[nsp] ip load-sharing per-packet - cef accelerated ?

[Scott Farber]

On Monday, March 10, 2003, at 07:06 PM, Dan Hollis wrote:
>
> Has anyone on this list actually done an analysis of CPU usage for CEF
> per-packet vs CEF per-destination?

At least one of the big boys, the "you don't terminate directly on my 
router unless you've got an OC-ohmygawd circuit" crowd found the 
utilization didn't scale well on boxes with several hundred customers.

Incidentally, to switch vendors for a moment, JunOS has a nice little 
knob to set the forwarding options hash to L4.

-s

> -Dan