"routing table slots" and the real problem

I have always subscribed to this fundamental concept; memory to
accommodate N number of routes has never been an operational
issue.

Not the case with one vendor's boxes. Memory _was_ the issue
after introducing CIDR. AGS/+s and early 7000s were running out
of memory, not out of CPU cycles.

--vadim

This message is brought to you by the society for history preservation :slight_smile: