On Tue, 10 Apr 2001 alex@yuriev.com wrote:
Try ~ 10ns ram which you can buy 256MB of for ~ $60-80 (www.pricewatch.com). At any rate, A raw 3 or 4 level mtrie FIB fully populated with the real 100k+ routes on the internet consumes less then 900kb, and all the interesting parts fit in the L2 cache of a Celeron A where you can do about 22,000 lookups per MHz. Hardly excessive memory bandwidth. The packet ram on the gige cards is also very fast, and could easily accomidate a dCEF approach.
CEF should be called Customer Enrangement Feature. It is a very very very bad idea to have linecards be anything else than forwarders. They should not make any intelligent routing decisions. There should not be a tons of copies of routing table on line cards. That is what creates problems.
They don't make intelligent routing decisions, the route processor does then pushes the FORWARDING table down to the individual cards. There is nothing wrong with distributed copies of the forwarding table in easy access of the hardware doing the forwarding, but you should make those copies over a bus which actually copies things correctly *coughmbussuckscough*. -- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/ras PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)