The throughput is 4GB/s full-duplex at 40 byte packets. The throughput is 5Gb/s full-duplex at 64 byte packets, and that's why on the GSR its a 2.5Gb/s line card (up to OC48).
I tested the SIP-400 with the OC48 SPA and had no problems running IMIX or random packet size traffic at line rate full-duplex. I couldn't use less than 64- byte packets with the tester so I didn't really get to see the point where it would break the SIP before I ran out of bandwidth on the OC48. I was also able to run one OC48 and one OC12 SPA on the same SIP-400 and achieve line rate on each full-duplex using an IMIX. However, the SIP will complain since it has the notion of bandwidth capacity given the SPAs that populate the SIP, a la bandwidth points on the VXR. It works, but it may be an unsupported configuration. I have inquired about that and haven't received an answer yet. The only real limitation with the PPS rate I could see is a DoS attack where the packet size is <64 bytes, then you could theoretically run out of processing power on the SIP before you fill your OC48. Phil On Jun 30, 2007, at 9:13 AM, Robert Blayzor wrote: > Cisco's datasheet doesn't seem to be as detailed on the SIP-400 as > it is > on the SIP-600. I'm trying to understand the limitations of the > SIP-400. The datasheet claims it has a CPU throughput of "4Gbps". > Does > that mean that the SIP-400 cannot handle a single OC48 POS SPA at > full-duplex/line rate? Or is the 4Gbps limitation full duplex (8Gbps) > as the SIP-600 is 20Gbps? > > -- > Robert Blayzor > INOC > [EMAIL PROTECTED] > http://www.inoc.net/~rblayzor/ > > Life's unfair - but the root password helps! > _______________________________________________ > cisco-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/cisco-nsp > archive at http://puck.nether.net/pipermail/cisco-nsp/ _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
