[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [FW1] Nokia IP330 vs. Compaq DL320 SolutionPaq
Chris Arnold wrote: ----------------------- >Do you have any results from your internal tests as well as the testing >criteria you could share with the group, Mark? All, I dug up the results of our benchmark testing of these two boxes, per Chris's request, and here they are. Hope you find this information useful. If not, remember what you paid for it. ;-) First, a few disclaimers: 1) We did this testing purely to satisfy our own curiosity, and the information comes with no warranties of any kind. Neither Compaq nor Nokia has verified our methods or our results. 2) YMMV. As with any benchmark, you should take it with a grain of salt. Different testing methods inevitably yield different results. That said, we made every attempt to ensure a fair, apples-to-apples comparison. Also, we tested on identical test beds in a totally isolated lab network, not a production environment. 3) These tests were performed in April on slightly older models from both vendors. Namely, the IP330 was equipped with an AMD K6-2 266Mhz CPU with 64MB RAM, and the DL320 was equipped with an Intel P-III 800Mhz and 256MB RAM. Nokia is now shipping their IP330 units with a 400Mhz processor (last I heard), and Compaq has similarly upgraded the DL320 to a 1Ghz chip. Test criteria and method: --------------------------- - HTTP transfer of 1MB and 5MB files using Mercury Interactive's LoadRunner, which is a popular commercial performance testing tool. Results from multiple test runs averaged together. - Check Point FireWall-1 version 4.1, and current OS revs from Compaq and Nokia. - We tested the Nokia both in the default configuration and with the FLOWS feature enabled. FLOWS accelerates throughput of regular IP traffic, particularly long-lived connections and small packet sizes. It does not accelerate VPN traffic and is not compatible with FloodGate-1. Some other problems have been reported with FLOWS, but I suspect Nokia and CP will eventually get most of them resolved, if they haven't already. - HA tests were performed using two units in a hot-standby configuration, state sync enabled. VRRP on the Nokia versus the built-in RainWall HA on the Compaq. Lower fail-over times are better. Passing a test means that the problem was detected and traffic shifted transparently to the backup. Failure means that problem was not detected and/or transparent failover did not occur. - Throughput here is defined very narrowly as unidirectional throughput between two fast Ethernet ports. Higher throughput numbers and lower response times are better. The maximum theoretical throughput possible in such a test is 100Mbps. Measurement of bi-directional or total aggregate throughput through multiple interfaces, which would allow higher throughput, was not attempted. The Results: ------------ Performance: Compaq Nokia Nokia w/FLOWS ------ ------ ------ Throughput: 89.4 52.4 89.4 (Mbps) Transaction Response Time: 2.01 3.45 1.99 (msec) High Availability: Compaq Nokia ------ ------ Average fail- over time: 3 sec 8 sec Power off: pass pass Cable pull: pass pass Stop FW-1: pass fail Unload policy: pass fail Conclusions: - Both boxes are more than capable of T3 speeds. In the default configuration, the Compaq was significantly faster than the Nokia, but with FLOWS enabled on the Nokia, there was no significant difference between the two in this particular test. - It appears that the simple unidirectional test method and 100Mbps wire-speed limit prevented both boxes from reaching their full throughput potential, as they both registered identical top speeds. This is likely because both reached the Ethernet saturation point before the CPU saturation point. It would be interesting to see an aggregate total throughput test that simultaneously sends files through 3 or more ports in multiple directions to determine the real maximum when the LAN bottleneck is removed. - Although we did not test VPN, Compaq will likely have an advantage for VPN traffic, since VPN is CPU-intensive and does not benefit from FLOWS. More information on Nokia IP330: http://www.nokia.com/securitysolutions/platforms/330.html More information on Compaq DL320 SolutionPaq: http://www.compaq.com/solutions/security/solutionpaq.html Best regards, Mark L. Decker Rainfinity - High Availability for [email protected] ================================================================================ To unsubscribe from this mailing list, please see the instructions at http://www.checkpoint.com/services/mailing.html ================================================================================
|