North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: IPv6 allocatin (was Re: ARIN Policy on IP-based Web Hosting)

  • From: Hank Nussbacher
  • Date: Sun Sep 03 08:33:06 2000

At 17:02 01/09/00 -0400, Deepak Jain wrote:

> The Internet as we know it took 30 years to come to everybody, and it was
> providing something that essentially wasn't provided before.
> Internet2 provides the same thing, only faster; so why do we believe it will
> replace the existing solution *EVER*, much less in a shorter time frame than
> 30 years?
> The Internet is going to be with us for a very long time. Emphasis should be
> on fixing it, not replacing it.

Is it just me or is the Internet2 not a very big deal? Its based on OC-48c
and OC-192c technology as far as I know, so why would its adoption change
_anything_ except the some switch or router gear that guarantees QoS?

I admit my information may be woefully out of date; I saw an article in
InternetWeek about transfering a file at 40Mbit/s. I wouldn't see the big
deal at 400Mbit/s.
A link to review first:

In any event, Internet-2 is *not* about having a bigger pipe (best left to James Cameron - if size really does matter) - it is about having a smarter pipe.

Smartness may manifest itself as:
- on paper: QoS end to end
- in practice - the ability to view an MPEG-1 2Mb/sec video with no packet loss from 6000 miles away. Try watching any of these at 2+Mb/sec:
It doesn't matter how fat a pipe you have - it depends how loaded it may be. Let me know how well you see it. I see it is fine - each and every time - over a mere T3.
- on paper: going at 800Mb/sec end to end
- in practice: how many TCP tuned systems do you know out there that can handle 800Mb/sec? I'll make it easier - how many systems that can handle 10Mb/sec TCP streams (sending) from coast to coast? Try:
- in practice: ADSL users suddenly got huge performance boosts when they realized they needed to tune their Windows TCP stack. I consider that a trickle down affect of I2 research, already.


Deepak Jain