hello..

I just read about POE very recently and I am intrigued ;)

I have a sort of open ended question or point for discussion if anyone wants to 
comment.  Appologies if this is not the appropriate place to ask.

I am building a load testing tool for web/http.  I currently have a tool developed 
that does functional/acceptance testing (www.webinject.org).  I am in the process of 
designing a future version that will have load testing capabilities.  It will send 
many HTTP requests to a server.  It is important for me to have a tool that is as 
scalable as possible.  I need to be able to launch MANY http requests at once and 
manage them separately (unique cookies, timers, and response validation for each 
"virtual user"), without blocking between them.  

My initial design idea was to use forking.  A main process forks many children that 
each go off an do their work (sending/receiving http and dealing with responses).  I 
am just wondering if anyone has any thoughts on how POE would compare to forking in 
terms of performance and scalability?  I know this is sort of an impossible question 
and the answer is most likely "it depends", but any initial thoughts would be great.  
It looks as if POE does not fork, but just uses non-blocking event loop.. correct?  
Any clue on how this compares to forking?

-Corey Goldberg




-----------------------------------
 Corey Goldberg
 [EMAIL PROTECTED]
 [EMAIL PROTECTED]
 [EMAIL PROTECTED]
-----------------------------------

Reply via email to