How we did it

We measured the Sun Secure Application Switch N2120V in terms of connection capacity and connection setup rate for both HTTP and secure-HTTP traffic. For all tests, we used Avalanche traffic generator/analyzer system from Spirent Communications to emulate clients and servers. We used two pairs of Avalanche 2500 appliances for these tests running Avalanche 6.54 software.

We tested Sun device as a content switch with layer-7 inspection enabled and again as a layer 4 load balancer. For both sets of tests, we configured the Spirent Avalanche to emulate 204 clients and four HTTP and four HTTPS servers. We also configured the Sun device to present one virtual IP address to all clients.

In the connection capacity tests, we configured clients to request a 1K byte object, then wait 60 seconds, then request another 1K byte object. Each client repeated the process 16 times, with the 60-second waits allowing us to build up connection count. We gradually built up connection count over a 5-minute ramp-up period, and counted the average number of connections open during a 60-second steady state phase.

In the connection rate tests, we used HTTP 1.0 to force TCP connections to be set up and torn down for each request. Again, we configured the clients to request 1K byte objects, this time without the 60-second delay between requests. We built up the connection rate over a 5-minute ramp-up period, and counted the connection setup rate during a 60-second steady state phase.

The HTTPS (SSL) tests were identical to those for HTTP except that clients issued HTTPS requests. The emulated clients offered 11 different cipher types, but the one selected by the Sun switch in all tests was RC4-MD5.

Back to Clear Choice Test: Sun Secure Application Switch N2120V