Ascend Communications commissioned The Tolly Group to compare the Fast Ethernet performance of its GRFTM 1600 and GRF 400 (version 1.3) to two models of the Cisco 7513 router (version 11.2(4)), using two versions of its routing engine, the Route Switch Processor (RSP) 2 and Cisco’s latest, the RSP4. Testing was performed from April through July of 1997.

The Tolly Group determined aggregate throughput by sending streams of traffic through each device. Additional streams were added until the maximum throughput of each chassis was reached. Additionally, to approximate a more real-world routing scenario, The Tolly Group determined how both routers performed while forwarding IP packets to varying Class C IP destination addresses.

The results show that the aggregate throughput of the GRF 1600 was nearly triple (262%) the Cisco 7513/RSP4 processor, and 389% greater than the Cisco 7513/RSP2. Further, The Tolly Group found no significant degradation in performance of the GRF 1600 when routing packets to the same destination address or packets to varying IP addresses.

Performance of the 7513 was adversely affected when routing packets to varying destination IP addresses. The 7513’s aggregate throughput (RSP4) dropped to nearly one fifth of its single-destination packet performance when forwarding frames to varying IP destinations.