Hi Koen,
I also doubt that database is having problem of scalablity. But I need to
prove it.
so that only i can go ahead. So the only way is to simulate requests to Wt
app.
Is there any way?
What about siege[1]?
[1] http://www.joedog.org/siege-home/
When I tried sending request using JMeter i can see request coming in Wt
app log but work is not getting done (eg: a db insertion)
but when I tested using WTestEnvironment linking libwttest.so its creating
a sesssion inserting data and destroying session
but Its not binding to a ip:port thing.
Am I heading in right way?
Maybe my limited knowledge about web programming is the issue?
Can you make it clear why we cant mae/simulate AJAX request or a browser?
There will be some way Web Browser communicates with Wt. I think If I know
the underlying mechanism then i can do ?
What do you think?
Waiting for your reply even if I asked something stupid
On Mon, Jun 11, 2012 at 5:01 PM, Koen Deforche <[email protected]> wrote:
> Hey Mohammed,
>
> 2012/6/11 Mohammed Rashad <[email protected]>:
> > I want to test loading Wt Application including the load generated by
> > Wt(rendering)
> > Yes I need to simulate actual AJAX sessions.
> >
> > I tried using WTestEnvironment but is not what I really need.
> >
> > One thing I use JMeter is that i can simulate number of users accessing
> it.
> >
> > So consider the simple case:
> >
> > I had a Wt application. It works like a charm online.
> > But when more users are accessing it becomes slow. So I want to find:
> >
> > When the application becomes slow?
> > Is it because of no of users?
> > Is it because of data in db (because my app make a lot of db
> > operations(insert/update/select)
> >
> > If I can simulate no. of users I can find out easily that upto what
> number
> > of users my application will be in good state?
> > There may not be tools available for 100% real simulation but its fine
> If I
> > can get something closer to it.
> >
> > So basically I want is testing load on Wt App which runs on server and
> need
> > to find wheather it really need to be in a cloud
>
> How about the suggested online tools? They have (or at least one of
> them) has a free option which you can use to evaluate.
>
> The overhead of Wt itself is entirely scalable and thus you can easily
> extrapolate from data using a small number of users. I think you need
> to evaluate what in your application is not as scalable, and that is
> typically your database layer. So you could just as well stick to
> plain HTML sessions.
>
> You should look at time used for individual requests (as printed out
> by Wt) -- if this time grows then you have something that is not
> scaling, and you also know that it will limit the possible throughput
> for a fixed number of threads. It might be that increasing the number
> of threads already solves things (removing head-of-line blocking
> because of waiting for I/O from e.g. a database).
>
> Regards,
> koen
>
--
Regards,
Rashad
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
witty-interest mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/witty-interest
