[CLUE-Talk] "increase throughput of yer broadband 'net connection - windows users need not apply"

[black at galaxy.silvren.com]

Smoke and mirrors.

I'd venture that the way this software works is primarily by messing with
the TCP window size. This smacks of every other "connection optimizer" on
the market, for Windows or otherwise.

"In tests using off-the-shelf hardware, the CalTech scientists achieved a
speed of more than 8 Gbps, which they claim is the fastest throughput
ever."

What on earth are they talking about? Nobody has ever pushed more than
8gbps of traffic? I beg to differ.

Way too much hype in the article vs. practical information. If someone
actually used this, I'd be curious whether they can tell ANY noticeable
difference.


On Thu, 7 Aug 2003, Evan Widger wrote:

> bill ehlert wrote:
> >
> >
> > "Internet on Steroids
> >
> > "The fastest hardware in the world will not do
> > you much good if your software can't take
> > advantage of it. The same principle applies to
> > the Internet. Researchers from the California
> > Institute of Technology recently demonstrated
> > this with a zippy new protocol that can squeeze
> > more horsepower out of high-speed Internet
> > connections--up to 6,000 times more horsepower.
> >
> >     . . .
> >
> > "Oh, and there's one more catch: Pokey Windows
> > servers need not apply. For now, FAST runs only
> > on Linux."
> >
> >
> > http://www.pcmag.com/article2/0,4149,1195048,00.asp
> >
>
> that's cool, i was just thinking earlier today looking at a download
> that was transfering at 4.3k of file data but 5.5k was passing through
> my nic and i was wondering if there was some way to make the TCP/IP
> overhead lower (i'm on dialup but i think i'd notice the speed
> difference even more if it was a leaner protocol).
>
> -Evan Widger
>
> _______________________________________________
> CLUE-Talk mailing list
> Post messages to: CLUE-Talk at clue.denver.co.us
> Unsubscribe or manage your options: http://clue.denver.co.us/mailman/listinfo/clue-talk
>