[Disksim-users] Simulating multiple processes (applications) with trace-driven simulation

[John Bucy]

I think the way I'd do this is use a lightweight threads package like
gnu pth or "state threads", run disksim in one thread through the
interface and model each of your processes with a separate thread.
You'd have to write the glue to synchronize all this but I wouldn't
think it would be more than 100 lines of code or so.




john


On 5/19/07, Yipkei Kwok <yipkeikwok at gmail.com> wrote:
> Hi,
>
> My goal is to simulating the I/O system activities under multiple
> (simulated) application workloads. We have looked at the
> synthetic-driven simulation and, now, we want to look at the
> trace-driven simulation.
>
> I have checked the format of the 3 types of traces: ascii, validate
> and raw. None of them allows specifying the generating process
> (application) of each request.
>
> According to the user manual, at most 1 trace can be used for each run
> of simulation. I am wondering if it is possible to supply n traces to
> the simulation if I want to simulate n processes (applications). Or,
> if it is possible to tag the process ID of the generating process to
> each request in the trace.
>
> Looking forward to hear from you. Thank you!
>
> Regards,
> Yipkei
>
> --
> Web site:
> http://360.yahoo.com/yipkeikwok
> http://www.mcs.csueastbay.edu/~ykwok/
> email addresses:
> yipkeikwok at gmail dot com
> ykwok2 at utep dot edu
> ICQ UIN:    2309842
> Google Talk ID:    yipkeikwok at gmail dot com
> Yahoo! Messenger ID:    yipkeikwok
> MSN Messenger ID:    yipkeikwok at hotmail dot com
> _______________________________________________
> Disksim-users mailing list
> Disksim-users at ece.cmu.edu
> https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
>