Performance Report for J2468A, J2469A, J2499A, J3573A, and J2806CA ATM/155 and ATM/622 Adapters
Performance Report on ATM adapters 5
The MTU size makes a difference in the scalability. See Table 6 below.
Table 6: Maximum TCP Throughput and Interface Type on K580 (four CPUs, six 155 Adapters)
Classical IP
MTU 9180 bytes
LANE
MTU 9218 bytes
LANE
MTU 1516 bytes
Outbound Mbps 654 611 247
Inbound Mbps 589 589 252
Bidirectional Mbps 666 584 249
The maximum throughput and request/response performance on the K580 is shown in Table 7
below. The figures compare the performance when using six 155 adapters versus four 622
adapters.
Table 7: K580 (four CPUs): Maximum Figures for 155 and 622 adapters
The request/response performance is CPU bound. To get an additional transaction rate, it is
better to add/allocate CPU resources instead of adding adapters.
six 155 adapters four 622 adapters
Outbound TCP Throughput Mbps 654.19 812.96
CPU Utilization % 76.59 86.43
Bidirectional TCP Throughput Mbps 666 820.11
CPU Utilization % 85.8 83.92
TCP Single byte R/R Transactions/sec 22432 30296
CPU Utilization % 100 100
UDP (1K/2K) R/R Transactions/sec 15696 19579
CPU Utilization % 100 100






