[6605] in North American Network Operators' Group
Re: NAP/ISP Saturation WAS: Re: Exchanges that matter...
daemon@ATHENA.MIT.EDU (Brett D. Watson)
Sun Dec 22 04:01:49 1996
To: David Schwartz <davids@wiznet.net>
cc: Tony Li <tli@jnx.com>, nanog@merit.edu
From: "Brett D. Watson" <bwatson@genuity.net>
Reply-To: bwatson@genuity.net
Date: Sun, 22 Dec 1996 01:55:40 -0700
>
> On Fri, 20 Dec 1996, Tony Li wrote:
>
> > You should be more specific about _why_ you want a MUCH bigger CPU. IMHO,
> > the box needs more packet switching capacity (and more backplane
> > bandwidth), but there's enough CPU there for the OS.
>
> Recomputing larger route tables, especially OSPF. Tunneling.
> Encryption. Acess list filtering. Finer grain accounting (I'd love to see
> usage stats for each interface over a day or so right from the router).
> And most important to me, not becoming unusuable during a major routing
> change.
well, there is netflow. not quite the granularity you're probably
looking for but it helps. and since frontier has signed an exclusive
deal with cisco i suspect they might start moving more of the "probe"
functionality into the routers and switches. the cat5k now supports
a mini-rmon in the box.
-brett
