Re: BGP/4 (fwd)

System Manager (root@oberon.glo.be)
Mon, 9 Dec 1996 12:16:00 +0100 (MET)

> Keep in mind that the IOS also eats a lot more RAM than ComOS, and the
> room needed to store routes depends in large part of the format you store
> them. You cannot compare the two directly.

True. But there is also the matter of CPU power, I fail to see why cisco
needs dedicated RSP and SSP cards and IRX needs nothing to do the job.

> >case current portmaster hardware is not capabable of dealing with such
> >enormious routing tables let alone simply using 2 fully loaded T1 lines.
> This is not true. IRX-112 and IRX-114 are routinely used with 2 T1s
> maxed out and they handle it fine.

With a very small routing table and no filters I assume :-) I have seen
the performance of a pm2 declining when increasing bandwidth and so at
384kbps I switch but instead of going IRX I buy cisco since I also *NEED*
classless routing, keepalives, custom and weighted fair queueing, isdn
backup, clockrate, header and payload compression, ...

> >seeing it (still waiting for pm2er30) my guess is that Livingston will have
> >BGP4 running by the time it is ancient history, I'd rather see some effort
> We already have it running internally.

Now this gets interesting ... when will *WE* see it running ?

> We have seen almost no demand for IPv6. And looking at the industry it
> seems to be almost studiously ignored. And looking at the IETF docs on it
> I can see why - there is still much to be worked out before it is a useful
> replacement for IPv4. And now we have splinter groups arguing that v6 is
> flawed and proposing new specs for other systems. This flares up on the
> IETF mailing list from time to time.

It will be ignored upto the point the whole thing is on the verge of collapse,
THEN those who have a working implementation of something capable of doing the
job will set the standard and eat the cake.

--Steven.