We are looking to terminate a GigE circuit running about 700Mbps during peak hours. We will be using three interfaces and running iBGP with a half dozen or so peers and a eBGP session with our upstream with a full table. Last month Claudio mentioned routing 450kpps, what sort of performance can we expect in the above situation? Is this do able under a 3.9 GENERIC i386 kernel? How much a performance hit should we expect with a full table vs. decoupling the FIB and having a few static routes?
Dustin Lundquist Optic Fusion
