RE: Routing System Scaling - Disaster Looming, but Medium-Term Fi xes Known

From: smd@clock.org [mailto:smd@clock.org]
Sent: Monday, April 02, 2001 5:59 AM

So, there are several Deaths of the Internet which are possible:

  -- it's too expensive to keep up with growth, so utilization

  -- we blow up on one of the scaling axes

    -- too much dynamism: poof, our memory isn't

  -- we blow up because while we can handle some of the

                -- inefficient/broken routing: we don't have the power

One observation here, PC133 RAM is getting ever cheaper, as are CPUs. If
routers were designed around commoditized components their COGm would be
lower. Of course, the router vendors would no longer be able to get the
premium prices for their boxen, either.

What I wish I had finished @ Sprint: a web-based form that lets
one arrange an exception to access-list 112 for some small fee,
like $50/month/change. For your 50 bucks, you see no filtering
on one of your long prefixes for a whole month (or year or whatever).

These days it may be that your $50/month will make your
long prefix less likely to be damped out of the routing table
for a long time, however there is scope for further innovation,
such as charging only $5 to clear your prefix out of the flap-damping
penalty box of one router.

I, for one, would gladly pay for such a service. Provided that I don't have
to pay it to 32 ISPs per month.

From: Roeland Meyer <rmeyer@mhsc.com>
Date: Mon, 2 Apr 2001 11:21:47 -0700
To: "'smd@clock.org'" <smd@clock.org>, nanog@merit.edu
Subject: RE: Routing System Scaling - Disaster Looming, but Medium-Term Fi xes
Known

From: smd@clock.org [mailto:smd@clock.org]
Sent: Monday, April 02, 2001 5:59 AM

So, there are several Deaths of the Internet which are possible:

-- it's too expensive to keep up with growth, so utilization

-- we blow up on one of the scaling axes

-- too much dynamism: poof, our memory isn't

-- we blow up because while we can handle some of the

-- inefficient/broken routing: we don't have the power

One observation here, PC133 RAM is getting ever cheaper, as are CPUs. If
routers were designed around commoditized components their COGm would be
lower. Of course, the router vendors would no longer be able to get the
premium prices for their boxen, either.

Do you think it is the CPU on the box that makes it expensive? Do you
realize that most modern routers are ASIC based? Do you realize that there
is no "off-the-shelf" hardware that can do OC192c forwarding? Cost of the
control processor CPU is not what makes the price of the router expensive.

Bora