I would like to say that in the area of SOAP attachments the DataHandler class's use of memory is very good compared to specifying the return type as a byte[].
To send a 1MB document tomcat grows from ~22MB at startup to ~34MB after one call. And after repeated calls it settles at about 40MB.
To send a 5MB document tomcat grows by roughly the same amount.
Where as with a standard byte[] return type a 5MB document was blowing tomcat upto 125MB.
So thanks for all the good work! Stuart.
Davanum Srinivas wrote:
<rant>
of course they don't want to explain the issues. ALMOST EVERYTIME i ask someone even on this
mailing list to come up with sample/test case to show the problem so that we could fix it, there
is ABSOLUTE silence from those people after that. AND EVERYTIME someone shows HOW WONDERFUL their
soap stack is as compared with, invariable the comparison is with Axis 1.0 or some such ancient
version. (See http://marc.theaimsgroup.com/?l=axis-user&m=105890139828409&w=2) for example. </rant>
Bottom line, If anyone has a performace problem, log a bug report with sufficient information and we will try to address it ASAP.
-- dims
--- Jani_V�is�nen <[EMAIL PROTECTED]> wrote:
Hi,
Has anyone seen any comparison documents between different web services toolkits? I have been training the use of AXIS for few weeks now and I have liked it. But today I heard that I should stay away from AXIS because it has some performance issues. The person who told that, didn't want to explain those issues.
I would like to find comparisons between toolkits and some performance statistics.
So if anyone could help me, it would be really nice.
Jani
===== Davanum Srinivas - http://webservices.apache.org/~dims/
