Another Internet2 Speed Record Broken

rdwald writes "An international team of scientists led by Caltech have set a new Internet2 speed record of 101 gigabits per second. They even helpfully converted this into one LoC/15 minutes. Lots of technical details in this press release; in addition to the obviously better network infrastructure, new TCP protocols were used."

Comments

one LoC/15 minutes

[SpaceLifeForm]

One Line of Code every 15 minutes? Seems slow to me.

Re:one LoC/15 minutes

[hummassa]

Yeah, it's my aproximate productivity in days I keep posting things in /. ... :-)

Re:one LoC/15 minutes

[fredrikj]

Seems accurate if these are lines of Perl, and they are converting them to Java before sending...

Re:one LoC/15 minutes

[macaulay805]

One Line of Code every 15 minutes? Seems slow to me.

Is this the current requirements to become an EA coder also? (No wonder they're overworked)

Re:one LoC/15 minutes

[Anonymous Coward]

No no no, you got it all wrong!

Actually, the speed record is so amazing that it causes one Loss of Consciousness every quarter of an hour, probably Slashdot readers that can't handle their adrenaline.

Re:one LoC/15 minutes

[David McBride]

Depends on whether you're transferring Perl code or not.

Gigabits...

[wittj]

The speed is 101 Gigabits per second (Gbps), not Gigabytes.

Re:Gigabits...

[shawn(at)fsu]

Right now the MPAA is trying to figure out how many movies that converts to....

Re:Gigabits...

[LSD-OBS]

Around 1.5 full length DVDs including extra features, per second. Yikes!

Re:Gigabits...

[relaxmax]

And since 1 Gigabit = 0.125 Gigabytes;

101 Gbps = 12.62 GBps

Re:Gigabits...

[dtfinch]

For most MPEG4 compressed movies it comes out to about 18 per second.

Oye

[NETHED]

Bring on the Porn comments.

But remember, never underestimate the bandwidth of a 747 full of Blueray disks.

Re:Oye

[nadadogg]

Or a nerd on a bike with a backpack full of CD-R's

Re:Oye

[KiloByte]

I'm afraid that you're _overestimating_ it.

Let's see... it's a 16h trip vs 12.6GB/s. With 25GB per disk, you'll need to take 29k disks on the plane. I would say that 30 disks weight around 1kg (it's an estimation by anal extraction, I don't have any scales at hand).

So... in order to match the bandwidth, you would need to haul around 2 tons of disks per plane. The cargo capacity of a 747 is around 30 tons (if I read the brochure I found correctly).

In other words, we already have 1/15th of that -- and the blue ray is _not_ a deployed technology yet. If you use DVDs instead, the 747 still comes ahead, but for ordinary CDs, you would need three jumbo jets doing the route...

Re:Oye

[City Jim 3000]

Cargo capacity of a 747-400 is 53000 kg and 160 m3
I assume around the same size and weight of a blueray disc as of a DVD disc which is 1.2 mm thick, 12cm in diameter and weighs a maximum of 20 grams. Also consider a plastic sleeve which maybe adds .2 mm and 3 grams.
Space needs for a disc with sleeve is thus 120x120x1.4mm = 20160 mm^3 = 0,00002016 m^3
Weight is 23 grams = 0,023 kg.
Thus:
discs/plane (volume) = 160 / 0,00002016 ~ 7936500 pcs
discs/plane (weight) = 53000 / 0,023 ~ 2304300 pcs
maximum discs per plane is then about 2300000 pcs
Blueray stores 50GB = 400 Gbit
Plane stores 400*2300000 Gbit = 920'000'000 Gbit

Not counting the time to load, burn and read the discs, a non-stop flight from Pittsburgh to LA takes around 5 hours = 18000 seconds
This amounts to 920000000/18000 =~ 51000 Gbit/sec

Considering a very approximate cost of $1/kg for the transport, and $2 for each disc it amounts to around
$4653000 total.
Which is about 0.04 $/Gbyte, or around the same price per GB as a cheap 160GB Hard drive.

Re:Oye

[City Jim 3000]

Pah.. the transport cost was not approximated based on sending it by air. Well anyway, even if the cost increases to $10/kg it doesn't change much (result is about 0.05 $/Gbyte I guess)

Re:Oye

[Anonymous Coward]

A 747 full of Blueray disks is simply a datagram (source address, destination address, payload). Bandwidth is the amount of datagrams that can be transmitted across a certain link over a duration of time.

Thus, your bandwidth calculation should be something more along the lines of a given link (say DFW to LAX). Don't forget, you can have many many many 747s in flight at any time. The bandwidth is truly staggering.

Re:Oye

[tompaulco]

That sounds like the figure for available cargo capacity in a 747 used for passenger service. The 747-400 F (cargo only) has a 124 ton cargo capacity. The proposed A380-900F cargo plane is expected to have a 148 ton cargo capacity.

Re:Oye

[tompaulco]

>The bandwidth is truly staggering. Yeah, but the latency's a bitch.

Writing speeds

[Norgus]

>. if only my HDD would write that fast!

Re:Writing speeds

[Anonymous Coward]

one wonders if my CPU would die trying to process that info alone...

Too Fast for its Own Good

[omghi2u]

Has anyone every stopped to think this might be too fast for its own good?

Isn't there a point when we've reached a speed where rather than deciding what to send from one place to another, we become lazy and start sending everything?

And won't that just lead to massive researcher mp3 swaps? :P

Re:Too Fast for its Own Good

[RAMMS EIN]

``Isn't there a point when we've reached a speed where rather than deciding what to send from one place to another, we become lazy and start sending everything?''

You mean like broadcasting radio and TV?

Re:Too Fast for its Own Good

[oexeo]

> Has anyone every stopped to think this might be too fast for its own good?

Has the infamous Bill Gates quote not taught you anything?

Re:Too Fast for its Own Good

[grid geek]

Has anyone every stopped to think this might be too fast for its own good?

No, this isn't a car, it doesn't need human intelligence after the code has been developed to keep it from turning into a wreck.

Isn't there a point when we've reached a speed where rather than deciding what to send from one place to another, we become lazy and start sending everything?

This data transfer was part of the ramp up for the start of the LHC which will be taking data at 40TB/sec, recording approx 750MB/sec data to disk - which totals about 6 Petabytes of data a year. Added to this the simulation data which may produce 12PB/year of data (based on really conservative estimates - the BaBar experiment at SLAC produced upto 6x as much simulation data as recored each year). Since all this data needs to be copied off site then no, its unlikely we could get the network to a point where everything could be sent - processing still increases faster than network bandwidth.

And won't that just lead to massive researcher mp3 swaps?

Not if they want tenure ...

Re:Too Fast for its Own Good

[LSD-OBS]

The Swedish LHC (CERN) guys are going to need to be sending *petabytes* of collision data around the world for analysis over the course of their experiments. It's crazy to think it, but some people really do have a need for this. I suppose this is why the Internet2 is largely restricted to research and education purposes.

Re:Too Fast for its Own Good

[LSD-OBS]

which totals about 6 Petabytes of data a year

At 100Gbps that's still 5 months per 1Pb. These guys sure are demanding!

Re:Too Fast for its Own Good

[Anonymous Coward]

This poster and others have asked if this is such a great idea to be able to send this much data at such speeds. This, however, is dealing with Internet2, which is used exculsively for universities and research institutions sending data back and forth. It might seem like a lot if you're talking e-mail, pr0n, and other normal Internet traffic, but not so much if you're sending real-time high-def images to be used by a remote robotic surgeon or data for weather modeling.

Re:Too Fast for its Own Good

[lisaparratt]

Yes. Yes, it is. It's far too fast for it's own good. Also, it's a good thing we don't travel about 30mph, otherwise our skeletons would turn to jelly! Oh, wait a minute...

Re:Too Fast for its Own Good

[MyLongNickName]

Does your surgeon really need these speeds? Or does he need a connection with near-zero latency, and very consistent near-zero latency?

Real-Time Video can be acheived with today's bandwidth (albeit expensive solutions). What we can't deal with is a five second hiccup (sorry about cutting all the way through you ma'am... lag is horrible today).

Does Internet 2 solve the latency issue?

Re:Too Fast for its Own Good

[vidnet]

For a couple of million users, this is around dialup. You'd need a whole bunch of these if you wanted to give a small country TV on demand.

Re:Too Fast for its Own Good

[Meostro]

Am I missing something, or did you just make that up?

Please correct me if these are wrong or feel free to use 1024s in place of the 1000s, it won't make a difference.

8b = 1B
* 1000 ^ 3 = 1GB
* 1000 = 1TB
* 1000 = 1PB

100Gbps / 8bpB = 12.5GBps

12.5GBps * 60s/m = 750GBpm

750GBpm * 60m/h = 45TBph

45 TBph * 24h/d = 1.08PBpd

So it's about 6 days for 6PB of data, or about 18 hours if it's Pb instead.

Re:Too Fast for its Own Good

[EkkiEkkiShiwaddle]

And won't that just lead to massive researcher mp3 swaps? :P

Ok, someone has to say it...

Won't somebody please think of the RIAA???

Re:Too Fast for its Own Good

[LSD-OBS]

Hmm, remind me not to take order of operators for granted when using MS calculator. Silly piece of crap software (>_)

Re:Too Fast for its Own Good

[elem]

Its not speed for its own sake.

One of the major reasons why this is important (and higher bandwidth is good) is for scientific data. Things like data from CERN and other particle accelerator. These produce -huge- amounts of data every time they run and this will allow researchers to be able to access this data without actually having to go to places like CERN.

Re:Too Fast for its Own Good

[Anonymous Coward]

Right now, this is only really meant for particle accelerator and other mass data producing scientific tools.

Right now, the only other alternative is to go to the particle accelerator. It gets expensive flying to Geneva just to reanalyze some data.

Re:Too Fast for its Own Good

[Charcharodon]

Duh! Until it's so fast that it gets there before I even decide to send it then it's not fast enough!

Give me a terabit a second transfer rate and a year and I'll show you someone who is sick and tired of the damn wait for things to download/upload.

It's all what you are used to.

Re:Too Fast for its Own Good

[LordOfYourPants]

When people jumped from 56k to 1Mbps, the only thing that really changed for the *average joe* was grabbing mp3s and checking out more trailers.

Contrary to popular belief, most people are not out there downloading a 9GB collection of Friends, season 1 or grabbing a 20GB MAME set with flyers and cabinets. Most people will just go buy the DVD or grab Midway Arcade Treasures and be happy.

When people jumped from 1Mbps to 5Mbps, I've seen them take advantage of it by shopping on amazon 2ms faster than before.

I think the real "danger" with higher speeds would lie in the realm of more annoying/higher def advertising. When the day comes that it becomes trivial and technically possible on a large amount of computers to download and display a 1920x1080 30 sec interstitial ad before you can view a webpage, it *will* be done.

You can already see this transition happening with lower res video as people try to pack a highly-compressed 30 second FMV ad into a flash box.

Re:Too Fast for its Own Good

[Dr. Kinbote]

> --Last yeer I kudn't spel Injunear...Now I are
> one!

Try spending the next year on learning the
difference between Switzerland and Sweden.

Re:Too Fast for its Own Good

[embo]

Just food for thought, this isn't so much about speed as it is about size (and you all thought that didn't matter).

Think about it this way. If you have a 1" pipe, and you send a little bit of water down it, the water reaches from one end to the other in a certain amount of time. Now take that up to a 4" pipe. Does the water travel any faster simply because it's a larger pipe? No. But the difference is that you can send MORE water in the same amount of time, not that you can send it from one end to the other any faster. (And don't bore me with stats about friction or using a pump, and if the pipe is at capacity, etc. I know this. I'm trying to make a general point here.)

By the same token, I would bet that ping times on a T1 are going to be relatively the same as they will be on an OC-192, assuming the same number of hops, etc. You just can't send as much data through the T1 as you can the OC-192.

This isn't about how fast the connection goes, as much as it is about how much we can push through that particular connection. This matters greatly as backbone routers expand and have to push more and more traffic. And yes, I know that when pipes get full, things DO get slower because the pipe is full. They do go hand in hand, but they're not exactly the same either.

When my ping times between Pittsburgh and LA are in the 1 - 5ms range, then I'll believe it's faster, instead of just bigger.

This really should be titled "Internet2 CAPACITY Record Broken".

Re:Too Fast for its Own Good

[grid geek]

No, lets stick to base 10 (which is what disk manufacturers tend to use anyway).

1 GB = 1,000,000,000 Bytes

1TB = 1000 GB

1PB = 1000 TB

650MB/s (ATLAS/CMS/LHCb/ALICE detectors) * 60 sec * 60 mins * 24 hours * approx 100 days = 5,616,000,000MB = 5.6PB / year

Unfortunately the LHC needs a nuclear power station (and a hydroelectric one for the computers 8)). I'm not joking, it really does. Running for more than 100 days a year + 30 for some other stuff is not practical unless you want Geneva to freeze during the winter. To put this in perspective the Computing farm is going to use 7 MegaWatts of power and is classed as a small power user

Re:Too Fast for its Own Good

[LSD-OBS]

I know the difference. I just couldn't remember where Geneva was. After a nap I now remember it lies across the borders of both France and Switzerland. I'd be ashamed if it weren't for the weekend I just had.

Re:Too Fast for its Own Good

[dpuu]

I remember watching a lecture on the reseach channel, where a comparison was made of growth rates of different technologies: cpu, storage and network bandwidth. The bottom line was that cpu performance growth follows Moores Law (i.e. the perf increase is dominated by manufacturing issues) while network performance is increasing at 10X the cpu rate (disk is somewhere between). The talk discussed the implications of this.

The summary was that we'd need to revisit system tradeoffs. We currently compress data on the cpu, because the bottleneck is the network bandwidth. But given the rate of growth of the different components of the system, the bottleneck is likely to move from the network to the cpu (or storage). At that point, we increase system performance by sending raw data and eliminating the CPU compression stage.

Things get more complex again when you look at offloading various tasks from CPU onto dedicated hardware. But my basic point is that performance of the network shouldn't be considered in isolation. If we can improve performance by "being lazy and sending everything", then where's the harm?

Re:Too Fast for its Own Good

[FrYGuY101]

Have the throngs of people correcting you that Bill Gates didn't actually say it taught you nothing?

Re:Too Fast for its Own Good

[oexeo]

Have the throngs of people correcting you that Bill Gates didn't actually say it taught you nothing?

That's not important, it's still has a valuable moral behind, regardless of it's authenticity

Re:Too Fast for its Own Good

[Anonymous Coward]

That's not important, it's still has a valuable moral behind, regardless of it's authenticity

It is important when it's attributed to him. He didn't dispute the lesson imparted by the quote, he disputed the attribution of the quote to somebody who didn't say it.

Simply dismissing a factual error as 'not important' because other aspects of it have value is to demean the facts.

Re:Too Fast for its Own Good

[Anonymous Coward]

Very valid points, just one nit though: 5ms from Pittsburgh to LA would probably require re-writing the laws of physics. Even at the speed of light in a vacuum (c~186Kmiles/sec, and bear in mind light in fiber is 0.6c), Pittsburg to LA is about 2145 miles (air distance, see http://www.losangelesalmanac.com/topics/Transport/ tr53.htm) so the time for light to get there and back (RTT) is 23msec. So short of inventing tachyons that transmit information faster than c you will have a long wait to get to 5ms ping times.

Re:LoC?

[RAMMS EIN]

Perl, obviously. Ever seen whole one line programs in any other language?

** Spoiler **

[Stokey]

Cue the gags about "Finally, I shall be able to download my pr0n collection".

Cue questions about whether is gigabytes or gigabits.

Cue questions about "How can I get a such gaping-a$$ bandwidth.

One of these days I will write the ultimate FAQ to /. posts including all the possible combinations of arguments started by SCO stories, how politics is treated here and a whole chapter on non-funny memes.

Go! Pedal faster.

Re:** Spoiler **

[Anonymous Coward]

I for one welcome our new FAQ-writing overlord.

Re:** Spoiler **

[Anonymous Coward]

In Soviet Russia, those imagine a beowulf cluster of you!

Re:** Spoiler **

[David McBride]

Said ultimate FAQ already exists:

http://en.wikipedia.org/wiki/Slashdot_subculture

Feel free to add to it.

Re:** Spoiler **

[Anonymous Coward]

More like que the "Que the jokes about..." Thanks for the obvious notice.

747

[Fullaxx]

or the cost ;)

Re:747

[kfg]

or the cost ;)

Never overestimate the cost per bit of a 747 full of blueray disks.

KFG

Doesn't make sense

[oexeo]

new TCP protocols were used

TCP is a specific protocol, a "new" TCP protocol would suggest a different protocol, unless it means a revision of the current protocol.

Re:Doesn't make sense

[TopShelf]

Isn't "TCP protocol" redundant anyway?

Re:Doesn't make sense

[PhrostyMcByte]

They are talking about "Fast" TCP, which AFAIK just consists of a better routing algorithm and using multiple TCP streams at once.

Re:Doesn't make sense

[ari_j]

Aren't there other concerns, such as window size? When you're piping that much down the line, an ACK every 48 bytes just isn't right.

Re:Doesn't make sense

[oexeo]

> Isn't "TCP protocol" redundant anyway?

I think acrynums where invented for the hilarity of watching stupid people prepend/append the abbriviated words to them (yes, just like I did above).

Re:Doesn't make sense

[MikeBabcock]

Insightful?

What is HTTP? Oh yes, a TCP protocol. For proper semantics, you might say its a protocol that sits on top of TCP, or go into network layers and bore the reader to tears.

On the other hand, lets say we're talking about a new revision of the TCP protocol ... TCP version 42 ... wouldn't that be a new TCP protocol as well?

I'm surprised you can't figure out what that would mean.

Re:Doesn't make sense

[The-Bus]

No, this is a TCP Protocol for Transfer.

DUH.

...Oh.

Re:Doesn't make sense

[oexeo]

where = were

Re:Doesn't make sense

[oexeo]

What is HTTP? Oh yes, a TCP protocol.

It's an application protocol, that functions over TCP, you can't really say it's a TCP protocol.

On the other hand, lets say we're talking about a new revision of the TCP protocol ... TCP version 42 ... wouldn't that be a new TCP protocol as well?

Yes, it could be, like I said in my post.

Re:Doesn't make sense

[meethade]

I think what was meant by "new" was a more recent revision of TCP along it's implementation.

Re:Doesn't make sense

[RichN]

acrynums = acronyms
abbriviated = abbreviated

Re:Doesn't make sense

[CProgrammer98]

He meant that the statement was redundant as in "Transfer Control Protocol Protocol" (in the same way that people say PIN number (Personal Identification Number Number)

Re:Doesn't make sense

[LabRat404]

if ((TCP->streams > 1) && (pr0n.streams > 1)) { do { pluggin.bearshare.download.randomPr0n(); }while (1); } well, damnit, its about time someone said something about porn instead of "say something about porn!"

Re:Doesn't make sense

[justforaday]

Yes, but if it's using TCP/IP the two Ps cancel each other out, so you'd need to append 'protocol' to that, right?

Or at least that's what I remember from algebra...

Re:Doesn't make sense

[LabRat404]

now, why is it that I lose all my formatting when I submit my posts? sorry about that :-p

Re:Doesn't make sense

[Anonymous Coward]

Posts like the parent make me wish there was an "incorrect" moderation option. Routing takes place at layer 3; TCP is used at layer 4. You don't 'route' TCP. Ergo the description above is incorrect and does not warrent a +5. Though I'm sure the +5 is for nothing more then the link.

Re:Doesn't make sense

[Mondrames]

My friend calls this an "Extronym" - when you append the word for wich the last initial stands to the acronym. For example - ATM Machine or PIN Number. She's be using it for a couple of years now.

Re:Doesn't make sense

[Anonymous Coward]

Isn't "TCP protocol" redundant anyway?

I use the TCP Protocol over my NIC Card

Re:Doesn't make sense

[Fatchap]

There are several different forms of the TCP protocol. Most of them deal with the way that congestion is managed or deal with networks that either have very high or very low latency or reliability.

They are "new" in that they have not been around for very long and people are still researching them. They are not really revisions thought, it is not like there will be a TCP 2.5 specification in the future. The individual protocol used will be different for different links.

Re:Doesn't make sense

[Anonymous Coward]

It's a modification of the scaling/congestion alogorithms in TCP/IP:
http://netlab.caltech.edu/FAST/index.html

You can download it and try it yourself.

Enjoy

Re:Doesn't make sense

[Tribbin]

TCP/IP = TC/I

TCI, which stands for Telecommunications, Incorporated, was a cable television provider in the United States.

The company came into being in 1968, following the merger of Western Microwave, Inc. and Community Television, Inc.

After going public in 1970, the company grew rapidly, and became the top cable provider in the United States. It was purchased by AT&T in 1999.

Re:Doesn't make sense

[turm]

I've recently heard it referred to as "RAS Syndrome" or, Redundant Acronym Syndrome Syndrome.

Re:Doesn't make sense

[Anonymous Coward]

Great, now I have a word to explain it. If your wondering I usually do this one and always get attacked for it.

RPG Game.

Re:Doesn't make sense

[Y2]

They are talking about "Fast" TCP, which AFAIK just consists of a better routing algorithm and using multiple TCP streams at once.

No, FAST has nothing to do with routing or multiple streams.

(Users of FAST may also be users of multiple streams for data transfer, as in GridFTP.)

FAST (follow the link in the parent) is one of many approaches to improving the throughput of a stream by changing the algorithms and timing for packet transmission.

Re:Doesn't make sense

[tompaulco]

Is MP3 an example of an acrynum?

Re:Doesn't make sense

[tompaulco]

No point in buying in now. I'm going to wait for FAST-WIDE ULTRA TCP-2.

Re:Doesn't make sense

[MikeBabcock]

"Its an application protocol that functions over TCP" has the same semantic function as "Its a TCP based protocol" or even "Its a TCP protocol" although the latter can have other meanings as well.

My point was that English semantics do leave a lot of room for what someone "means" (thus deconstructionism).

Library of Congress?

[NardofDoom]

Is that with or without the pictures?

Re:Library of Congress?

[KronicD]

Hmm so, 12.625 * 60 * 15 = 11362.5GBytes....

Damn, that would be one big torrent... Oh how I wish it were available :(

Best read this way....

[Himring]

Best read using Christopher Lloyd's voice from Back to The Future, e.g.:

"101 jigowatts per second!!!" --Professor Emmett Brown

Re:Best read this way....

[Anonymous Coward]

"101 jigowatts per second!!!" --Professor Emmett Brown

Uh, yeah, that's pretty much, uh, completely wrong. If you are quoting someone, it works best if you include at least one word that was in the original quote.

Re:Best read this way....

[humbads]

That should be, "One point twenty-one Jigawatts!, One point twenty-one Jigawatts!." -- Professor Emmett Brown, with his hands flying in the air. (just saw the movie again last night)

Gigabits

[Macfox]

Jeez... When you're talking about new world records you think you'd stop to double check those IMPORTANT facts which appear in the first paragraph. :)

But then again this is slashdot.

Bytes'n'Bits

[mx.2000]

speed record of 101 gigabytes per second.

Wait, isn't this supposed to be a nerdy tech magazine?

I mean, I except this kind of Bit/Byte confusion on CNN, but on slashdot...

Re:Bytes'n'Bits

[Anonymous Coward]

Maybe you should accept the fact that people make mistakes occasionally.

Re:Bytes'n'Bits

[Anonymous Coward]

like me accepting things like accept/except on /.

Re:Bytes'n'Bits

[spectrokid]

I would accept this accept/except confusion on Slashdot, but not on CNN!

Re:Bytes'n'Bits

[magefile]

I think you mean that you would "except this confusion on /.".

Sustained transfer?

[Anonymous Coward]

How did they sustain a transfer like that? Unless my math is wrong, that's 11GBps ... what has that kind of read/write speed?

Re:Sustained transfer?

[Chatsubo]

/dev/null

Re:Sustained transfer?

[RAMMS EIN]

``what has that kind of read/write speed?''

The network, obviously. And that's the only part that needs it - they don't need to be sending useful data.

Re:Sustained transfer?

[Anonymous Coward]

They sent a 1-bit file and multiplied.

Re:Sustained transfer?

[noselasd]

Uh, they are measuring network speed, not harddrive/etc. speed.

Re:Sustained transfer?

[a3217055]

The way to write 11GBps is to use a distributed array of disks. A parllel filesystem can easily handle it. Over a 100 networked computers with a parallel filesystem like Lustre, GPFs or PVFS( 1,2 and 3 .... is there a 3 ?) can do it. I mean there are disk arrays that have sustained throughput of over 55GBps. Also the 11GBps that we see now may one day used for having all sorts of communication going through it so in a way it is a way of the future.

Re:Sustained transfer?

[Anonymous Coward]

a triple raid 5 array of 15 scsi U320 drives across 3 controllers... basically a raid 5 of a raid 5 array.

Hell we have a burst speed of 20Gbps with that kind of setup.

Re:Sustained transfer?

[Idarubicin]

... what has that kind of read/write speed?

For the first time, a comment that starts "Imagine a Beowulf cluster..." might actually be on topic.

More seriously, the Internet2 is designed for transferring massive scientific data sets between research institutions. The folks at CERN are planning to run experiments that generate terabytes of data per second. They're no doubt going to be using buckets of RAM and monster arrays of drives operating in parallel to keep on top of that. They wouldn't be developing these big pipes of they didn't think they could fill them in the next few years.

Re:Sustained transfer?

[Octagon Most]

"How did they sustain a transfer like that? Unless my math is wrong, that's 11GBps ... what has that kind of read/write speed?"

Good point, but that's the aggregate throughput of the data pipe and not necessarily generated or used by any two single end-point devices. They may test it this way as a proof of concept, but it's more likely that 1000 computers in a lab on one coast would send that total data through such a link to a lab on the other coast.

Re:Sustained transfer?

[m50d]

A 4-drive raid0 array on a PCI-X card.

Re:Sustained transfer?

[Falrick]

As this was an experiment, it is likely that they merely sent pseudo-random data. Probably even just the same blocks of data repeatedly. You don't need a large data set to generate traffic for testing purposes, just something that is easily verifiable.

Re:Sustained transfer?

[Anonymous Coward]

Actually, just transferring raw scientific data (like you might get from a particle accelerator) could suck up all that bandwidth. Which was kinda the whole motivation for Internet 2, it was a big pain to download dumps from CERN. More familiar uses would be for transmitting X-rays and MRIs, which are sampled at very high resolution for obvious reasons, so are also pretty dang huge. So you probably wouldn't necessarily involve a whole lab's workload (which wouldn't tend to peak that high that consistently, anyway, I'm thinking), but you'd need a serious storage array backing it up.

Porn

[LifesizeKenDoll]

You could send so much porn at 101 Gbps...*drool*

Re:Porn

[Anonymous Coward]

hey! now we can just download the actresses instead of just their movie!!
omgah! I can't wait...
hubababababa dr000l.. where do I sign up for this new tech?.

My Car Gets Forty Rod to the Hogsgead

[The-Bus]

Yeah, I'm not really sure what the Library of Congress unit does for me. I'm more used to the European metric measurement of Geburninged Volkswagen.

Nowhere in the article does it say how long they ran the test for. A second? A minute? An hour?

I mean that's a full terabyte almost every minute and a half. What has so much data?

Re:My Car Gets Forty Rod to the Hogsgead

[SlayerofGods]

I mean that's a full terabyte almost every minute and a half. What has so much data?

The library of congress perhaps?

Re:My Car Gets Forty Rod to the Hogsgead

[MikeBabcock]

tcpclient www.slashdot.org 80 cat /dev/urandom

Jeez

Slashdot sucks ;-)

Re:My Car Gets Forty Rod to the Hogsgead

[pjf(at)gna.org]

> What has so much data?

/dev/zero ;P ?

Re:My Car Gets Forty Rod to the Hogsgead

[Binestar]

> What has so much data?

/dev/zero ;P ?

Yeah, but that compresses pretty easy.

Re:My Car Gets Forty Rod to the Hogsgead

[Fortran IV]

from the this-is-Slashdot-so-why-remember-the-past dept.

Wal-mart, remember?

Re:My Car Gets Forty Rod to the Hogsgead

[jahill_isu]

I mean that's a full terabyte almost every minute and a half. What has so much data?

Internet 2 as well as the Engineering Science Network are being setup to move massive amounts of data from science labs to computational labs within the US Department of Energy.
Labs like Fermilab are expecting to produce hundreds of terabytes per day of data from research when they come online.

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

Well in that case /dev/random

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

Longhorn source code ?

Re:My Car Gets Forty Rod to the Hogsgead

[perdu]

Labs like Fermilab are expecting to produce hundreds of terabytes per day of data from research

I'm sure the NCBI, EBI and other genomics institues could use some of this for genome, gene expression and variation data and such...

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

Here's a link http://computing.fnal.gov/news/chictribune090203.h tml

CERN Conseil Europeen pour le Recherche Nucleaire (European Laboratory for Particle Physics)

Important Point:
When CERN comes online in about 5 years, it's expected to churn out petabytes of data. Yeah. I meant that. Petabytes, as in 1024 terabytes. Fermilab is already turning out terabytes but it will be surpassed greatly by CERN.

A particle accelerator is basically taking very high resolution images in 3 dimensions hundreds of times a second. It's pretty easy to see how so much data is accumulated.

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

a pseudo random-number generator.

Re:My Car Gets Forty Rod to the Hogsgead

[Jorkapp]

A few DVD's perhaps?

There is a disturbance in the force. As if millions of MPAA employees suddenly cried out in terror, then were suddenly silenced.

Re:My Car Gets Forty Rod to the Hogsgead

[bathmann]

warez? porn? errrrrrrrrrr, not that there is any difference between those 2.

Eventually, there is always a use for so much bandwidth. Just run a ftp-server on the box with anon login et voilà.

Re:My Car Gets Forty Rod to the Hogsgead

[Yottabyte84]

/dev/urandom. /dev/random will run out of entropy pretty fast.

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

> What has so much data?

Spam.

Re:My Car Gets Forty Rod to the Hogsgead

[Thuktun]

I mean that's a full terabyte almost every minute and a half. What has so much data?

Full-immersion remote presence, perhaps? Common use of audio/video emails? Use your imagination.

I would imagine most people 20 years ago wouldn't be able to conceive how one would use hundred-gigabyte hard drives and multi-gigabyte memory cards.

Re:My Car Gets Forty Rod to the Hogsgead

[Decimal]

What has so much data?

I'm sure we could overload it with a good Slashdotting.

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

My porn directory.

C:\pron\

Re:My Car Gets Forty Rod to the Hogsgead

[Anonymous Coward]

Wal-Mart has 460 terabytes of consumer-tracking goodness stored on mainframes at its headquarters: http://developers.slashdot.org/article.pl?sid=04/1 1/14/2057228&tid=187&tid=221&tid=198&tid=1

Porn

[TheKidWho]

Here comes new super high quality porn on the internet!

Alternative High Bandwidth Solution

[isa-kuruption]

A station wagon hauling backup tapes. Too bad the latency is so high!

Re:Gigabits!

[compro01]

still, thats 12.625 GBps. it's still plenty fast.

that's my entire hard drive moved in 10 seconds.

Re:LoC?

[slapout]

Actually yes. Back in the day when Basic allowed multiple commands per line with the colon (:) seperator. Oh wait, you meant useful programs...nevermind.

15 LoCs?

[Chemisor]

15 Libraries of Congress in 15 seconds? Great! Anybody got a copy?

I can beat that!

[Anonymous Coward]

I can transfer one and a half terabits from one end of the room to the other in less than a second in two easy steps.

Step 1. Fill 200MB hard drive with data
Step 2. Fling aforementioned hard drive in a frizbee'esque motion across the room.

Expect some data loss however.

Take that Caltech!

Re:I can beat that!

[tokenhillbilly]

That's the problem with unicast protocols. What you need to do is recruit a friend an have them stand on the other side of the room. Their job is to catch the hard drive. This protocol addition to your network will greatly reduce the chance of data loss errors.

Also, you might consider a 300GB drive to increase network throughput.

Re:I can beat that!

[relaxmax]

Data loss can be further reduced by having a 'catcher' wearing a customised outfit, so as to minimise the sudden impact on the disk. Something like a sumo wrestler wearing baseball gloves

Re:I can beat that!

[Anonymous Coward]

1.5Tb != 200MB

Re:I can beat that!

[zx75]

Wow, you can fill 200MB hard drive with data in less than a second?

People comment about the bandwidth of a card full of DVDs or Tape Drives and the like, but do they ever stop to think about exactly how LONG it takes to write information to the medium? Driving from one place to another with the data is trivial, but converting the raw data into the transportable message takes absolutely forever.

Re:I can beat that!

[LucidBeast]

This might be bit degratory, but I've heard that in England they toss midgets (some sort of bar game) and surely the information content of a midget is much more than 200MB. So the Brits have transfered information in their drunken stupor for centuries faster than these dudes.

If anybody shorter than me (5'11") is offended by midget tossing, blame the Brits not me.

Re:I can beat that!

[Anonymous Coward]

Correct you are, clearly you've mastered the intricacies surrounding logic, though you've yet to breach that pesky "humor" problem.

Good luck.

Re:I can beat that!

[Anonymous Coward]

Hm? Bandwidth refers to specifically one thing, how much data you can stuff into a transport mechanism.

Converting data into the format appropriate for transport isn't part of the calculation, the time it takes to do such is generally calculated as part of the latency.

Re:I can beat that!

[magefile]

Blame the Brits? Unfortunately, this bar game is around in the States, too. The problem is not just that most dwarfs (we call ourselves "dwarfs", or "short-statured") see this as degrading, but that it is dangerous. Not that we're particularly worried about the dwarfs that subject themselves to this - they are probably aware of the risks, even if they are ignoring them - but the fact that this is seen as "acceptable" creates a danger for Joe Dwarf walking down the street, in that some day, some drunk asshole will think, "oh, hey, there's one of them guys I can throw!" and try it.

Re:I can beat that!

[frankvl]

1 byte != 7500 bits

Re:I can beat that!

[dogmatixpsych]

Don't you mean 200GB? I do remember my first 200MB hard drive though (actually 240). Man, it was hard to fill!

Re:I can beat that!

[tompaulco]

We don't hear about how long it took to load up the data for the 101 Gb transfer either.

Re:I can beat that!

[Anonymous Coward]

I think a 200GB hard drive might help with getting the appropriate amount of data across the room.

They could get better speed

[CastrTroy]

They could probably get much better speeds if they compressed it first. The Library of Congress is quite compressible, as there is a lot of redundant data. Text in general is known to be quite compressible.

Here's a question. Sure, you can send 101 Gigabits per second. But what kind of power do you need on either end to send or interpret that much data? I know my hard drive doesn't go that fast. I don't even think my RAM is that fast.

Re:They could get better speed

[CProgrammer98]

>>The Library of Congress is quite compressible

I do hope you mean that the data content within the library is compressible... The building itself is quit tough to compress!

Re:They could get better speed

[Anonymous Coward]

I doubt you could compress at 101Gbps. Therefore, it'll take less time to just transfer it than to compress->transfer->uncompress.

Re:They could get better speed

[DerPflanz]

I know my hard drive doesn't go that fast. I don't even think my RAM is that fast.

They are testing a network, not harddrives or RAM (as said in other posts as well). Consider the incredible amount of data generated by big clusters or machines used at CERN for example. You generate the data in Switzerland, and move it to your friendly colleague in the USA, where his supercomputer will render some nice images from it. And, I am not a supercomputer-guru or anything (haven't even seen one for that matter), but I can tell you, they are nothing like your PC at home.

And that is only one usage of it. Consider the internet connection between Amsterdam and New York. That are a lot of connections together, all in one stream. If that could be a 101Gbit/s line, it will make a lot of connections faster, think of real-time video(phone) streams, normal phone connections, etc.

Re:They could get better speed

[McWilde]

You'll find you cannot send the building over a network in 15 minutes either.

Re:They could get better speed

[SphereII]

It is compressable... you just can't uncompress it without some data loss.

Re:They could get better speed

[Anonymous Coward]

The building itself is easy to compress but unfortunately it's lossy.

Re:They could get better speed

[Detritus]

Often times it isn't worth the effort to compress data, especially when your network bandwidth greatly exceeds the rate at which your system can compress the data.

Re:They could get better speed

[CastrTroy]

That was kind of my point. It would be nice to have this kind of connection running underneath the ocean to link continents. However. We first have to have a line under the ocean that can handle that kind of bitrate. We also need a computer (or a cluster of computers) on either end that is capable of dealing with all this data, parsing the packets, routing it, and doing anything else that may be necessary.

Re:They could get better speed

[dogmatixpsych]

They just need this on either end controlling the network. It would then be a sentinent species as well.

Re:They could get better speed

[Anonymous Coward]

i duno, the WTC compressed pretty nicely..

Re:They could get better speed

[DerPflanz]

Well, what I understand is that such machines exist already. Terabit switches do exist and are used as well. They connect several highspeed (gbit) lines to eachother and route everything accordingly.

This record is not about high speed on short distances (as used in terabit switches) but about long distance sustained rates. It is so to speak the 'head of the research'. More useful systems will emerge from technologies used here.

Why still TCP , what about SCTP?

[Viol8]

SCTP was specifically devised as a replacement for TCP as it can emulate the 1 -> 1 connection of TCP but can do connection based 1 -> N too. I thought it has been designed with high speed in mind too. Does anyone know whether this protocol is being used more and more or has it just become another good-idea-at-the-time that got run over by the backwards compatability steamroller?

Is it needed?

[Kombat]

This is great and all, but has anyone stopped to ask why we need such fast networks? The stock-frenzy driven surplus of unneeded bandwidth was a major contributing factor to the dot-com bust. I remember when I was working on a multi-gigabit, next-generation optical switch, and the project manager was assuring us that in just a few years, people would be downloading their movies from Blockbuster instead of actually traveling there to pick up a DVD. We were all supposed to be videoconferencing left and right by now, with holographic communications just around the corner. A massive growth in online gaming was supposed to cripple the existing legacy networks, forcing providers to upgrade or perish. All of this was supposed to generate a huge demand for bandwidth, which were were poised to deliver.

Well, as we all know, that demand never materialized. We had way more bandwidth than the market needed, and when the bandwidth finally became stressed, providers opted to cap bandwidth and push less-intensive services rather than pay for expensive upgrades to their infrastructures.

I think we should instead be focusing on technologies that can a) generate real new revenue to the providers that we're trying to sell these ultra-fast networks to, b) have obvious and legitimate research or quality of life improvements, and c) are sure-fire hits to attract consumer attention (and $$$).

Don't get me wrong, this is very cool and all, but until Netflix actually lives up to its moniker and sends me my rented movies through my phone/cable line rather than UPS, then it doesn't really matter to me if the network is capable of 5 Gbps or 500 Gbps. Slashdot will still load in a second or 2 either way. We need real products to take advantage of this massive bandwidth, and that revenue will drive research even further, faster. I fear we're going to stall out unless we find a way to embrace these faster networks and make money off of them.

Re:Is it needed?

[Anonymous Coward]

We need such fast networks so we can parcel the total bandwidth out amoung many consumers.

Re:Is it needed?

[Kombat]

We need such fast networks so we can parcel the total bandwidth out amoung many consumers.

The existing network is more than adequate for that goal. You missed my point entirely.

Re:Is it needed?

[HarveyBirdman]

We were all supposed to be videoconferencing left and right by now, with holographic communications just around the corner.

These are the same sanke-oil futurists that once told us we'd have flying cars, fully automated homes, vacations in space, sexbots and televised death sports.

OK, maybe only Norman Jewison predicted televised death sports, but you get my point. They would righteously rock, though. Especially watching televised death sports while fucking my sexbot in my flying car.

Re:Is it needed?

[Anonymous Coward]

Internet 2 is for use by scientists not this 'we' that you mention. it is exclusive to them. YOU dont need that bandwidth, THEY do. YOU dont generate terrabytes of data, THEY do. I dont think existing networks are capable of efficiently transfering this information... keyword EFFICIENTLY.

Re:Is it needed?

[Anonymous Coward]

From your original post...

"We had way more bandwidth than the market needed, and when the bandwidth finally became stressed, providers opted to cap bandwidth and push less-intensive services rather than pay for expensive upgrades to their infrastructures."

And from your last post...

"The existing network is more than adequate for that goal"

Both cannot be true. I think demand will always eventually fill the available bandwidth of the network, and it is in the interests of the ISPs to regulate that use.

When we do get films on demand, the ISP will need enough bandwidth to accomodate a thousand other users doing the same thing, so they need a much faster network, *and* to cap the usage of individual customers.

More bandwidth will allow films on demand, but it will not change the structure of the network.

Re:Is it needed?

[rsmith-mac]

Just because "the future" isn't happening in N. America yet, doesn't mean that it isn't happening elsewhere. N. America is constrained by its last mile problem, but Asian nations like S. Korea and Japan don't suffer this, which is why they already have multi-megabit fiber drops to homes and businesses. Sure, we on our miniscule ADSL and Cable hookups may not see the need for such massive bandwidth since we can't use it, but when you have a 1000 unit apartment complex with 100Mb fiber drops, this kind of intense technology is necessary.

The use for it exists, it just hasn't reached us yet.

Re:Is it needed?

[Coolmoe]

Well many of the things on your list would sell if the infrastructure was there. Do you think cars would sell as well if only 10% of the country covered by roads? I doubt it.

The problem is that cheap high speed net access needs to be the norm for many of these services to fly. Without the infrastructure your service will not sell. That is what led to the dot bomb was rolling out services that could not support themselves with existing infrastructure and bad marketing. As soon as everyone has a big pipe to thier house at a reasonable price many of these services will actually take off.

Re:Is it needed?

[Anonymous Coward]

You are bumping heads with a chicken/egg problem. In the entertainment industry, the supply generates the demand, and without the infrastructure, the supply doesn't exist.

Color TV had similar headaches. Why broadcast in color when nobody owns a color TV? Why buy a color TV if nobody broadcasts in color?

Anyways, we're getting close to downloading movies from Blockbuster. Charter's OnDemand service is functionally similar, but still has bandwidth problems. Extra bandwidth will be useful.

Re:Is it needed?

[Anonymous Coward]

Did you RTFA? If you had, you would know that these ultra-fast networks are needed for the next-generation of high-energy physics experiments. These experiments are going to generate a high amount of data (petabytes per year) and the analysis is to be distributed to labs all over the globe. An ultra-fast network infrastructure is required to transmit the data.

Future optical networks, incorporating multiple 10 Gbps links, are the foundation of the grid system that will drive the scientific discoveries. A "hybrid" network integrating both traditional switching and routing of packets, and dynamically constructed optical paths to support the largest data flows, is a central part of the near-term future vision that the scientific community has adopted to meet the challenges of data intensive science in many fields. By demonstrating that many 10 Gbps wavelengths can be used efficiently over continental and transoceanic distances (often in both directions simultaneously), the high-energy physics team showed that this vision of a worldwide dynamic grid supporting many-terabyte and larger data transactions is practical.

Re:Is it needed?

[soliptic]

Way to get +5 insightful despite not displaying the slightest evidence of what the article is about. It's Internet2 - not Internet1. Thus, it is not aimed at consumers.

Perhaps I'll say that again - not aimed at consumers.

Thus, that point about Netflix - irrelevant.
Online gaming - irrelevant.
Videoconferencing - irrelevant.
Providers, markets... - irrelevant.
"Consumer attention" - irrelevant.

Internet2 is a RESEARCH network for scientists and academics. Stuff like CERN (nuclear / atomic physics data), which actually DOES need this sort of bandwidth.

I mean, you have a decent enough point were this article ever remotely talking about creating these bandwidths for the resale-to-consumers market. But it's not.

Re:Is it needed?

[tompaulco]

>not aimed at consumers
We'll get it anyway. Just like Internet1

Might as well correct all of your mistakes...

[Anonymous Coward]

acrynums = acronyms
abbriviated = abbreviated

Re:Might as well correct all of your mistakes...

[oexeo]

I was educated in a foreign language, so my spelling in English sucks. I only pointed out the misspellings I noticed.

Re:Might as well correct all of your mistakes...

[SlayerofGods]

And we're here to help you find the rest. ;)

Better wait

[koi88]

I dunno, my internet seems still pretty fast.
I guess I skip this internet2 thing and just wait for internet3.

Re:Gigabits!

[3terrabyte]

What I want to know is how to fill up a hard drive that size in 10 seconds. This speed is nice and all, but it means that our hard drives are now the bottleneck.

I mean what's the speed on a RAID-5 SCSI?

Re:Gigabits!

[Anonymous Coward]

that's my entire hard drive moved in 10 seconds.

Ha., I bet I'm able to move your entire hard drive in less than a second, just hand it over to me and I'll show you... ;-)

Possible uses?

[yetanothermike]

Instead of looking at the possibility of beefing up your catalog of Futurama episodes, think about the new uses for it.

Medical imaging produces very large files, and the need to transfer them over distances quickly to save lives is real.

The possibility for video is great as well. Imagine getting multiple feeds of the next WTO event from different sources on the ground. Or quality alternative broadcasting that isn't just some postage-stamp-sized, pixelated blobs. Torrents are nice, but there is something to be said for being jacked in live.

And for those who didn't RTFA, it's 3 DVDs a second.

A 200MB HD will only get you 1.6 gigabits

[PornMaster]

Thanks for playing the home game. Unfortunately, due to a math error, we miscalculated the entry fee and have deducted $18000 from your bank account. Please come again.

Entire hard drive

[dark-br]

Yep, your entire hard drive moved in 10 seconds but the question is: How do they got those read/write speeds?

Your HD would never reach that... hdparm gives me 40mb/s if I am lucky.

Maybe they have a *LOT* of RAM :)

Re:Entire hard drive

[InfiniteWisdom]

Read from differents HDs on the network in parallel

Re:Gigabits!

[RetroGeek]

This speed is nice and all, but it means that our hard drives are now the bottleneck.

Moving hardware is always the bottleneck, if for no other reason than you have to wait for the data to reach the read/write head.

When we were setting up out first networks, it was faster to load eveything off our Netware servers, than to load it off the local hard drive.

The Netware server cached the files in RAM, then served them up over a 10BaseT LAN. This was noticebly faster than going via the local hard drive (pre-IDE days).

Nah, who needs a station vagon

[Anonymous Coward]

Just get that backup tape from place A to place B,
so that you write to it at place A, then it scrolls to place B, then read it at place B, then gets written to, then scrolls to place A. Of course, that would take some 10-1000 km of tape with some exotic routing, but that would be cool :)

--Coder

Why TCP?

[ReeprFlame]

Why is the Internet2 still developing protocols based on TCP. I thought that they were proven to be somewhat inefficient with others proving better. I could have also sworn I read I2 was developing a new protocol that was almost a blend on TCP and UDP. Maybe that is what this is? At least I am happy with the new speed records... Shows development!

Re:Why TCP?

[Anonymous Coward]

Because unfortunately we are stuck with TCP/IP there are better protocols but trying to change people is quite hard so its much better to either change the stuff below TCP/IP (make better networking devices) or above it (add overlay networks, etc).

Really not sure how you could say blend UDP and TCP since UDP is basically just TCP minus flow control, message ordering, and message resending.

Re:Why TCP?

[ReeprFlame]

There was something that I believe was on Cisco now that I recall. What it did was mixed the reliability of TCP with the speradict Transmission of UDP into some package that was useful. Of course that is with several modifications to completely work. I do not think the protocol is public, but was worked on in a lab.

Re:Why TCP?

[cottrell]

There is work going on both to improve TCP so it works better on Fast Long Distance networks (see for example http://www.slac.stanford.edu/cgi-wrap/getdoc/slac- pub-10402.pdf for a comparison of the performances of some of these new versions of TCP, and to come up with replacements such as SCTP and UDT (see http://www.ietf.org/internet-drafts/draft-gg-udt-0 1.txt) which is a reliable transport protocol built on top of UDP. We did make tests with UDT on the SC2004 testbed and it perfomed very well (~4.5Gbits/s) which was not as good as we saw with TCP but it is early days for the UDT so it is looking hopeful.

What I want to know is...

[daveschroeder]

...how fast this could transfer the sum of all data (DNA, memory, etc.) contained in a human.

Yes, I'm kidding. But only half kidding. In some crazy future where we can reconstitute energy into matter, how much bandwidth would be needed to do this practically? Do we even have any ideas or estimates on how much storage would be needed to accurately represent the nature of the human body in terms of data? And no, I'm not talking about the "memory" of the brain - I'm talking about the physical manifestation of the body itself, of which the memory of the brain is a part.

Re:Gigabits!

[neverutterwhen]

So what? Progress is progress people. We do it anyway and sooner or later, in a couple of decades, perhaps we'll all have holographic storage and optical computers all shoving huge masses of knowledge/power entangled bits through hyperspace to the outlying colonies. Trust me, we will work out ways of using that bandwidth, even if it is just protect servers from slashdot.

Re:Gigabits!

[Anonymous Coward]

Time to get a new hard drive...

can the disks cope?

[CProgrammer98]

Such a blazingly fast connection is amazing, but how the hell do they get the data onto and off the pipe? Are the disk read and write speeds up to that? What about the ram? how the hell do they do that!!!

Re:can the disks cope?

[Zemplar]

"Such a blazingly fast connection is amazing, but how the hell do they get the data onto and off the pipe? Are the disk read and write speeds up to that? What about the ram? how the hell do they do that!!!"

The article didn't specify that the data was unique data, only that it was equivalent to such things as DVD or Library of Congress data. The easiest way to genertate that throughput is just have a script generate random [or semi-random] data for the transer. Hell, you wouldn't even necessarily need a hard disk to do it this way.

Re:can the disks cope?

[Anonymous Coward]

RAID 0 across a good number of these bad boys.

Re:can the disks cope?

[Anonymous Coward]

Such a blazingly fast connection is amazing, but how the hell do they get the data onto and off the pipe?

Well, at a guess :

PC1 # cat /dev/zero |nc 10.0.0.2 9000

PC1 # nc -l -p 9000 >/dev/null

Come to think of it, if you change the first line to read:

PC1 # cat /dev/zero |bzip2 |nc 10.0.0.2 9000

You could probably acheive a record breaking 100 Gigglybytes per second on your home network!

Re:What I want to know is...

[Anonymous Coward]

There are on order of 10^27 atoms in a human (6.022*10^23 per 12 grams or carbon so about 10^26-10^27 for a 100kg carbon blob)

That at even 100Gigabytes per second assuming say 100bytes per atom is 10^16 seconds or about 2% the age of the universe (100million years)

We need another 10^9 increase

So what?

[tattoi.nobori]

Alright, so Caltech made a special pipe transfer a hojillion bits in a few seconds, shattering all previous records. This sort of thing will become exciting for me when Comcast pulls up in front of my house with a spool of fiber...

NOT 101Gbps SENT OR RECIEVED,

[Anonymous Coward]

But 101 Gib Sent AND recieved
Avg IN - 18.79 Gbps PEAK of 45.59 Gbps
Avg OUT - 26.81 Gbps PEAK of 53.82 Gbps
Avg In&Out - 45.59 Gbps PEAK of 101.6 Gbps
Still Pretty Impressive.

Re:Too Fast for hard drives

[Anonymous Coward]

Canadian researchers at CaNet3 did an interesting experiment around this very question.

What do you do when your network is faster than your drives?
You turn the network itself into a drive - a giant drive made of light and 1,000 miles in diameter.

Basically, the idea is that instead of accessing data relatively slowly from a server's drive, you instead keep the data spinning around the fibre network at the speed of light. If anyone wants something - a DVD quality movie for example - they peel it off as it comes whipping by. I'm not sure what speeds they were working with, but I do recall that a DVD would take less than 1/4 of a second to download. Once you hit these kinds of speeds, everything is always everywhere.

Nope.

[hummassa]

FAST TCP = Better window handling, active congestion monitoring and optimization of link use.

Re:LoC?

[KontinMonet]

APL, a very terse language that required a special keyboard: sigAPL

Get the MPAA access fast

[Sabalon]

That was obviously some new movie that came out being transfered. We better get the MPAA access to Internet2 even faster now, otherwise box offices across the country will shut down.

Problems inherent in high speed networks

[F4Codec]

A paper I wrote a while ago...

Some Perceived Problems with the Introduction of Terabit Network Technologies.

This short paper attempts to highlight some potential problems associated with the introduction of high speed networking - specifically at the Terabit per second level. These problems are still in the theoretical arena as practical Terabit networks are probably still several weeks away from fabrication.

Introduction.

The primary problem when considering Terabit networks must be the enormous speed that the packets on such networks will be traveling. Naturally there are problems at the protocol level with very large window sizes necessary for useful throughput, and enormous quantities of data "in flight" at any one point. However, these problems are encountered at the Gigabit level and are solvable in principle (by appropriate window and packet size negotiation for instance).

The major problem that is perceived at such high speeds is that data is now flowing at a significant fraction of the speed of light. This brings into play a number of relativistic effects that must be taken into account when designing such high speed networks.

Physical Considerations.

There are firstly a number of physical considerations that must be taken into account. These are problems associated with any body traveling close to the speed of light (c).

1. A large amount of energy is required to accelerate the packets to the required velocity. However, the closer one approaches c - the more of that energy is transformed into mass. Thus packets will become heavier. A related problem is the slowing down of packets, when they enter conventional lower speed Megabit networks. The large amounts of energy that have gone into accelerating the packets and giving them extra mass will be lost. This will require large heat sinks. Cable fractures may also be explosive in these cases (which is in keeping with the abbreviation TNT Terabit Network Technologies). Alternatively, a special large coil of cable could be used to allow the packets to naturally slow down.
2. Networks often need to be laid to fit the physical shapes of buildings and other infrastructure. When any body traveling close to c undergoes acceleration it tends to emit "breaking radiation" or bremsstrahlung. This is particularly noticeable when bodies have to undergo angular acceleration when turning corners. Thus any bends in the cabling will have to be heavily shielded with lead plates to stop the intense burst of high energy particles. At high enough speeds, the curvature of the Earth may also prove a significant loss of energy.
3. Whilst traveling at high speeds, the packets will undergo time-dilation effects. Thus whilst two ends of a connection may agree on a round-trip time for a packet, this may well be different to the packets perceived RTT. The packets estimate of the RTT will be shorter than the measured delay. Therefore if times are kept in the packet this will lead to confusion.
4. When a body is traveling at high speeds, it tends to shrink in the direction of the travel. This means that a packet taking 1400 bytes, might actually take up 1300 bytes space on the network. This leads to more capacity being available than might first be perceived. However all routers must be able to handle packets at speed to stop them suddenly growing. This leads to circuit switching being possibly a better base technology.

A perhaps more serious problem is the case of collisions on a network technology such as ethernet. The collision of two very high speed packets could give rise to many spectacular effects, equivalent to those seen in current particle accelerators. In par

Swoosh!

[magefile]

I think the subject says it all.

now that's fast!

[UltraAyla]

I don't care how long they ran that test for! where can I get one of those? I guess I'll have to wait 10 years...

Asking question when not required...

[zijus]

I read a lot of : is this needed?, let's be clever and ask oneself what we are doing...

Frankly, it is hilarious from folks who probably jumped on GMail, IPods, stupid phone which does all but work when needed, and other devices which are arguably the most un-needed space on the planet. (No you won't get me to believe your 200MB emails are worth keeping...)

Ciao

As a reminder, the ALICE experiment at CERN will produce per year 1 PB ( Peta Byte ) of _raw_ data. This is only _one_ experiment out of _four_. Add DB overhead and you start getting the picture. And no: there won't be backups: too big. The nature of particle physics is to be statistics. The search is for slight deviations from what is predicted. So the amount of raw data is huge. It is also that the amount of (raw) data per second produced will be in some case magnitude of order bigger.

It is thought that some data will not be stored at all at CERN, but sent straight to remote storage farm. Too much data to be stored localy.

The people analysing those data will be scattered over the planet, involving indeed the need of big transfers.

Ha ha ha: is this needed ? Hi hi let's think about it... Please dump all the crap data you pretend to need and ask again the question.

Re:Asking question when not required...

[a8o]

Well said and I think we can extend this to the advent of cheap hard drives and CD/DVD-R/Bluray discs. We keep things because we can, not because we can ever use all that data. My ADSL runs at 256/64 kbit and with that I can get more music, more movies and more data than I will ever know what to do with. The internet is so vast, one could spend his entire life using it or consuming its products (files from P2P.) The question is not whether it's needed, but whether what we have achieved is enough and I think the answer is that until we have instantaneous access to all of the internet's resources, we have not completed its construction.

Nice monitoring tool

[duffbeer703]

http://boson.cacr.caltech.edu:8888/

A Jini-based, self-discovering network monitoring tool. That's pretty damn cool too.

And I thought that Jini was totally ignored after I bought "Jini in a Nutshell" for $0.50 at a church book sale!

Re:Nice monitoring tool

[costing]

They also have a nice map with the Internet2 traffic:

http://boson.cacr.caltech.edu:8080/

Re:What I want to know is...

[Idarubicin]

...how fast this could transfer the sum of all data (DNA, memory, etc.) contained in a human.

Another poster has already provided an excellent summary of how long it would take to transfer a whole 'human', assuming 100 bytes per atom.

I will note that DNA is actually easy. Since it's massively redundant--just about every cell has a copy of the same stuff--you only need to send it once. The entire human genome is three billion (3E9) base pairs. Each base is one of only four possibilities, so that's just two bits each.

Without annotation, you can fit the entire genome into about 750 megabytes--it will just barely squeeze on to a CD. Actually, there are a number of repetitive features, so it can be compressed further. The genome is big, but it's not huge.

Movie delivery to theaters

[silverbax]

I've wondered for some time when speeds get fast enough, what's to prevent movie delivery to the theaters from being digital?

How about this...a movie theater that can pay per-viewing to the studio?

OF course, most people will just stream movies across anyway...

Oh, well.

QuipWire - Bad Press Run Amok

Units, Units, Units

[Limburgher]

What's that in Terabytes/Fortnight? Not as cool as the speed Furlongs/Fortnight, but. . .

Re:Units, Units, Units

[bbdd]

consult the oracle

Internet2?

[Sourdough]

This is not an Internet2 speed record (lsr.internet2.edu), which is measured between single host pair. This demo was done for the SC2004 bandwidth challenge using a large number of hosts.

Surprise!

[frankvl]

It's about the size of 60x15x3 DVD's

A practical application

[davidwr]

I'm a professor. I need to run a calculation against a terabyte of satellite imagery data that's stored at sites scattered across the world.

The calculations aren't that complicated, the big headache is reading the data - the faster I can read it, the faster I get my answer.

If I can suck that data to my local mainframe at 1 TB/sec, I can get the answer in about 2 seconds. If it takes an hour to suck the data down, I'll get the data in about 1 hour and 1 second. Even at 100Gbit/sec, that's still only a few minutes. Compare that to the 100Mbit/sec. or so I probably had 10 years ago.

This is a one-off project and it's not worth trying to figure out how to move the calculations to the data, not if I have very fast data pipe.

I think mainframes can do it

[davidwr]

I think a suitably-fast mainframe can do the job:

Connect a suitably huge bank of hard disks to the mainframe either as suitably-fast raid-0 or on separate suitably-fast channels, and send the data to the disks as fast as it comes in.

For a sufficiently large value of "suitably" this is guarenteed to work.

Did I miss something?

[Tribbin]

Do you mean that we have no reason to be impressed or that the readers can't read.

Or do I miss something?

It's a good markup but not really worth (Score:5, Informative), right? There must be something I missed in your post.

Re:LoC?

[MasamuneXGP]

Actually Basic (VB now) still does that. But that doesn't make the concept of a single-line program any less stupid.

This is a serious problem for some networks

[davidwr]

In particular, RFC-1149-compliant networks suffer from this problem, as do 747-based networks.

Actually a very good remark

[Anonymous Coward]

Meant as a joke, this is a very good remark. I have never understood hopw you can compare transfer speeds when no distance measure is involved...

Re:What I want to know is...

[numist]

I work in an AI lab, and we do a lot of work on FMRI scans. Each of them is a gigabyte or two.

Since in orderf to find patterns between brains in different states we need multiple brains for multiple brain activations, this translates to a messload of data. We use dual Opteron systems here with a terabyte of disk each (7 workstations), and they are barely sufficient for the task.

I, for one, would see a significant improvement if I could upgrade past the gigabit we already have running in the lab.

LoC/15 minutes

[doktorstick]

LoC/minutes is a horrible measurement. With all the trashy romance novels, self-help, and diet books being published every day, we'll still be at a LoC/15 minutes in a millennia from now.

Re:Gigabits!

[3terrabyte]

I am not a neigh-sayer.

I am just curious what they used. Obviously they used something if they're touting "sustained" bandwidth at this speeds.

Ssshhh...

[GrodinTierce]

Don't let the MPAA hear, they'll try to use it as evidence of piracy.

obligatory...

[hakkikt]

I, for one, welcome our massive bandwidth pr0n distributors overlords.

At last, we can get info faster than we can save it!

640K ought to be enough for anyone!

[PhYrE2k2]

640K ought to be enough for anyone!

I seem to have 380MB in use with just XP and this Mozilla Window... I wonder if Bill Gates' computer just has really good compression :)

-M

Correction

[Y2]

When CERN comes online in about 5 years, it's expected to churn out petabytes of data. Yeah. I meant that. Petabytes, as in 1024 terabytes. Fermilab is already turning out terabytes but it will be surpassed greatly by CERN.

Fermilab has already turned out a small number of petabytes (2 and a bit). CERN is expected to produce on the order of 10 pB/y.

SCTP "Applicability Statement"

[anti-NAT]

From RFC3257 - Stream Control Transmission Protocol Applicability Statement"

SCTP is also connection-oriented and provides all the transport services that TCP provides. Many Internet applications therefore should find that either TCP or SCTP will meet their transport requirements. Note, for applications conscious about processing cost, there might be a difference in processing cost associated with running SCTP with only a single ordered stream and one address pair in comparison to running TCP.

So you're right, SCTP can perform an equivalent of TCP's functionality, although with additional features, and therefore additional complexity. Of course, increased complexity usually decreases performance. As performance was one of the major goals of this speed test, SCTP would probably have not have been appropriate.

Re:Gigabits!

[Anonymous Coward]

The fallacy is thinking you're just going to use one drive, or one SCSI controller. In reality, if you were serious about trying to read/write everything on the network from/to disk, you'd be using hundreds of disks. And they'd easily do the job. Easily. It's not latency, it's bandwidth. Bandwidth scales.

LoC/15 min

[torrents]

they should just use most dsl providers fav metric... how many times faster than 28.8 dialup???

Re:What I want to know is...

[Anonymous Coward]

It depends on whether you believe quantum state is necessary to store. If so, I'm sorry, all the bytes in the world won't help you :)

Re:Wal Mart -- 400+ Terabytes of data

[Anonymous Coward]

Subject line gives at least one source of "lots" of data. Now, whether they're willing to share, that's another story. Perhaps if they were offered some quick form of backup.

Then why does my local net only get 8Mbs max ?

[Anonymous Coward]

When transfering files locally on my suposed 100Mbs ethernet, I can only get about 8Mbs max, this is with two linux boxes and ethernet cards that bus master. So what gives?
I asked some friends and they have the same problem, so it isn't just me. I also tried it on several machines here, all over 700mhz processors. It shouldn't take that much CPU to transfer data when bus mastering does most of the work.
So what is the 100Mbs spec for anyway? Raw bits on the wire? And its reduced that much by time tcp is applied or something?

More details on the bandwidth challenge

[cottrell]

In answer to the question: "Nowhere in the article does it say how long they ran the test for. A second? A minute? An hour?"

On many of the 10Gbps paths we were able to sustain sending over 99% of the available bandwidth for hours at a time. We sustained an aggregate of over 100Gbits/s for about 2 minutes. The median aggregate bandwidth over 48 minutes was about 66Gbits/s. The HEP bandwidth challenge test ran for about 48 minutes.

Much more information is available at: http://www-iepm.slac.stanford.edu/monitoring/bulk/ sc2004/hiperf.html

Re:What I want to know is...

[Anonymous Coward]

for a 100kg carbon blob

Interesting post - ta.

Random nit: most times I see an average figure (e.g. max load in elevators) they use 75kg or 80kg.