Joshua,
What sort of problems are you having?
My experience with LoadRunner (from a prior job) is that it captures TCP/IP traffic and replays it. While this can work reasonably well for some applications that use static data, it becomes rather difficult for dynamic data. While LoadRunner allows for some editing of the data it sends (to make it somewhat dynamic), the control is largely on the part of the LoadRunner driver.
The reason this does not work for Seaside is that almost all of the URLs are dynamically generated on the server, and LoadRunner cannot simulate them.
What sort of performance needs do you have? Using the no-cost edition, Dale was able to get consistent loads of 15 to 50 pages per second (depending on configuration), with peaks higher. It turns out that continuations can consume a lot of disk space so the limit on long-term performance is disk as much as anything else.
As to GemStone/S itself, the whole design is to scale and if you are willing to spend a lot of money on hardware (and some on software), you should be able to handle a few thousand pages per second. We have one existing customer who recently commented, "We easily handle transaction rates approaching 6000/sec with about 8000+ daily users. Our average data center round trip times are in the 2-3 ms range."
What sort of performance are you trying to demonstrate?
James Foster
GemStone Systems, Inc.
P.S. 'Smalltalk' has only one capital letter, the S at the beginning.
On Apr 7, 2008, at 11:39 AM, Xiaoshu Chen wrote:
Hi there,
I have been trying to run a performance test for a client-server application based on SmallTalk+GemStone/S, however so far, we have NOT been able to get LoadRunner tool working for such environment.
Wondering if you can kindly provide any advices on how to do performance test for such applications?
thanks a lot.
Joshua
Locked
