Re: Shared Memory, semaphores and a new "box"

Well, I'm not sure of the implementation details (a couple of  
colleagues wrote the application that interfaces with Kannel), but  
it's simple a Java web app running in a tomcat instance making fairly  
standard requests via apache HTTP client to Kannel. We've benchmarked  
it at ~1200-1400 MOs and/or MTs per second concurrently on a 2 x dual- 
core 2.0 Ghz opteron (though only one or two cores were maxed out, of  
course). Like you said, most telco's throughput limits mean that this  
level of performance is more than adequate.

I guess the nice thing about HTTP as a protocol is that it's well  
understood, easy to trace/debug, and there's plenty of tools to proxy/ 
log/mangle the HTTP requests as they come and go :)

On 04/03/2008, at 8:49 PM, [email protected] wrote:

> Hi,
> Im getting interested with this topic. So Giulio what concept do you
> use? Is it a complicated one? Or just using a simple sockopen with
> scripting language? In my cases, currently is still useless to be able
> to inject thousands of MT in a second. Because the telco usually limit
> the quota of MT that can be injected in a second. Other opinion? Many
> thanks for the replies.
>
> Regards,
>
>
> Willy

--
Giulio Harding
Systems Administrator

m.Net Corporation
Level 2, 8 Leigh Street
Adelaide SA 5000, Australia

Tel: +61 8 8210 2041
Fax: +61 8 8211 9620
Mobile: 0432 876 733
MSN: [email protected]

http://www.mnetcorporation.com