8.6 GB Internet? 382
prostoalex writes "Caltech computer scientists announced the protocol, capable of delivering 8,609 Mbps over the Internet, using 10 simultaneous flows of data. The research project was conducted in partnership with CERN, DataTAG, StarLight, Cisco, and Level 3. The practical applications, according to the press release, is ability 'to download a full-length DVD movie in less than five seconds'. There is a number of papers and scientific publications available."
watch out! (Score:5, Funny)
Re:watch out! (Score:5, Funny)
Re:watch out! (Score:5, Funny)
"Back... and to the left... back... and to the left... back....... and to the left..."
Already started... (Score:5, Funny)
Re:Already started... (Score:5, Funny)
This sounds like what the Pentagon needs (Score:3, Interesting)
Argh! 8Gb (Score:5, Funny)
Re:Argh! 8Gb (Score:5, Informative)
Saying 8.6GB is off by an order of magnitude.
Sigh..
Re:Argh! 8Gb (Score:4, Funny)
Re:Argh! 8Gb (Score:2)
http://www.vendian.org/envelope/dir0/oom.html
Which basically says you round (decimal) orders of magnitude at 3. 256 is order 100, 365 is order 1000.
Or.. maybe I was referring to an octal order of magnitude
Re:Argh! 8Gb (Score:5, Informative)
Re:Argh! 8Gb (Score:5, Insightful)
Because we're humans speaking English. The assumption when humans speak english is that all numbers are base 10 positive whole numbers, unless otherwise noted.
Re:Argh! 8Gb (Score:3, Funny)
speak for yourself. this is a turing test complient computer speaking ASCII.
Re:Argh! 8Gb (Score:3, Funny)
Re:Argh! 8Gb (Score:3, Funny)
Re:Argh! 8Gb (Score:3, Informative)
Look it up: Dictionary.com definition of order of magnitude [reference.com]
Re:Argh! 8Gb (Score:3, Informative)
Base 2 is pointless because comms people use SI prefixes properly*. 1 megabit = 1,000,000 bits. Base 10.
*Probably because not all platforms use 8 bit bytes (encoded to 10 bit bytes for transmission). The comms mfr's only care about the rate on the line.
what are you talking about? (Score:4, Insightful)
No one uses 10 bits/byte.
Re:Argh! 8Gb (Score:2)
Re:Argh! 8Gb (Score:2)
Hmmm. Given that I only get 100 kilobytes a second, that is
Eight gigabit per second throughput
Oh yea, Monkelectric - it is off by exactly an order of magnitude if you are counting in octal
Re:Argh! 8Gb (Score:3, Informative)
Actually when talking about DVDs, the density of each layer cannot change.
I believe you are referring to SSSL (Single-Sided, Single-Layer) DVDs, as each since on a regular red-ray DVD disc can have two layers of data.
I'm not sure if that's the same for the famed blu-ray discs or not, I would appreciate it if someone could enlighten me.
Re:Argh! 8Gb (Score:5, Interesting)
Otherwise, who needs Internet connections if you can carry a copy of the whole Internet on 2 discs?
Re:Argh! 8Gb (Score:4, Funny)
I recall a frame wherein Dilberts boss ordered up a printout of the internet. So, you don't need to use discs, either.
Re:Argh! 8Gb (Score:5, Insightful)
Ah, but it isn't! More and more often, there is a direct 10x correlation. Serial ATA, Serial Attached SCSI, Fiber (Fibre) Channel all use 8b/10b encoding, so each 8 bit byte takes up 10 bits on the wire (or equivalent medium). Hence, on a 1Gbps link, the max transfer rate is 100MB/s.
Of course, if you have a trinary computer, the algorithms may have to be adjusted slightly...
CalTech (Score:5, Funny)
CalTech's Motto: Enabling Faster Porn and Slashdoting Through Technology
Bless those people
Re:CalTech (Score:5, Insightful)
Marge: "Does anyone need that much porn?"
Homer: "MMMmmm... one million times.... (gurgle noise)"
Okay, now to say something serious. Even with broadband, most files download painfully slow because no one can afford to constantly upgrade their servers to dish out large volume of data to the public. If you ask me, 8.6Gb ethernet would be a lot more useful. After all, huge file transfers on your ethernet are at least common place.
Re:CalTech (Score:5, Informative)
The problem is that the fastest hard drives on the market today are Ultra320 SCSI, which have a throughput of 320MB per sec... or about 2.5Gbps. Even that's theoretical, of course. And few people have an all Ultra320 datacenter.
Just pointing out that the cabling is hardly the bottleneck when you reach that kind of speed, even at the LAN level. I've seen so many people upgrade their switches to gigabit ethernet then scratch their heads wondering why the network is still slow... when the server in the closet hasn't been upgraded in 5 years. Storage will continue to be the bottleneck on the LAN for a long time to come.
Re:CalTech (Score:5, Informative)
Re:CalTech (Score:3, Insightful)
Before you out-geek me (Score:5, Funny)
Yeah, yeah. I meant Gb/s. Still not fast enough to get you laid.
Nothing new (Score:4, Informative)
Its just that optical routing is expensive and so would the switch at such a high speed.
Re:Nothing new (Score:5, Interesting)
Re:Nothing new (Score:5, Insightful)
If I were you I'd be looking for a cheaper way to make crystals, or another non-polymeric approach. Quantum dots sound promising.
Re:Nothing new (Score:5, Interesting)
Re:Nothing new (Score:5, Informative)
This isn't talking about delivering raw speed over a point to point connection but delivering a large amount of data over a shared network. It's talking about a protocol rather than a transport medium, which must account for problems such as error, latency, bandwidth, and flow control.
Re:Nothing new (Score:5, Informative)
If you read the article, you'll notice that they're talking about a new transport protocol, not a physical layer. Yes, physical layers have been able to run faster than 8 gig for quite a number of years. But if you run a TCP session across such a link, the window sizes, congestion control algorithms, ack timers, and such features keep you from getting anything like the speed of your physical link.
To illustrate with a point from the article, TCP over the 1 Gbps link they were apparently using typically maintained a throughput of about 270 Mbps, only 27% efficient even when that one session had the whole link to itself.
Previously, no one much cared, because you used those multi-gigabit links by multiplexing a whole bunch of different users; each user saw a much slower link. So, the old, slow TCP was adequate. (To a point, of course; people have noticed problems with TCP acks even on DSL links.)
One of the purposes of the "Internet 2" research and other projects was to develop protocols that would allow _single connections_ to run at very high speeds.
The announced protocol, according to the article, managed 95% efficiency and 925 Mbps with a single flow, and 88% efficiency and the 8.6 Gbps of the headline with 10 concurrent sessions over a 10 Gbps link.
So, yes, it is something new.
Re:Nothing new (Score:5, Informative)
They achieved this by modifying the TCP protocol to be more efficient. It's a huge improvement.
Of course you'd need a big pipe to the net to be able to handle this bandwidth; it will be quite a while before home users could see this kind of speed. But the new protocol sounds like a big win over regular TCP and hopefully it can be rolled out relatively quickly.
What's wrong with IP? (Score:2)
IP can scale, can't it?
Re:What's wrong with IP? (Score:5, Informative)
Re:What's wrong with IP? (Score:5, Interesting)
Ethernet's an unreliable short-distance transit protocol.
Both depend on a higher layer to endow reliability on any carried datastreams. This means a number of things -- protection against dropped packets, corruption, and especially out-of-order delivery (which turns out to be a much more difficult problem than you'd think). TCP provides these services, but not without cost. Getting TCP to scale to gigabit speeds turns out to be non-trivial. Even this experiment only achieved 1Gbps -- they had to interleave across ten datastreams to achieve the maximum possible performance.
Not to take away from their achievement -- this is cool work. I'd like to see more done to handle dropped packets, though. I was thinking about MP3-style overlapping windows to allow any individual packet to be dropped, but it turns out drops clump together...
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
MPAA Surrenders (Score:2, Interesting)
Re:MPAA Surrenders (Score:5, Funny)
i was looking forward to downloading DVD's as i watched them.
Shit, they're not talking about downloading the DVD while watching them, they're talking about downloading the entire movie before MPlayer is even finished recognizing your hardware!
Unhandleable (Score:2, Interesting)
Sigh (Score:3, Funny)
Re:Sigh (Score:3, Funny)
A full-length DVD movie in less than five seconds (Score:5, Funny)
Oh, they meant legit full-length DVD movies...
Re:A full-length DVD movie in less than five secon (Score:5, Funny)
gosh (Score:4, Funny)
That sound you hear (Score:5, Funny)
Give me units I can understand! (Score:5, Funny)
Don't know... (Score:4, Funny)
(stood on end, that is...)
Re:Give me units I can understand! (Score:5, Interesting)
Also, this page [digitalenterprise.org], at least, says it would take 88 TB = 704 Tb to digitize the LoC.
That's correct; 30.992 Tb/hour. With 1 LoC = 80 Tb, we now get 30.992 Tb/hour / 80 Tb =With the much larger figure of 1 LoC = 88 TB = 704 Tb, we get 30.992 Tb/hour / 704 Tb = .044 LoC/hour.
How long will it take for hard drives to catch up? (Score:5, Insightful)
RAID (Score:2, Informative)
Not only can a high-end storage array handle that sort of throughput, but it can do it without any bugs [bugfreeliving.com].
Re:How long will it take for hard drives to catch (Score:3, Insightful)
Who needs hard disk capacity if you can stream a movie in realtime? *eg*
BTW, even if hard disks eventually reach the required capacity, you wouldn't be able to store it on disk anyway thanks to MPAA's DRM initiative...
Re:How long will it take for hard drives to catch (Score:5, Funny)
Who needs hard disk capacity if you can stream a movie in realtime?
Why would anybody want to watch an entire movie in 5 seconds, certainly my ability to absorb information is not as good as that and I regard it to be rather high(Toung in cheek).
Re:How long will it take for hard drives to catch (Score:3, Funny)
Too bad the other 16 minutes are utterly, utterly wasted.
Re:How long will it take for hard drives to catch (Score:2)
Of course, the question remains as to whether the MPAA will wake up and realise that they can use the internet to make profit rather than vainly trying to thwart file traders.
Re:How long will it take for hard drives to catch (Score:3, Insightful)
Re:How long will it take for hard drives to catch (Score:3, Insightful)
honestly, what's the big deal? help please. (Score:3, Interesting)
bearing that in mind, isn't 10 gigabit TCP in the getting-done stages?
i don't know, maybe i missed something 'golly-gee' about this. this just seemed superfluous.
Re:honestly, what's the big deal? help please. (Score:2)
You probably mean 10Gbps Ethernet which AFAIK works already, using fiber.
Five Seconds? (Score:5, Funny)
Five seconds?? Ohhhh... but I want it NOW!
Bottleneck (Score:5, Insightful)
Re:Bottleneck (Score:5, Funny)
Well,
Re:Bottleneck (Score:2, Funny)
Re:Bottleneck (Score:5, Interesting)
Assumin its actually 8.6 bytes/sec and not bits like another poster suggested, the pci bus would become oversaturated since it can only transfer 3.2 gb/sec ( correct me if the transfer rate is wrong).
I wonder if a Sun or IBM unix box could handle this. My guess is this speed will only be used as a backbone anyway so only large unix mini's or dedicated routers will send and recieve at 8.6/gbs. Sorry Johny you can not download porn at that speed.
Re:Bottleneck (Score:2)
It all depends on how the I/O busses are laid out. For a single channel, I believe 2GB fibre channel is the fastest they can do, but if you stripe a bunch of channels together, you can probably sink that much data.
Re:Bottleneck (Score:5, Informative)
A 33 MHz/32 bit PCI bus which is standard in most PCs will transfer 133 MB/sec.
A 66 MHz/32 bit PCI bus which is in quite a few Intel and UNIX servers will transfer 266 MB/sec.
A 66 MHz/64 bit PCI bus which also is quite common in UNIX servers (and becomming in Intel) will transfer 532 MB/sec.
A 133 MHz/64 bit PCI bus which is the current standard for big UNIX servers will transfer, you guessed it, 1 GB/sec.
Mind you that these numbers are pr. PCI bus, some of the lager Intel servers, and most UNIX servers have more than one PCI bus.
Re:Bottleneck (Score:5, Funny)
Bloody repeats (Score:4, Informative)
There's nothing "Practical" about that. (Score:3, Insightful)
In other news, the DCMA was appended to... (Score:5, Funny)
I love recursive acronyms. (Score:3, Interesting)
Looks pretty slick... is it a replacement for TCP, or an overlay of TCP? The article says that it uses 10 simultaneous TCP connections....
I can't find on a cursory glance whether or not it can run on IP, but I assume so. Hopefully it will work with IPv6, when (if) it becomes mainstream.
The article mentions that has an average throughput effeciency of 95% (meaning that if you have a Gigabit connection, it can send/receive stuff at around 950Mbps). Does it drop TCP's congestion handling and do something similar to UDP then?
Eight giga-whats? (Score:5, Funny)
Re:Eight giga-whats? (Score:3, Funny)
Re:Eight giga-whats? (Score:4, Funny)
1076.125 MegaBytes per second
3.87405 TerraBytes per day
Assuming an average of 5 kB per spam comes to ~77.5 Million spam emails per day.
Yes I have no life.
HDD Speeds? (Score:5, Insightful)
We must fear such a technology (Score:5, Interesting)
In other words, such a technology would give a boost to legal attempts to allow hard DRM - as is today illegal under the liberty-preserving legislation of a lot of countries, especially in Europe.
Do not answer that the media lobbies aren't asked to give their opinions here. Because it is part of Microsoft's, Intel's and AMD's (to cite only 3 members of the vast TCPA alliance) strategy to maintain good relationships with the media companies in order to enlarge the computer market.
You know what I'm talking about - Palladium. I don't think it's necessary to insist on the fact that it would be a bad thing for us.
Re:We must fear such a technology (Score:2)
I haven't heard anything from them about that though....
Someday even MPAA will see the commerical benfit (Score:5, Insightful)
That thing is a hazard! (Score:2, Funny)
DMSCA (Score:2, Funny)
Jack Valenti has heart failure at announcement.
Hilary Rosen Responds by announcing DMSCA bill.
Berman and Hollings, after campaign contributions and honorariums, announce co-sponsor the Digital Millenium Scientist Control Act, stating that the only reason scientists could possibly have for developing the technology for such fast downloads is to support porn, piracy, and terrorists.
The Digital Millenium Scientist Control Act is written to allow scientific research, unless it can be used to deny unjust enr
This just in... (Score:3, Funny)
How to pronounce it.. (Score:5, Funny)
That would be FAQMSTCP..in other words pronounced
FAH Q MS TCP...
as read on FARK...
So what? (Score:5, Insightful)
And I won't add the problem of hard drive speeds which can't handle it. Of course, big RAID arrays and the like can, but not consumer drives.
Of course, eventually, when we use a better quality encoding method for video/audio, the datarate may have to increase, but right now, it's useless.
I see two things... (Score:5, Interesting)
And I figure that by the time this becomes mainstream, the amount of data needing to be transferred will also have increased by 1 or 2 orders of magnitude, and you'll still be stuck waiting hours for the latest HoloVideo downloads. Just like you wait hours to download Attack of the Clones over DSL and Cable, and like you once waited hours to download that 5 meg shareware program over your 56K modem.
Seems like the amount of data being stored is always 1 step ahead of the amount that can be conveniently transferred... We need a war on program bloat.
This reminds me of a conversation I had last night (Score:4, Funny)
DVDs/Second? (Score:5, Funny)
In other news... (Score:4, Funny)
In other news, the MPAA has filed a lawsuit against Caltech for "aiding the piracy of copyrighted movies". The RIAA is expected to file a similar lawsuit as many wonder why they haven't already milked Caltech for all the money it's got.
Now wait a minuet! (Score:3, Interesting)
The only machine that I know of that could even utilize a connection this fast is a Cray X1 [cray.com]
Isnt 10 streams cheating? (Score:4, Insightful)
But what if this was done on a mass scale, and everyone used 10 streams to increase their transfer rate? I imagine the combined bandwidth would be as bad as a single stream was. Which I find questionable if its efficient, or a good thing for the Internet.
Re:Megabits per second (Score:2, Informative)
Overhead (Score:5, Informative)
It is just over a gig a second.
Not all of that is data. Some is packet headers. Some is error correction. That's why you can't push 6 KB per second over a v.90 dial-up connection at 48 kbps.
Re:Megabits per second (Score:5, Funny)
When you count in octal, there are 10 bits to a byte. After a few months doing coding on old big iron I accidently balanced my checkbook in octal. Took me a WEEK to get that straightened out.
Honestly though, this doesn't eliminate the bottleneck, it just moves it from the cables to the Server, or to your hard drive. Given that we can pretty much
I still would like to get it to my house, though.
Re:Too fast! (Score:2)
Not if you stay on dial-up.
Mmmmm... LPB's.. (Score:3, Funny)
Re:no thanks (Score:5, Funny)
Why would you want to do that? Don't you understand that this connection would be much, much faster than your current modem? It would be at least 24 times faster, with the potential of being up to 57 times faster. Especially during off-peak hours like Thursday at 8pm.
That would make me 24 times the pirate I am today, with the potential to be up to 57 times the pirate I am today. Then they would use that peculiar method of reasoning to assign me a sentence that would require 24 lifetimes to serve, with the potential of requiring up to 57 lifetimes to serve.
Of course, I'm all for upping the stakes, here.
Re:Slashdot 2020 (Score:3, Funny)
Re:Raping of the Acronyms (Score:3, Funny)
Maybe they're Welsh.