in the way the data is transmitted on other levels of the OSI.

~Jarik

Re:Who needs it?

[Belial6]

You only see it that way because you don't understand how computers work. A C64 that did swapping out to magnetic media was going to easily hold every paper you were going to write. Today, we don't run out of memory when we load a DVD because we swap out data. most machines today are still 32-bit Windows. That means 3.5 gigs of ram. A single layer DVD is 4.7 gigs. If you try to load your entire DVD into memory, the way you did your paper on your C64, you are going to easily run your machine out of memory. Disk swapping is an ugly software hack that we have to deal with because memory has not even come close to keeping up with the amount of data we want to put in our computers. Back in the 8 bit days, it was even common to use some of your memory as a ram disk. This would be unheard of today.

Re:Who needs it?

[rastoboy29]

Two words: Game physics.

We've got ALL KINDS of room for improvement in physics realism in games.  I bet we could use that terabit ethernet if our processors were fast enough to handle all the data.

Re:Who needs it?

[Profane MuthaFucka]

If I had the bandwidth I wouldn't carry my laptop to and from work. I'd just make a VMWare virtual machine with 500 gigs of disk space and shuffle a copy of that back and forth every day to computers at home and at work. And if I needed a copy of that machine somewhere else, I'd just take 30 seconds to download it.

Re:Who needs it?

[svunt]

Average people are also moving away from LANs because wireless everything is so much cooler, regardless of performance. I've been amazed at how often I see people being thrilled with their pathetically slow wireless connections between non-portable machines in the same room.

Re:Who needs it?

[Anonymous Coward]

You haven't worked at a 3DFX company, where the standard project is around 300GB. Move that kind of stuff around on your network, and gigabit ethernet starts to feel slow. Also, when I do hard drive network backups, I don't want it to take an hour, I want it to take 5 minutes.

Re:Who needs it?

[porpnorber]

10Gb is about three channels each way of uncompressed studio quality HD - uncompressed being the easy way to get low latency. Been there, as a matter of fact, done that across the continent - it's *so* 2006 ;). I must say it looks very nice... but it's not HDR and you can still see the pixels. We can always use more bandwidth!

Re:Who needs it?

[Da Web Guru]

Gigabit ethernet connections have been the standard for years. Long gone are the days when an ISP/web hosting company standardized on DS3 or OC3/12/48 connections for their uplinks. All the major network providers are now supplying optical gigabit and 10 gigabit links to their customers. Even long haul links are sold in gigabit increments. *Everything* does gigabit ethernet. When you look at the network setups of the largest hosting companies, you see multiple gigabit (and 10 gigabit) connections everywhere, both on internal connections and external uplinks. When you look at the top-end networking gear from manufacturers (Cisco, Juniper, Foundry, etc.), you see line cards with multiple 10 gigabit ports spread throughout the chassis.

Re:Who needs it?

[Eivind]

Possibly. But the thing is, you're still going to use your ears and eyes for ultimately accessing the information. Which puts upper limits on what is needed.

Gigabit ethernet is cheap, reliable and available everywhere now. Lots of people and companies are still running a lot of stuff over 100Mbit ethernet because the improvements just aren't that important.

We're also at that point with storage. People buy laptops with 160GB storage rather than desktops with a terabyte, because it's /enough/ for their needs.

As technologies mature, eventually the breakneck pace of increases in performance will slow down. Yes, we're still growing very quickly, but large masses of people are already starting to care less. Upgrades become -less- frequent even as prices drop. Not because the brand-new isn't still 10 times more powerful than the 5-year-old, but because frankly, for many jobs 10-time-more makes no difference at all.

The single-core, 3-year-old Celeron laptop at home work just as well as the new Core 2 Duo at work for 99% of the stuff I actually do. The main exception being for those people who like to play 3D-games.

Re:Who needs it?

[PONA-Boy]

Long gone are the days when an ISP/web hosting company standardized on DS3 or OC3/12/48 connections for their uplinks. I'd like to know wherein you base this statement. Do you work in the industry? I can tell you from a number of years of experience that Telco's, even for large-capacity links, still very routinely sell NxDS-1 and NxDS-3 and OCxxx to their subscribers. ISP's and Web hosting company's purchase their Internet connectivity from these selfsame Telco's. Price being a consideration, only the most well-equipped ISP or hosting company is going to be able to afford even the NxDS3 and OCxxx connectivity.

I will certainly agree with you that the best-advertised gear from the top names in network gear are well invested with 10GB options. It is only natural that the big chassis-based routing switches _come_ with 10GB links considering their bread-and-butter market of large network Ethernet.

I still stand by my assertion that Ethernet is _not_ a solution for reliable long-distance data circuits.

Re:Who needs it?

[jandrese]

Way to put such a condescending first sentence in your post when the second line is so laughably wrong. No C64 word processor that I ever saw did swapping to magnetic media (which by the way was only 170kb, a reasonable person could fill it up with a paper) for several reasons. The biggest of which is that the C64 had no MMU, so doing that was painful. The second reason is that the C64 floppy drive was slower than molasses in January, unless you're talking about swapping out to cassette tape (the only other mainstream magnetic media available to the C64).

Also, memory disks are far from uncommon, although they tend to be used for things like CD based installers that need some scratch space or for times when you want to benchmark some IO intensive operation but don't want to get the disks involved.

I'm not sure why you want to load a full DVD into RAM either. I can't think of any good reason to do that. There are people who need 4+GB of main memory (those people tend to buy 64bit machines to avoid the evil that is PAE though), but your average consumer does not, at least not yet. We aren't to the point on memory where we are on network technologies. There are still some consumer applications that can make use of 2GB fairly easily (mostly games), and it's not like your average Wal-mart machine ships with 16GB (which is what it feels like when you buy one that has Gig-E). Eventually we'll need to move past it, but right now Gig-E is pretty hard for your average consumer to fill with any regularity.

Re:Who needs it?

[Belial6]

Your right that I was unnecessarily condescending. Sorry. That being said, you are still wrong in your comparison. The fact that C64 word processors did not swap to disk was in no way even close to a hardware limitation. It was a software choice. The same can be said for the choice to swap out a DVD. The point is that the ratio of memory size to data size was far better in the C64 days than it is today. Whether the software was as well designed in the C64 days compared to today has no bearing on the hardware resources available.

"I'm not sure why you want to load a full DVD into RAM either. I can't think of any good reason to do that."

For the same reason that you would want to load an entire paper into memory. So that you don't have to have to swap out to magnetic media when you move around in the data. Just because we have figured out some software tricks to ease the pain that data is increasing faster than memory, doesn't mean that it isn't happening.

I'm not saying that most people are filling Gig-E. I'm just pointing out that saying X is enough, has historically not been accurate, and your C64 analogy was backwards.

Re:Who needs it?

[qmaqdk]

One terrabit per second is roughly:

6 x as fast as 32-bit 2.8GHz HyperTransport
16 x as fast as x16 PCIe 2.0
60 x as fast as 20GFC fibre channel
400 x as fast as SATA-300
700 uncompressed 1080p HDTV streams (24bpp, 30fps)
15 million telephone calls
Other than the LHC, who the hells needs that kind of bandwidth? Think volumetric video. At 1920x1080x700 that is exactly one uncompressed stream...

For those of you playing at home, a TB is

[Anonymous Coward]

For those of you playing at home, a TB is a lot more than you can ever use in a million years...unless you link off the pirate bay, then it's not quite enough.

Re:For those of you playing at home, a TB is

[Macrat]

1Gbs is a bit slow when backup up a 1TB hard drive to the network server at home. ;-)

Re:For those of you playing at home, a TB is

[jcnnghm]

And I'd be willing to bet you're IO bound, not bound by the speed of your network. It is hard to saturate a gigabit link at this point, at least with consumer equipment.

Re:For those of you playing at home, a TB is

[brunascle]

really? i have a Gigabit network with a NAS and it seems to be just as fast as local drives (both are 7200 SATA 2). they both max out around 50MB/s. unless you have faster hard drives, I think the bottleneck on a gigabit NAS would be the hard drives and not the network.

Re:For those of you playing at home, a TB is

[Dan Posluns]

Not if your drive is still 7,200 rpm...

Re:For those of you playing at home, a TB is

[TubeSteak]

1Gbs is a bit slow when backup up a 1TB hard drive to the network server at home. ;-) 1 Gb is 128 MBs. According to Storagereview.com the Seagate Barracuda ES.2 is the only terabyte drive that has a transfer rate (104 MB/s) which maxes out high enough to even come near filling a gigabit pipe.

The bottleneck is your hard drive.

Re:For those of you playing at home, a TB is

[trickno]

I disagree. As technology improves in the media field, communication of your work will become important not only to the person trying to make a living off of it, but also the everyday user. Currently, it isn't uncommon to see a 10 Megapixel camera out, and personal HD-DVD camcorders. Chances are, sharing this media is done in a degraded state, not because of choice, but because people don't want to wait a day or more to upload there video.

Bottom line: Give the users a bigger backbone, and they will FIND a way to use it.

Re:For those of you playing at home, a TB is

[Anonymous Coward]

1gbs is 125 megabytes per second. Your problem is the speed of your hard drive(16megs/sec probably), not the network connection.

Re:For those of you playing at home, a TB is

[Anonymous Coward]

1 Gb is 128 MBs. According to Storagereview.com the Seagate Barracuda ES.2 is the only terabyte drive that has a transfer rate (104 MB/s) which maxes out high enough to even come near filling a gigabit pipe.

The bottleneck is your hard drive.

If you use just one hard drive, that might be true. These days, though, most people with those sorts of storage transfer issues have RAID-5 sources.

I know that the network is my bottleneck when copying from a 7-drive RAID-5 (6x SATA-300 effective data throughput...easily 200MB/sec) through a PCIe 8x host adapter (2000MB/sec) to the gigabit ethernet (125MB/sec).

On the other hand, if I'm not backing up multi-gigabytes of data, normal access over the network isn't much slower than local drives for real-world performance, especially with the large caches in the servers that avoid the disk altogether.

Re:For those of you playing at home, a TB is

[Anonymous Coward]

> 1Gbs is a bit slow when backup up a 1TB hard drive to the network server at home. ;-)

A 40MB/s (~320Mb/s) 1TB hard drive is slower.

Re:For those of you playing at home, a TB is

[Slim Backwater]

RAID-5 is EVIL! Never Ever, EVER use RAID-5, You will LOSE DATA! RAID-1 or RAID-10 only for production use.

If you are thinking about Raid-5, forget it, just stripe your drives in a RAID-0 and enjoy the performance benefits and keep frequent good backups and test your restores.

Why RAID 5 stops working in 2009
Why aren't disk reads more reliable?
End of Raid 5

finally, BAARF - Battle Against ANY Raid Five BAARF

HTH, HAND, don't cry.

Re:For those of you playing at home, a TB is

[WuphonsReach]

If you use just one hard drive, that might be true. These days, though, most people with those sorts of storage transfer issues have RAID-5 sources.

RAID 10 is a lot more popular in data centers. With RAID6 running a close second.

RAID5 (and RAID6) has crappy rebuild times (the more disks in the array, the longer it takes to rebuild). Which means that the chance of a double-failure killing the entire array during your recovery window go up as you add more disks.

RAID 10 recovery windows are based on the size of an individual disk in the array. Doesn't matter if you have a 4-disk array or 16-disk array. A rebuild of a failed drive always takes the same amount of time. It's basically RAID0 laid over top of a bunch of RAID1 mirror sets. Each mirror set gets rebuild individually.

You can even do interesting things with RAID10 where you put the RAID0 over top of triple-mirrored RAID1. (Three disks per RAID1 element. Which means you have to lose three disks before the array crashes.) Granted, it's not very space efficient when you do that (33% net space), but could be useful for the truly paranoid.

Other then that... yes, RAID 0/5/6/10 does a wonderful job of filling up network pipes and giving you a higher sustained transfer rate then a single disk could provide.

ethernet and collisions

[Anonymous Coward]

The original idea behind ethernet was to enable shared media through collion detection. Other than low performance SOHO hub use, isn't collion detection obsolete? Hasn't the need for collion detection been largely eliminated through switching technology?

What do the netork gurus have to say? With switching technology couldn't we chose another more efficient standard than ethernet?

Re:ethernet and collisions

[TeknoHog]

Wikipedia says: "10 Gigabit Ethernet abandons half duplex links and repeaters (and the CSMA/CD that goes with them) in favor of a system of purely full duplex links connected by switches as was already the normal practice with gigabit Ethernet."

I'm SHOCKED, shocked I tells 'ya

[SengirV]

Really, you mean one of these nebulous 5, 10, 20 years from now predictions actually hasn't come true? Amazing.

By now, I'd have thought that that with all the blown predictions like this, that it would only be a story if one actually came true.

I'd sooner have...

[Channard]

.. a technology that lets homes receive fast internet no matter where they are. My area's not cabled up, and thanks to me being too far from the exchange.. I just live in a normal street .. I can't reliably get more than 512KB a second. Fix that, and you'd be laughing. Powerline networking, maybe?

Re:I'd sooner have...

[Anti_Climax]

Well if you're in the US, the greatest probability has you in the ATT/SBC/Bellsouth regions. Regardless of your ILEC, did they stick you with a shitty CPE like an actiontec or speedstream? If so switch it for something with a decent AFE like a 2Wire and you might be able to push that more toward the 1Mbit range. Not much of an increase but better than nothing.

Also remember that, even if you get a decent DSL modem, they may still have you allocated under a lower performance profile just out of average expectation of attainable speed. If you contact tech support and you're lucky enough to get someone knowledgeable, they can get you in contact with the group that actually manage the provisioning at the DSLAM and get them to try some higher rate settings on their end.

Good luck.

Re:I'd sooner have...

[bazorg]

well I'd sooner have teleportation or a vaccine for dental caries but what the hell...

Re:I'd sooner have...

[mikael_j]

Also remember that, even if you get a decent DSL modem, they may still have you allocated under a lower performance profile just out of average expectation of attainable speed. If you contact tech support and you're lucky enough to get someone knowledgeable, they can get you in contact with the group that actually manage the provisioning at the DSLAM and get them to try some higher rate settings on their end.

I don't know if this is one of these situations where the US ISPs seemingly attempt to be different for the sake of being different (for example, AFAIK PPPoE is fairly rare outside of the US but in the US it's almost impossible to get DSL without PPPoE) but with the ISPs that I have worked with the standard DSLAM profile has always been 8M if the customer is paying for 8M, 24M if the customer is paying for 24M and 2M if the customer is paying for 2M (or any other speed they're paying for). If there's a problem the first step from the ISP's side is generally to activate interleave on the connection, if that doesn't work and there are problem they'll start lowering the speed but there's no point in setting the speed to lower than what the customer is paying for from the start as the predicted max speed isn't very reliable, some people living practially nextdoor to the DSLAM can't get more than 4-5 Mbps and there are people living several kilometers from the DSLAM getting speeds 20-40% higher than what would be expected on an average phone line...

/Mikael

Re:I'd sooner have...

[Anti_Climax]

I'm sure it depends on the carrier that's configuring it. Most of the people that are paying for basic DSL service in the US - I think ATT sells a sub-$20 1.5MBit line, I've gotten the same from Qwest for $17 - but are far enough out that they may not see all of it fall under the verbiage in their contract that says "Up to 1.5Mbit". From there it is a question of company policy and motivation. If the install and config is done by a tech, they might or might not call in to their provisioning group and try different profiles if they don't hit max. If it's a self install, the DSLAM may be provisioned before the CPE is even connected to a phone jack and they'll have nothing to base expected performance on outside a table of expected throughput based on their CPE of choice. Can't hurt to ask. And of course there's the possibility of the telco wiggling around in that "Up to X" promise, loading slower profiles to allow greater overselling of bandwidth to a given DSLAM, but I've seen no reason to believe that is happening... yet

The likely hood of an ILEC employee being lazy is pretty high and, like I said, can't really hurt to ask.

Porn measurement unit:

[Anonymous Coward]

Hotties per nanosecond

Misleading name, "Ethernet".

[pla]

"He's also suggesting that this be done in a non-standard way"

No, he suggested that five years ago

We don't yet have the technology described (wave division multiplexing) in our homes because very, very few of us want to bother with fiber in our homes at all.

You can push an amazing amount of data over glass, no one would claim otherwise. You can't, however, drape it across the floor and up the stairs to your switch for a quick LAN connection... Not only does terminating a fiber suck, the first time the dog steps on that little yellow wire, end of connection. By contrast, I've used Cat5 as a structural material (tied a PC to a hook on the ceiling with it) WHILE USING IT for data.

So no, we won't see terabit ethernet anytime soon, unless someone figures out a way to push it over copper.

Re:Misleading name, "Ethernet".

[plague3106]

Sorry, why would you leave any cables out where people can step on them?

Re:Misleading name, "Ethernet".

[TheRealMindChild]

Are you serious? Haven't you ever, say, had a guest come over with their computer up into the guest bedroom when OOPS! you realize you don't have any cable run to the room... just get a really long cable, drag it around the corner, up the stairs and into the room. You don't always have the time, patience, or foresight to run cable through the walls to every room in the house.

Re:Misleading name, "Ethernet".

[plague3106]

A dog stepping on a cable sounds like a more "permanent" run. I wouldnt worry about the dog for the occasional guest. Besides guests are what 802.11 is for...

Re:Misleading name, "Ethernet".

[mollymoo]

If you can lay fibre across the bottom of an ocean and drag it through conduits you can sure has hell run it up your stairs. You're just using shitty fibre leads. I work in a telephone exchange and trust me, it take more than a dog stepping on it to kill one of the fibre patch leads used there. Terminating it does suck, but most people don't terminate their own CAT-5 either.

Re:Misleading name, "Ethernet".

[pla]

Sorry, why would you leave any cables out where people can step on them?

Because very few of us have the luxury of hardwired data jacks in every room of the house. Personally, I have exactly three hardwired drops in my house, with a switch at all three, but that still requires running a cable (at least) down the hall if I want a connection anywhere other than my office, library, or living room.

And even if you do have God's Own LAN, you probably still don't want to run a new wire or three through your walls just so a few friends can come over for a LAN tournament of your favorite FPS.

Re:Misleading name, "Ethernet".

[H0p313ss]

just get a really long cable, drag it around the corner, up the stairs and into the room. You don't always have the time, patience, or foresight to run cable through the walls to every room in the house.

Yes... but I stopped doing that years ago when wireless became cheap, convenient and ubiquitous. Surely I'm not the only guy who'll remote desktop from the kitchen to the living room?

Re:Misleading name, "Ethernet".

[ChrisA90278]

I disagree. I bought a spool of fiber for some users who needed to deploy a temporary network and then roll it back up and use it again later. We bought a one kilometer roll on a wooden spool and they would just as you say pull it down stairwells through doors and toss it up in trees. Once they hung it over a freeway in Germany from some utility poles (had to hire local linemen for that one) and then after a few day rolled it back.

I told the fiber cable sales guy I was going to test their sample by placing it in the parking lot and letting cars drive over it for a while. The cable was tough basically it was a bundle of kevlar around a thin fiber strand. The kevlar absorbed all of the abuse. After all they lay fiber cable in the ocean. If it can take being dumped off a ship into the ocean it can take a dogs stepping on it. The trick is to specify the correct cable and don't just buy whatever is cheapest.

Re:Misleading name, "Ethernet".

[djp928]

Fiber is actually surprisingly tough. Yeah, it's glass, but as long as the light gets through, you've got a good signal. Where I work we once accidentally rolled a 700 pound cabinet over a fiber run. The fiber rolled up under the caster and got crunched a second time (once from the wheel going over it, once from it curling over the wheel and getting jammed in the housing) and we had to slowly back the cabinet off the fiber while gently tugging on the free end to get it out of the wheel housing. It came out all mangled looking, and we thought there was no way they would still work. But we got the cabinet in place, plugged the fiber in, and what do you know? It worked fine. We used those mangled cables for four years, and they were still working when we retired those boxes this year.

So yeah, in general fiber is less sturdy than standard copper cables. But they're not as brittle as all that.

Re:Misleading name, "Ethernet".

[pla]

If you can lay fibre across the bottom of an ocean

You mean to compare 7+ layer armored "sharkbite" cable with the sort of single nearly-naked fiber you'd use as a patch cable? You have waaaaay more money than I do!

I work in a telephone exchange and trust me, it take more than a dog stepping on it to kill one of the fibre patch leads used there

You probably have much higher quality fiber (or more accurately, fiber with much stronger jacketing) than the than what average Joes would use. But that example about the dog came from real life... A 50ft run (which at the time cost some ungodly amount - well over a dollar per foot), destroyed with one misplaced paw.

Re:Misleading name, "Ethernet".

[Anonymous Coward]

There are different types of fiber cables (well, the actualy fiber doesn't really change much.) The really thing stuff that's basically just some kevlar and plastic is NOT what is used in more hostile environments. I used to use tactical fiber in my last unit (USAF) which had several strands within a pretty covering. Supposedly it could handle being run over by our trucks and whatnot but we never really torture tested it. The stuff they lay in the ocean is definately nowhere near what you see in homes and on spools that fit in/on your vehicle.

Re:Misleading name, "Ethernet".

[plague3106]

What, you never heard of wireless? Seriously, I'm not an early adopter and I've had 802.11b since 2002. I guess I'm not lame enough to host a LAN torny..

Re:Misleading name, "Ethernet".

[pla]

What, you never heard of wireless?

Yup. Heard of it, tried it for about a year (and with a decent WAP, too, not the $19.99 Linksys special), got sick of having my connection flake every few minutes, and went back to wired.

Re:Misleading name, "Ethernet".

[sexconker]

You mean one misplaced cable.
The dog was exactly where it was supposed to be - walkin' around doin' it's thing.
You were the one who decided to sloppily lay some cables down.

When the consumer market needs fiber, they'll make it thicker. It's not really an issue.

Re:Misleading name, "Ethernet".

[WithLove]

You must be new around here bro, most of the guys here prefer their basements...

Proprietary protocols

[esocid]

So, without going into too much detail, he said he expects a technology revolution, during which proprietary and innovative approaches to Terabit Ethernet will rule, at least at first.

At first? With the way patent trolling is going right now, I wonder if anyone will see it all, but I commend him for trying to break out of the strangle hold that corps have on the standards in favor of innovation rather than profit and stifling competition.

Interrupts

[Anonymous Coward]

If each incoming packet results in a CPU interrupt, and the OS has to process these, wouldn't terabit ethernet make the CPU a considerable bottleneck? I know that for gigabit ethernet it's already something of a problem, now picture that situation being 1000 times worse.

Bob Metcalfe, hater of open source

[Bogtha]

Has this guy done anything relevant in the past couple of decades? Here's a choice quote of his:

Unix and the Internet turn 30 this summer. Both are senile, according to journalist Peter Salus, who like me is old enough, but not too old, to remember. The Open Sores Movement asks us to ignore three decades of innovation. It's just a notch above Luddism. At least they're not bombing Redmond. Not yet anyway.

The hard part of being down on Linux and the Open Sores Movement is worrying about that menace hanging over us at year's end. No, not Y2K, but Linux's nemesis, W2K, Windows 2000, the operating system formerly known as Windows NT 5.0.

W2K is software also from the distant past -- VAX/VMS for Windows. But it will overpower Linux. NT, now approaching 23x6 availability, is already overpowering Linux. NT and NetWare constitute 60 percent of server software shipments. All Unixes make up 17 percent, and Linux is a small fraction of that. When W2K gets here, goodbye Linux.

Re:Bob Metcalfe, hater of open source

[adam.dorsey]

Wow, that was fantastic. I needed a laugh.

I can't take anyone seriously who uses terms like "Open-Sores," "Micro$oft/M$," "Windoze," "Linuzzz," etc. It makes a person sound like a 13-year old script kiddie.

Re:Bob Metcalfe, hater of open source

[anticypher]

I had the pleasure to work on projects associated with Metcalfe at the beginning of my career, notably the migration of Ethernet I to Ethernet II standard. He was an autistic, anti-social, self-centered, egotistical curmudgeon from the start, and despite those charming qualities he nevertheless adopted an ivory-tower academic approach in his later life of hating anything created since his 15 minutes of brilliance.

He can always point to DJB as a worse curmudgeon, so there is that solace in knowing he isn't the most disrespected hasbeen still seeking the limelight.

the AC

I don't think a smattering of emoticons in this post will stave off the imminent hater responses, and there isn't really anything I'd put a smiley to.

stranglehold of standards

[thinktanked]

Stupid sexy standards.

Wait... you believe Metcalfe WHY?

[1336]

As in the Robert Metcalfe whose Wikipedia article has an "Incorrect predictions" section listing where he wrongly thought that "the internet would suffer a catastrophic collapse" in 1996 and this gem:

Metcalfe is also known for his harsh criticism of open source software, and Linux in particular, predicting that the latter would be obliterated after Microsoft released Windows 2000:

The Open Source Movement's ideology is utopian balderdash [... that] reminds me of communism. [...] Linux [is like] organic software grown in utopia by spiritualists [...] When they bring organic fruit to market, you pay extra for small apples with open sores - the Open Sores Movement. When [Windows 2000] gets here, goodbye Linux.

Just because he did something really cool 35 years ago doesn't make him an expert on related matters now.

Re:Wait... you believe Metcalfe WHY?

[TheRaven64]

he wrongly thought that "the internet would suffer a catastrophic collapse" in 1996 To be fair, IE 3 was released in 1996...

Re:Wait... you believe Metcalfe WHY?

[dbIII]

With malware at proportions you wouldn't expect in bad scifi it is amazing that the internet hasn't suffered a catastropic collapse.

That's what they said about a GB/s back in the say

[davidwr]

25 years ago we were using 300 and 1200 bps modems. Now it's like 1.5-6mbps for most people. That's 5000x faster to the outside world.

That's roughly a 5.5x increase every 5 years.

Within our homes it's gone from 100K-or-so serial cables to 100-1000Mbps Ethernet, and that's not even counting specialty protocols like SATA. That's a factor of 4-6.3 every 5 years.

So, by 2013 we should be in the 8-33 Mb/s for typical home users, and in the home we should be between somewhere near 63Gb/sec to network the gaming boxes together.

It's about Shannon's law too.

[Z00L00K]

For everyone that has been working with communication since the early datacom ages Shannon's law has been important. It's still important, and it means that you can't just push everything through, you have to consider the media used.

In a way it can be tweaked a bit, and that has caused a confusion among those that aren't well into the technological difference between Baud (modulation changes per second) and BPS (bits per second).

Anyway - The classical phone modems can have a speed up to 56kbps, but effectively they stay at 28 to 33kbps. And that on a line that actually only provides 3kHz bandwidth. The trick is that in the 3KHz bandwidth you can have a carrier with less than 3000 modulation changes per second, often 2400. In each modulation change you not only have one bit transferred, but multiple bits. This is achieved by having a variation in both phase and amplitude of the signal.

So to utilize the cabling at the extreme speeds that a terabit Ethernet is you may have to resort to the same technique.

There have also been other techniques in use like using multiple carrier frequencies, like what the Telebit Trailblazer modems did. That technology was very resilient to interference compared to the CCITT standards, but it had other disadvantages instead.

Front-end handler

[davidwr]

Who says the CPU has to handle all the load.

You can design your hardware so the CPU only gets interrupted when it needs to.

If you have a smart front-end processor, you can have the front-end processor bundle up IP- or insert-your-own-protocol packets and send them to the CPU as needed. Heck, if it's really smart it can even handle entire TCP streams on its own. Imagine only interrupting the CPU when it had the results of an entire HTTP GET request in hand. Or imagine downloading your favorite movie and having the front-end processor do all the work, shoving the data to RAM directly and alerting the CPU every MB or so.

Hmm, come to think of it, didn't the Internet begin with front-end processors or dedicated devices the size of a large refrigerator?

Re:Front-end handler

[Anonymous Coward]

If you have a smart front-end processor, you can have the front-end processor bundle up IP- or insert-your-own-protocol packets and send them to the CPU as needed. Heck, if it's really smart it can even handle entire TCP streams on its own. Imagine only interrupting the CPU when it had the results of an entire HTTP GET request in hand. Or imagine downloading your favorite movie and having the front-end processor do all the work, shoving the data to RAM directly and alerting the CPU every MB or so.

That's really ridiculous. Under that scheme, every time a new killer app comes out, everyone has to buy lots of dedicated hardware.

It's easier to just interrupt the CPU only once when short bursts of packets come in.

Progress!

[DragonWriter]

Five years ago, we were talking about using Terabit Ethernet in 2008. Those plans have been pushed back a bit, but Ethernet inventor Bob Metcalfe this week is starting to throw around a new date for Terabit Ethernet: 2015.

So, 5 years ago, Tb-E was 5 years away, and now its 7 years away. So by 2015, it should be about 10 years away, and by 2025 it should be about 14 years away, etc.

Re:Progress!

[1336]

Sounds familiar ;)

Meanwhile, at Disco Stu's "Can't Stop the Learnin' Disco Academies"...

Disco Stu: [making indescribable body motions] Did you know that disco record sales were up 400[%] for the year ending 1976? If these trends continue... A-y-y-y!

Re:Progress!

[Hugonz]

So, 5 years ago, Tb-E was 5 years away, and now its 7 years away. So by 2015, it should be about 10 years away, and by 2025 it should be about 14 years away, etc.

Talk about exponential backoff...

Just got back

[Anonymous Coward]

I just realized that slashdot doesn't have a way to post pictures in the forums. I'm surprised by this since it's a common feature at many other sites. Is there a suggestion box or I just email Cmdrtaco about it?

Hmm...

[brkello]

Let's see, 10 Gb Ethernet came out about 5 years ago. It started out only being able to do about 2 Gb/s with tweaks to PCI-X. PCI-X, at best, I have only seen capable of doing around 6Gb/s with a NIC. PCI-e x8 does much better and you can get very close to ~10 Gb/s. Theoretically, 8x can do 16 Gb/s...and there is a 16 lane version as well. Current standards that are being developed are for 40Gb/s and 100Gb/s. 40 isn't too bad since backplanes can already do that speed, of course, if you have multiple ports you would have to be doing blocking. I'd still say that was a few years off from any real products and then a few years of it sucking before it gets good.

Quite frankly, 2015 is a joke. I see a lot of these people in high places make these incredibly wild predictions. I am not sure if they are just off in their own little land or if they make aggressive predictions to try to drive the industry. So maybe they can put out some strange switch that could go this fast by that time but nothing that would be practical. 100G isn't even out, or at least nothing I can touch. I'd say it would be at LEAST 2020 before any networking equipment will be capable of this, much less used.

It is an issue of balance.

[jellomizer]

It is not state that we will never need Terabit eathernet. It is an issue that we don't need it now... And our computers will need to run a full performace just to handle the data. Sure we could probably use Terabit Eathernet for such things like distributed computing or Speedy downloads. However for most tasks Gigabit is fast enough and still isn't being fully utilised.

Re:It is an issue of balance.

[Vancorps]

I would think terabit ethernet would become mighty handy on the backbone although you certainly wouldn't need it in the home. Distributed computing can and eventually will take up this bandwidth. You're right, for most tasks gigabit is enough. Course with my 72port gigabit switch connecting to my backbone, now I'd need 72gigabit at least to sustain full bandwidth which I can't achieve. So I bottleneck it at the IDF with dual 10gig links. Of course the main bottleneck is the server then as dual gig nics can't service even 10gig so now it's time to upgrade the NICs in the servers since PCI-E can actually handle the increased throughput. Of course the SAN back-end can keep up without a problem.

I will Settle For 1Mbps

[s31523]

Forget terabit ethernet. I will settle for full, actual 1Mbps (10,100, 1000, etc.) speed for both transmit and receive. Even on my home network, I rarely get full %100 utilization on my LAN. Some PC's are linux, some are Windows. Neither machine ever really reaches its full potential. I looked at other networks as well, even my work LAN, and they have similar issues. I am not a network guru and don't want to spend the time tweaking and configuring. The damn Gbps NIC and network drive I bought should just plug and go and I expect to see speeds reasonably close to 1Gbps, but nope. I see like 1% utilization. Seriously, lets make current technology work as advertised before we start claiming super-fast speeds on other vapor-ware technology. Please?

Re:I will Settle For 1Mbps

[ivan256]

Spend 5 minutes troubleshooting.

Consumer grade copper gigabit in crappy low-end PCs (made in the last 4 years) should be able to give you at least 300mbit of transferred data over TCP given 10 minutes of tuning, and using the correct cables.

Don't use a USB NIC. Don't transfer your data to/from a 4000rpm laptop hard drive... Etc..

You're not going to get 1Gbps though, 'cause your hard drive probably can't go that fast. The average low-end desktop drive isn't going to give you more than 30MB/sec. Depending on your system, the bus you have the NIC plugged into can't do 1000mbps. Your network can handle the advertised speed just fine though. If you've got high end gear (motherboard, disk array) you can peg a gigabit ethernet link in a point to point transfer... Right now it's not the ethernet holding consumer grade equipment back.

Re:I will Settle For 1Mbps

[Aram Fingal]

I've spent more time on this than it probably deserves and I have found that the order of performance can be the opposite of the advertised speeds. With all other hardware equal, I have found that, usually, FireWire 400 is faster than USB 2.0 which is faster than gigabit ethernet (with a cat 6 crossover cable).

Re:I will Settle For 1Mbps

[tknd]

Tip: use FTP or something really low level to determine the actual bandwidth you're getting over the wire. There are also some programs that will specifically test bandwidth without other limitations (disk speed).

I've found some protocols like SMB can be really flaky at high speeds. But FTP has reliably been able to hit the hard disk transfer speed limit which is much lower than 1gbps.

Re:I will Settle For 1Mbps

[ivan256]

GigE is really much faster than the other two, but it is suited to different kinds of traffic. FireWire and USB mass storage do small block transfers. GigE preforms optimally with larger packet sizes. You should use a protocol which takes this into account. Not SMB on Windows unless you want to do a *lot* of tuning... FTP with a client that supports a sliding window size (Not Internet Explorer, but almost every other FTP client), NFS, HTTP, SCP if you have a very low latency link and a fast CPU...

Re:I will Settle For 1Mbps

[kasperd]

The average low-end desktop drive isn't going to give you more than 30MB/sec.

The IDE drives I bought in 2004 were capable of 55MB/s. The SATA drives I bought last year are capable of 75MB/s (I even managed to get 80MB/s over eSATA once). If you copy across a 1Gbit/s link, the hard disk is still going to be the bottleneck in many cases. But if you only get 30MB/s it is because you use a file system which is not capable of utilizing the disk's full potential.

Re:I will Settle For 1Mbps

[ivan256]

You probably bought higher speed drives than the bottom of the line. Sure, you're getting 55MB/sec, but you're not using a low-end drive. Probably something more middle of the road. But buy a $299 Dell and test the speed of the drive that comes with...

When I run disk benchmarks, I don't use a filesystem. Since 2000, you've been able to buy drives that can hit over 80MB per second. You can also still buy SATA drives that have trouble hitting 30MB/sec. (3.5" drives. At 2.5" you can easily buy drives that have a hard time with 25MB/sec. You can also get 2.5" server drives that smoke...)

If you are losing 50% of your speed from a filesystem... Well.. You're not. It's not even worth contemplating.

Re:I will Settle For 1Mbps

[arcade]

I quite simply do not know what you're doing wrong.

I've managed to get up to 600Mbps out of my Gbps NIC's, with only a very small bit of tuning. That was from /dev/zero on one machine to /dev/null on another machine.

File transfers over NFS thypically maxed out at around 200Mbps.

Managed to tweak my NFS speeds up to bursts of 500Mbps for rather small files (less than diskcache * number of disks in the raidset) - guess the rest was overhead. My problem there was that I had to enable jumbo-frames and so forth.. which was a tad of a mess, never worked properly and so forth - so I went back to my "slow" speed of around 200Mbps and didn't care too much.

Re:I will Settle For 1Mbps

[kasperd]

You probably bought higher speed drives than the bottom of the line.

I have usually been going for the lowest price per GB. I think that usually means I pay about three times as much as the price of the cheapest drives, but I get more than three times the capacity. Speed was never a criteria I considered when bying drives.

When I run disk benchmarks, I don't use a filesystem.

Sure, that is the only proper way to test the speed of the drive itself.

Since 2000, you've been able to buy drives that can hit over 80MB per second.

Are you so sure about that? The machine I bought in 1999 could do 12MB/s (when everything was configured optimally). The first time I ever saw a drive above 65MB/s was in 2006. Sure I have never been using really high end drives. But I doubt there could be that much difference in raw throughput.

You can also still buy SATA drives that have trouble hitting 30MB/sec. (3.5" drives. At 2.5" you can easily buy drives that have a hard time with 25MB/sec. You can also get 2.5" server drives that smoke...)

I have never seen a 3.5" SATA drive that did less than 50MB/s.

If you are losing 50% of your speed from a filesystem... Well.. You're not. It's not even worth contemplating.

It sure does depend on file system and access patterns. Sequential reads and writes of large files should come close to raw disk speed. But I have not done any carefull measuring of that.

Re:I will Settle For 1Mbps

[ivan256]

I have never seen a 3.5" SATA drive that did less than 50MB/s.

That's because many manufacturers don't even sell their 5200RPM SATA drives on the retail market. They only go to OEMs. Plus, you've said yourself that you buy larger drives. Those drives typically perform better simply due to the increased bit density. I've got a pile of Dell Inspiron's at work from 2006 with 5200 RPM SATA disks in them that will hit 35MB/sec if you're reading from the outside sectors. Generally the performance is much lower than that.

I write storage drivers for a living. I see the high performance disks as soon as, or even before they're on the market. The 15k RPM fibre channel drives I was working with in 2001 could do 80MB/sec in sequential reads. I believe that they were on the market at the time, though a non-enterprise user would not have been buying them, and enterprise users would have been abstracted from the individual drive by an array... (Plus they were $1500 each, and only 18GB. We were actually using them for the low seek times, not the high transfer rates.)

Re:I will Settle For 1Mbps

[Aram Fingal]

Actually, I was using two Macs. Mac OS 10.4 and higher can easily configure TCP/IP over FireWire and that's what I compared to the ethernet crossover cable. I used AppleShare for the file transfer and that is often criticized for using small packets. So that's consistent with what you're saying. It's the USB cable that should have had the unfair advantage because that was a target disk mode connection (and Finder copy). There have been numerous occasions where FireWire just seems to work the best but I didn't quantify it. There have also been several times that I have used GigE, with various equipment (admittedly not the best), but where a gigabit connection was indicated and failed to even get 100 Mbps. That's using various protocols including SMB and SCP.

Put off in favor of wireless.

[scaryjohn]

I humbly submit that the R&D money that could have increased the upper boundary of Ethernet speeds was spent to bring wireless to the masses. Five years ago, if you'd told me WiFi would now be a year away from nominal speeds of 250Mb/s I might have thought you were talking about prototypes. The dorms where I was a tech had just finished upgrading from 10Mb/s to 100Mb/s Ethernet. The few laptops that were sold with external wireless cards had nominal speeds of 10Mb/s. But now we have 802.11g and next year we should have 802.11n on the store shelves.

I think we've gained much more by pushing out the median speed of wireless than we could have gained from pushing out the marginal speed of twisted pair.

Re:Put off in favor of wireless.

[powerlord]

But now we have 802.11g and next year we should have 802.11n on the store shelves.

I don't know about you, but I've been enjoying 802.11n for the past few months quite happily.

The AirPort Extreme BaseStation (and Leopard) even includes the drivers to upgrade earlier MacBook/MacBook Pros that have the hardware and not the drivers.

Where's Our Terabit Ethernet?

[nebulus4]

Here

Amateurs...

[Colin Smith]

Should have thought of that when you were buying your tape libraries...

Everyone's gay for wireless nowadays

[Anonymous Coward]

Unfortunately wireless is taking the spotlight. Too bad most people can't recognize it for the pile of fail and AIDS that it is.

Oh well, once they realize it's shit then maybe we can resume work on TerE.

I think it's funny how we have a story on the front page about a breakthrough for 100GigE, and this story about TerE not being here soon enough.

Fucking Slow Down Cowboy.

rediculous or not depends on the applications

[davidwr]

It's easier to just interrupt the CPU only once when short bursts of packets come in. How short is short? 1 bit? 1 packet? 10 packets? 1M packets? 1 "transaction" at a higher-level of the protocol? The ideal answer depends on your environment. The realistic answer will be "close enough" to the ideal often enough to be a market success.

If I recall, the C64 handled serial port interrupts on the CPU. One bit equaled one interrupt. It kind of maxed out at 1200 or 2400 bps. Someone came up with the bright idea of a UART. Woo-hoo, a front-end processor for a serial port, so you only have to interrupt the processor every byte and you get parity checking and byte-border checking for free.

Come to think of it, Ethernet cards have their own front-end processors to handle the Ethernet-packet decoding and encoding, sparing the CPU.

If and when the need arises, you will see higher-level front-end processing on the network interface devices. If the need never arises, you won't see it.

Um, I have an idea...

[rickb928]

Any chance Bob's patents on Ethernet are expiring? Don't need no steenking standards, mon. Need new ways, we do! My ways! My Money! MY MONEY!

ps- all you hotshot engineers; rotsa ruck beating the real thing.

Useful?

[TerribleNews]

How useful is terabit ethernet when the pipe to the outside world is effectively choked at a couple of Mbps?

Also, am I the only person in the world who hasn't got GigE at home? I mean, again, what's the point? I could see faster ethernet being useful for university campuses and businesses who can afford giant pipes to the internet, but really, can we please get reasonable consumer grade internet access and *then* worry about the terabit ethenet?

Re:Useful?

[Thundersnatch]

There's a lot more to networking that "the Internet". Even small companies like mine (150 employess) need faster *local* networking. Our Gigabit Ethernet links often become saturated during nightly backups, and our iSCSI SAN actually uses 6 load-balanced Gigabit links to prevent from being a bottleneck (4 Gbps wasn't enough).

Though we have just 100 Mbps to the Internet (that's 1/10 our nominal LAN speed), it is only utilized at about 10% on average during business hours.

break loose...

[Hymer]

...from the stranglehold of standards and move into some fun new technologies. Yes let's do it as a complete new network... matbe a dual fiber ring topology with packets traveling in opposite directions and it should maybe be with a small, fixed packet size for faster switching.
Oh wait... that sounds familiar.

where is it???

[logicassasin]

It's in that same universe where everyone uses "... free GNU on their 200 MIPS, 64M SPARCstation-5".

(http://en.wikipedia.org/wiki/Tanenbaum-Torvalds_debate#Erroneous_predictions)

Re:Misleading understanding, "Ethernet".

[chill]

"Ethernet" is layer 2, not layer 1. It can happily run over copper or fiber or any other physical medium.

Yes, once you get beyond GbE then you are most likely going to need fiber. That means much more expensive equipment, considering I can get an 8-port GbE copper switch that supports jumbo frames for $100. You *might* be able to pump 10 GbE over copper for a few meters.

Blame Canada!

[Anonymous Coward]

Where my unlimited Internet connection that doesn't cost me an arm and a ball?

What do YOU do with your networks?

[mcrbids]

Maybe. One of the things that I've noticed is that as the bandwidth increases it becomes harder and harder to fill it up.

Really? I notice a very, very clear trend in the *other* direction.

Web pages used to be 5k simple text, maybe a pic or two. Now they're routinely a 300k flash animation doohickey - for the HOME PAGE. Once upon a time, I didn't use my computer as a replacement for television. Today, it's normal to have 2 or 3 computers in my house watching different shows (a la Youtube, etc) concurrently.

Used to be that my 1.5 Mb DSL line was just unbelievably fast. Now, 7 years later, it's routinely maxed out. I've upgraded my home network from 10 Mb to 100 Mb to 1 Gb. Part of my home network is still 100 Mb, and IT'S SLOW when I copy over an ISO image or do network backups...

What do you actually DO with your networks?

Re:What do YOU do with your networks?

[vux984]

Used to be that my 1.5 Mb DSL line was just unbelievably fast. Now, 7 years later, it's routinely maxed out. I've upgraded my home network from 10 Mb to 100 Mb to 1 Gb. Part of my home network is still 100 Mb, and IT'S SLOW when I copy over an ISO image or do network backups...

I can't fill my gigabit lan. I'd love to know how you do! Because when I transfer a 6GB VMWare machine from server to the next it should take... what 48 seconds (6GB x 8 bits per byte = 48 gigabits) over gigabit + plus some link overhead.

Yet I average maybe 80 mbps. (10MB/s) using scp, or ftp. (and even slower if using Windows filesharing/samba)

I'm not certain where the 'bottle neck' is, but I'm not getting anywhere near 1000 mbps. Now I realize I'm using consumer chipsets, and my SATA hard drives will never sustain anywhere near even half 1000 megabits/s but I was hoping to see at least 160-200mbs+ (20-25 MB/s) after upgrading to gigabit. Even if i had a RAID at both ends, I can't imagine saturating a gigabit link.

What would the point of getting an even faster link? My files aren't going to move any faster. And if my 80mbps seems strangly low ... let me know. I'd sort of assumed it was normal...

cheers.

Re:What do YOU do with your networks?

[Anonymous Coward]

Re your bandwidth woes, do a search for jumbo frames. If you (can) enable them on all your Gbps hardware, you'll get a huge speed boost.

Re:What do YOU do with your networks?

[iphayd]

Ping flood the broadcast IP address.

Re:What do YOU do with your networks?

[vux984]

Ping flood the broadcast IP address.

I'm sorry, I meant useful uses for a faster link. I guess I should have specified. ;)

Re:What do YOU do with your networks?

[iphayd]

In all honesty,

When you have a whole bunch of machines hitting a server, Gigabit is nice.
When you are running VoIP, gigabit is nice- even though the phones themselves are 100 Mbps and the protocols are tiny. It's knowing that this data goes through.
When you are pushing video. (not illegal, and not compressed)
When you are running a SAN you _must_ have gigabit for the ethernet side of the SAN.

I _have_ recalculated my hard drive speed based on a large (15GB) file copy to my SAN.

Re:What do YOU do with your networks?

[Slim Backwater]

Your transfer limit is either:

1) Lack of GigE switch between the machines, or a really really poor GigE switch (80 mbps - that sounds like Fast Ethernet)
2) Poor GigE chipset(s) on the mobo.
3) Poor CPU - rsync and scp encrypt the transfer which uses alot of cpu at each end. Also, GigE generates a fantastic number of interrupts/sec. Intel's Pro/1000 GT has an "Interrupt Moderation" feature to help with that.
4) Your hard drive speed. GigE is faster than most single hard drives. Seagate's Savvio 15k could maybe push it, with a STR of 117-97 MBps. A 10k Raptor is only 84 MB/s STR.

Run iPerf between two machines and see what your hardware can really do. (Quick tutorial, run `iperf -s` on one machine and `iperf -c {ip of other machine}` on the other. Then reverse the roles. FYI for a Lan, no IP Stack tuning is necessary)

Yes, trying to get what Gig-E promises has pissed me off too.

Re:What do YOU do with your networks?

[Da Web Guru]

What would the point of getting an even faster link?

Because the difference in price is almost trivial? Even at a retail brick and mortar store (e.g., Best Buy, Circuit City, etc.) a gigabit card (albeit low end) can be found for about $30. (Many new motherboards are coming with gigabit standard.) A 5 port gigabit switch maybe $60. (You don't even have to put everything on gigabit, maybe just your fileserver and main workstations.) With a little investment you can exceed 100mbit and remain somewhat future-proof for a while.

Re:What do YOU do with your networks?

[Fweeky]

The latest batch of consumer drives will pretty much saturate it; Seagate 7200.11's push around 105MB/s, never mind having more than one. Even WD GreenPowers can push 75M/s at 5400RPM, which is more than most GigE chipsets will push comfortably.

I'm certainly looking forward to 10GigE; I remember being annoyed at Fast Ethernet because a single one of my HD's was 4x faster than it; I'm probably only a couple of years away from that with GigE, and already well past it if you take into account the aggregate bandwidth of all the drives in my filestore.

And work? I have a cluster of machines with 25 Savvio 15k's *each*, I'd certainly appreciate being able to sync them at anywhere near the bandwidth they can push out.

Re:What do YOU do with your networks?

[Fweeky]

I get 55MB/s using samba with my GigE network, and that's mostly limited by the STR of my now long-in-the-tooth 400G drives.

That's with a quad GigE card on a PCI-X interface one end (which is basically just 4 £15 Intel 1000/Pro's on a card), and an old motherboard chipset nFarce GigE PHY (PCIe) on the other. I'm pretty sure I got similar performance out of my Intel Mac Mini, but it's been a while since I tested that.

Are you still on an old 32bit PCI system? My last desktop was an AMD K7, and it performed not unlike yours; it'd sync at 1000Mbps but it simply didn't have the bus bandwidth to even make a dent.

Terrabit?

[Anonymous Coward]

Hell, I'd be happy if Verizon could figure out how to get a 3MB pipe to my home, rather than the POS 768k connection that barely averages 300k during the evening hours.

Exponential backoff?

[Anonymous Coward]

> Five years ago, we were talking about using Terabit Ethernet in 2008.
> Metcalfe this week is starting to throw around a new date for Terabit Ethernet: 2015.

If you ask me, we should go by ethernet's CSMA/CD protocol and give terabit ethernet until 2033.

Business needs drives it

[TheAxeMaster]

I regularly download 500+MB of data just to update my link to the drawing database. Actually pulling things from it instead of my local cache requires more. The less my company has me sit and wait for drawings to load the better for them so I'm sure they'll want this. At home? Almost useless for me. I don't even have GbE and I get along just fine. But we're talking about being way past the saturation point of a SATA bus. We'll have to get better in that area first.

so what's your point?

[commodoresloat]

14 is greater than 5, so things just keep getting greater. That's progress, baby, roll with it!

Re:That's what they said about a GB/s back in the

[rjhubs]

I don't believe it is safe to assume we will continue to consume bandwidth and a constant growth rate. Back in the day, we were just sending text back and forth over the internet. But even then, we had potential other applications of the internet, like images, sound and video, but the bandwidth did not allow for it.

Now fast forward to today, we can manage to transfer video and audio feeds, sure they could still stand to be of higher quality, but there aren't many other things that we can think of right now that we would need more bandwidth for. What more could you send? Scent data? Feeling data? Holographic data?

My prediction: Once everyone has access to high quality video and audio streams bandwidth needs will taper off. Until a new technology comes along that requires vastly more bandwidth.

We had a flying car in 1979

[Anonymous Coward]

We had a flying car way back in 1979. It was orange and it's horn played "Dixie" every time it flew. It didn't fly very far, and the landings were a bit rough, but what the heck.... it *was* a flying car ;-)

Re:We had a flying car in 1979

[stubob]

That's not flying, it's falling, with style!

ExabitEthernet

[SirLanse]

Terabit Ethernet has a nice ring to it.
Petabit would be better but Exabit Ethernet has a real rhyme to it.
Maybe my great grand kids will see it.
They will need those high def hologram movies downloaded in 2 seconds right.

I just Hope they get rid of the RJ45 connectors

[PenguinJeff]

RJ45 seems like one of the worst ever. The latching pin always seems to break off. Even with the latching pin latched it can still be out far enough for the 8 pins not to have a contact in some devices. I want a screw on connector maybe like high density scsi cable connectors. And if fiber I like the ST connectors.

IPv6

[LoStMaTt]

I think we need to figure out IPv6 and how to implement it before we even consider terabyte ethernet cards.

What he really means

[mosb1000]

What he really means is that the market for terabit Ethernet is not large enough to justify major investments in the necessary research and development. That means smaller nonstandard implementations will have to be the norm, at least until there is a bigger market for it.

Uselessly small MTU?

[TaliesinWI]

Is one of the "breaking standards" going to be something like a realistic MTU? 1500 bytes is technically too small for 100BASE-T, let alone something ten or even 100 times that fast. Let's do away with the jumbo frames bullshit once and for all. IPv4 can handle a MTU of up to 65KB, IPv6 can take that to something like 4.2GB.

Quick reading: http://staff.psc.edu/mathis/MTU/index.html

1Tbase T

[MSDos-486]

While older forms of Ethernet net used Manchester, MLT and PAM5 encoding. new 10TbaseT does away with any sort of encoding or electrical signals. 1TbBase t creates a 4 black holes on each side of the wire. Only 2 pair of holes are used while the other 4 can be used to carry physical objects. 10BaseT works with a different field of quantum craziness. Each tranciever contains a array of photons that were split from other photons which are on other cards. any action on a set of these photons will effect other photons on other cards. Half duplex operation is allow only. 10BaseT also require a new MAC layer which implements MAEM(MAC address encoding messaging). A MAC address is a 16*x10e28 array of photons. when a transceiver sets it photons to 1 or 0. A array of photons will be set to 1 or 0 on all machines, everywhere.

Don't think so...

[AriesGeek]

No, it's at layer 2, not 1. Ethernet is an L2 protocol and is independent of L1 (RJ45 is an example of L1, as is Cat5e).

This is not "way down". It's 1 layer below IP.

Without a standardized transport, networks will be mass chaos.

Standards work. It's their job to work. Imagine if the electrician who wired your home didn't adhere to standards. Would you want to live there?