Just realised that the calculation of the delays only works for small
values of cps - below a few K - I'll fix that later.

S.
On 18/11/05, sebb <[EMAIL PROTECTED]> wrote:
> The nightly code now has an option to set up a Slow Socket for the
> HttpClient Sampler.
>
> See jmeter.properties for details.
>
> S.
> On 16/11/05, sebb <[EMAIL PROTECTED]> wrote:
> > I've now had a look at the HttpClient code, and posted a question on
> > the mailing list - it appears that one can register a new
> > ProtocolSocketFactory, and use that to return a "slow" Socket. Not yet
> > sure if this is in 2.0.2, or only in 3.0.
> >
> > Rather than enable this feature via an extra GUI item, I assume it
> > would be sufficient to switch the code on via a property, so it will
> > apply to the entire JMeter testplan.
> >
> > S
> > On 16/11/05, Peter Lin <[EMAIL PROTECTED]> wrote:
> > > Here is a link to an old thread on /. about simulating different kinds of
> > > traffic, it might save you some time.
> > >
> > > http://ask.slashdot.org/article.pl?sid=04/08/20/2035211&tid=222&tid=185&tid=95&tid=4
> > >
> > > peter
> > >
> > >
> > > On 11/15/05, naveen gs <[EMAIL PROTECTED]> wrote:
> > > >
> > > > Hi
> > > >
> > > > Thanks for the reply Sebb and Peter i will search for the network driver
> > > > and try to work on it
> > > >
> > > > Regards
> > > >
> > > > naveen
> > > >
> > > > sebb <[EMAIL PROTECTED]> wrote:
> > > > As Peter says, it's not easy to simulate lower bandwidth.
> > > >
> > > > One can add delays to the scripts, but the connections will still only
> > > > be active for short periods of time, so as far as the server is
> > > > concerned it will appear to be the user who is slow, rather than the
> > > > network.
> > > >
> > > > One really needs to be able to slow down the actual network 
> > > > reads/writes.
> > > >
> > > > This does not appear to be possible at present using either the
> > > > default JVM or Apache HttpClient implementations, as neither gives
> > > > access to the underlying connection. I'm looking into this further.
> > > >
> > > > I suppose one could use a separate proxy to do the slowing down; not
> > > > sure if there any suitable applications.
> > > >
> > > > S.
> > > > On 14/11/05, Peter Lin wrote:
> > > > > currently that is not possible with JMeter. Measuring the performance 
> > > > > on
> > > > a
> > > > > variety of connections like dial-up, dsl, cable modem and T1 is 
> > > > > actually
> > > > > rather hard to do. Mercury has the ability to simulate lower 
> > > > > bandwidth,
> > > > but
> > > > > I don't know how "accurate" it is. dail-up connections are notorious 
> > > > > for
> > > > > dropped packets, so the only real way to measure real world 
> > > > > performance
> > > > is
> > > > > to get a couple dozen dial-up accounts and test the website.
> > > > >
> > > > > peter
> > > > >
> > > > >
> > > > > On 11/14/05, naveen gs wrote:
> > > > > >
> > > > > > Hi all,
> > > > > > I have recorded a transaction of a web application with 100MBPS
> > > > > > connection, now i want to execute the same transaction with 
> > > > > > different
> > > > Dail
> > > > > > up speed like 128,256,512kbps.
> > > > > >
> > > > > > Is there any option or element in jmeter so that i can provide this
> > > > > > various speed and execute for different Vusers.
> > > > > >
> > > > > >
> > > > > > Thanks in advance
> > > > > >
> > > > > > Naveen
> > > > > >
> > > > > >
> > > > > > ---------------------------------
> > > > > > Yahoo! FareChase - Search multiple travel sites in one click.
> > > > > >
> > > > >
> > > > >
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > > For additional commands, e-mail: [EMAIL PROTECTED]
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > ---------------------------------
> > > > Yahoo! FareChase - Search multiple travel sites in one click.
> > > >
> > >
> > >
> >
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to