New Internet Speed Record

Himanshu writes "Researchers have set a new data transmission record over the Internet2's high-speed backbone. The new record announced Tuesday at the Spring 2004 Internet2 member meeting in Arlington, Va., was for transmitting data over nearly 11,000 kilometers at an average speed of 6.25 gigabits per second. This is nearly 10,000 times faster than a typical home broadband connection. The network link used to set the record spans from Los Angeles to Geneva, Switzerland."

Comments

also....

[Anonymous Coward]

"Recent studies by the U.S. Department of Energy have shown that researchers in high-energy physics, astrophysics, fusion energy, climatology, bioinformatics and other fields will require networks in the terabit-per-second range within the next decade."... and games as well..

Re:also....

[moviepig.com]

...researchers in high-energy physics, astrophysics, [etc.] will require networks in the terabit-per-second range...

Trying to be cynical, I wonder how much of that new "requirement" is just using the Internet as a big, cheap backplane bus, i.e., for parallel processing.

Or maybe there's a newly crucial need for the conveniences of full-access telecommuting ...in which case the petitioning physicists may be joined by, say, Citibank.

How many wires?

[Adriax]

I can't see it in the story, how many wires/fiber pairs were used?

If it was a single pair, then DAAAAAAAMN...

How long until broadband speeds up for mainstream?

[trublue]

How long do /.ers estimate until available broadband will reach speeds such as these for the consumer?

Less than 10 years?

Re:Jack Valenti

[weave]

Brings up an interesting point. Because Internet2 is not a public net, can RIAA and MPAA's hired goons get into the net to look for file sharing activity?

Re:Awsome....

[strictnein]

I could fill my 250 gig drive with pr0n in 3 seconds!

Really? Hmm... the connection ran at 6.25Gbps. That's roughly 750MBps. At that speed it would take about 330 seconds, or 5.5 minutes to fill up your HD. Of course, there are some other problems as well (HD speed, etc).

If it's not important, why mention it?

[hal2814]

10,000 times faster than the average broadband connection, but even the article states that that kind of bandwidth is not useful to home users yet. Then why mention it? How about this:
This new transmission is 2,796,206 times faster than a 2400 baud modem!

That's an equally useless comparison but at least the number is higher. You don't get to see a useful comparison figure until the 3rd paragraph where it says that the previous record was 4GB/second. They really should first and foremost tout the 36% increase in speed over the previous record. That's pretty impressive.

Re:I just hope this never gets to Joe User.

[hooded1]

This has sort of happened, not at quite 5 gigabit a second. I am at a school that has Internet2 access. A couple months ago a security hole opened up in one of our unix networks. Several outside users gained root access and launched a DDoS attack against some computers at BU using the full gigabit pipe. It was fixed relatively quickly and security has since been tightened.

AccessGrid

[hot_Karls_bad_cavern]

Check out Access Grid for at least one reason that we may end up needing all of that pipline. Imagine for a second if you will everyone having a personal AG node...you can pump out 5-6MB/s without *doing* much of anything. Our campus bandwidth (which is Internet2 enabled) would be shot with less than 20 people. Imagine what we'd need if all 15,000 needed 5-6MB/s all the time?

Sounds far fetched, but then again a great many things sound "far fetched" when considered before their coming.

We'll need pipelines that big and bigger...just you wait!

Hard drive rate

[bonch]

If, say, you were downloading a 20GB file...would the hard drive even be able to keep up with 6.5 gigabits a second? What does that translate to in megabytes?

IPv4

[liam193]

Am I the only one who picked up on this? They portray this as a test of a standard IPv4 file transfer. They are only half correct. While I don't doubt they used IPv4, they certainly didn't utilize standard TCP which their implications of this being similar to other transmissions (unlike the IPv6 test performed earlier) is way off base.

For example, if this test were performed with TCP, the largest TCP window size is 64K. Since TCP transfers must have an ack every window, you can only send 64KB in the amount of time that light travels the 11,000,000 m and back through fiber cable.

Some basic information:

d = 11,000,000 m (each direction)
c = 199,861,638 m/s (in glass / fiber)
W = 65536Bytes (TCP Window Size in Bytes)

The theoretical bandwidth for a transfer over this distance using TCP window sizes:
bw = (W * 8 bits/Byte) / ((2d)/c)
bw = 4.763Mb/s

So basically they had to use something like a UDP file transfer. While this is not an uncommon thing, it certainly isn't anything as "typical" as it's made to sound.

Re: Pr0n (DVDs / sec)

[Thuktun]

At 6.25Gb/s, about 6 seconds. I can hear MPAA quaking in their boots. Or, if you prefer, you could stream about 600 DVDs simultaneously. *drool*

Alas, even RAM-based SAN devices can't keep up with that bandwidth by half. Time to use latency of network loops as a storage mechanism. =)