Re: Unexpected low CPU usage

> 1. The Apache Commons HttpClient's keep-alive functionality works and,  
> in general, is more stable than Sun's default implementation. The  
> expected usage of this class is to use use a single shared instance and  
> to use MultiThreadedConnectionManager when concurrent requests are being  
> made.

I have no doubts about the quality of Common HttpClient. I just  
encountered an issue when I switched to it. It turned out that the default  
configuration did not match my needs.

> 2. I'm assuming the following:
>         - that you started with -Dhttp.keepAlive=true

No, but this seems to be the default for both the Sun HTTP client and the  
Commons one.

>         - your code (or the code it calls) explicitly calls close() on  
> the HttpConnection's InputStream when it has finished reading the  
> response.

I assume XFire and Apache XML-RPC definitely do.

>         - you are not using HTTPS

No, I am using HTTPS.

> 3. Assuming you are now past the initial problem with excessive  
> connections being left in TIME_WAIT state, you didn't mention what this  
> system will do and how it will be used.

I am building an RPC server that exposes all services on multiple  
protocols so the applications running on it is as open as possible. This  
is not entirely true. Limitations of a protocol may prevent a few services  
 from being exposed on that protocol.

> In your first email you mentioned that your goal is to test the overhead  
> of your framework. That won't be realized with this setup, which isn't  
> really simulating a physical network connection. To understand the  
> overhead the test physical architecture, and specifically the network  
> latency should be roughly comparable. Published studies that compare the  
> performance of Soap and RMI typically report an order of magnitude  
> difference. I don't understand why anyone would choose to use Soap if  
> they have control of both client and server - choosing complexity,  
> bloat, and immature test tools.

I understand I need to run many more tests using different scenarios to  
test all aspects of my server.

RMI is limited to Java. The advantage of SOAP over XML-RPC is that XFire  
generates a nice WSDL which can be used to generate client code in for  
example .Net.

The real reason is I just don't want to force somebody to use a particular  
protocol to be able to use my application. I want the application to be as  
accessible to developers/administrators/hobbyists/... as possible.

> Good luck with your project. If your system has high usage you may want  
> to reduce the time_wait_interval to something better suited for web  
> servers. YMMV

Thanks. I have solved the TIME_WAIT issues also by configuring the  
components better.

As I sent in another post I found the configuration option that allows me  
to disable the delay. I have already run successful tests with my SOAP  
client. It is now doing 800+ roundtrips per second instead of 60+.

Johan Stuyts

===================================
This list is hosted by DevelopMentor®  http://www.develop.com

View archives and manage your subscription(s) at http://discuss.develop.com