> > I've read through the data sheets, and I also can't see any signficant > > differences. I was wondering if there was some hardware differences (like > > CAM table size, ethernet input/output buffer sizes), etc...
Is the packet buffering on the -Es significantly better than on the non-Es? It would seem that the buffering capabilities of a non-E are at best limited, based on my experience - granted we have bursty server loads that we were attempting to condense down into 4-port etherchannels, but I would have expected the 3560Gs to do better than they did. I suppose it's possible that if I split the ports up amongst the ASICs better it might be better, but it seemed like there was only one TX queue buffer for the entire switch, which if you did "mls qos" you could split up some but you still had a limited choke that everything went through. I really don't want to go buy a -E to find out. I never could get an answer from cisco as to the actual design of the internals of the 3560/3750s. Is the information around anywhere? (And why the heck does Cisco keep it such a secret?) > Just think of the 3750s as baby VSS-6500s or Nexus 7Ks :) Now _that_ is hard to imagine. :) _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
