Citrix NetScaler test summary
By
David Newman, Network World
January 16, 2006 12:04 AM ET
In a single box, the NetScaler Application Delivery System sums up much of what's right, and some of what's wrong, about the
Web front-end market. The device bristles with virtually every acceleration feature available (see features sidebar), including caching (which we didn't test), support for complex application-layer switching, SSLCi offload and HTTP compression. And it's a strong performer in some tests, scaling to the highest levels possible
when it comes to connection count, TCP multiplexing, and goodput (see scalability test results).
For all that, though, the appliance reached limits in some of our performance tests. In some cases, enabling HTTP compression
actually increased response time, even with very low transaction rates and user access speeds (see services test results). Citrix says compression results are better with Version 6.1 of its software (the company supplied us with Version 6.0), but we did
not verify this. Also, the device does very well in TCP multiplexing tests only when users have longer 'think times' between
page requests. With rapid requests from users, there's relatively little offloading of TCP processing from servers. For its part, Citrix attributed its subpar results to our test configuration, especially regarding our lack of caching on
Web front-end devices and in client browsers. Fair enough, but other devices did better running these same tests, suggesting
that perhaps there are limits to this otherwise capable performer.
To continue reading, register here and become an Insider. You'll get free access to premium content from CIO, Computerworld, CSO, InfoWorld, and Network World. See more Insider content or sign in.
In a single box, the NetScaler Application Delivery System sums up much of what's right, and some of what's wrong, about the
Web front-end market. The device bristles with virtually every acceleration feature available (see features sidebar), including caching (which we didn't test), support for complex application-layer switching, SSLCi offload and HTTP compression. And it's a strong performer in some tests, scaling to the highest levels possible
when it comes to connection count, TCP multiplexing, and goodput (see scalability test results).
For all that, though, the appliance reached limits in some of our performance tests. In some cases, enabling HTTP compression
actually increased response time, even with very low transaction rates and user access speeds (see services test results). Citrix says compression results are better with Version 6.1 of its software (the company supplied us with Version 6.0), but we did
not verify this. Also, the device does very well in TCP multiplexing tests only when users have longer 'think times' between
page requests. With rapid requests from users, there's relatively little offloading of TCP processing from servers. For its part, Citrix attributed its subpar results to our test configuration, especially regarding our lack of caching on
Web front-end devices and in client browsers. Fair enough, but other devices did better running these same tests, suggesting
that perhaps there are limits to this otherwise capable performer.
< Previous summary: Array product summary results | Next summary: Crescendo product summary results >
Read more about infrastructure management in Network World's Infrastructure Management section.