re: Cisco's HFR Gets ModTasman dudes, Procket is up for grabs and cheap. It makes double economic sense to not lose it to competition. HFR is a money-eating machine, dont shy away from that truth.
re: Cisco's HFR Gets Mod----------------- What routers with that much capacity meet power/weight/thermal for NEBs compliance?
Avici actually. ~~~~~~~~~~~~~~~~~
OH? The GR-63 says 1950/watts per square meter. Even your smallest box (the QSR) exceeds the limit measuring in at 2687/per square meter. That makes it NON-COMPLIANT. Did I miss something?
----------------- Very few sites need that kind of capacity of course, that's why it's kind of pointless to compare the number of OC-192 ports per rack for these boxes... 64, 48, great, how many of those fully-loaded boxes do you have deployed? ~~~~~~~~~~~~~~~~~
There are a couple of reasons why customers aren't fully populating their boxes. One of them being the packet blocking nature of crossbar architectures. The point is however that port density per rack is a metric that relates to total cost of ownership. As Tony mentioned in a later e-mail, port density very much matters when networks scale northwardly. VOIP deployments for example are growing at a steady rate.
re: Cisco's HFR Gets ModNot if you have graceful restart extensions. The IGP/BGP process will come back up and continue to work. Modular design gives you HA even when you have only one route processor.
Juniper would get 5 T640's plus the TX, for 160.
Procket would get 5 8812 plus , for 240.
Foundry gets 9 NI40G's for 288