Hi.
When considering to buy or not to buy Foundry routers, switch-routers and
load-balancers, one issue that I found was packet processing scalability.
It seems that all boxes uses a flow-based model, where the first packet of a
flow goes to the CPU to be software processed (routed, switched, filtered,
load-balanced), which generates a CAM entry in the line cards that is
followed by the remaining packets of the flow.
Is this true for the entire Foundry line (BigIron, NetIron, ServerIron) ?
The 16K entries limit has been hit by someone using NetIron or ServerIron ?
Rubens Kuhl Jr.
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:05 EDT