[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
|
[News.eclipse.test-and-performance] Re: What about analytical models?
|
- From: mgn@xxxxxxxxxxxxx (Mike Norman)
- Date: Wed, 21 Jul 2004 15:24:44 +0000 (UTC)
- Newsgroups: eclipse.test-and-performance
- Organization: http://news.eclipse.org
- User-agent: NewsPortal/0.25 (http://florian-amrhein.de/newsportal/)
Alexandre
This is interesting feedback. I understand exactly the relationship you
describe and I think it makes sense for it to be in scope of the T&P
top-level Project. My initial thoughts are that it would be an additional
project, although the models themeselves would be managed in the Hyades
Platform Project so that the necessary cross-model and cross-tool linkages
could be inserted. That said, it would require someone to commit
resources to the additional project and I think we should wait until the
dust settles on the top-level project creation before we go there.
Mike Norman,
Scapa Technologies
(Old) Hyades SubProject Lead
(New) Hyades Test Project lead designate
Alexandre Junqueira wrote:
> Hi,
> I've been reading Hyades documentation and future plans (beyond 3.0) and
> I miss one thing on performance evaluation: a framework for
> analytical analysis.
> As I could see, Hyades has it's strengths in test automation and
> application profiling,
> but doesn't have anything about modeling an application (or it's
> components) in
> a way that provides you a model that can be used to evaluate (or
> even predict) application's
> performance analytically, but only thru measurement and/or simulation.
> It'll be nice if this new project could provide a standard way to
> model such analytical models
> (Queue Network, Petri Network, etc.). Doing so you could provide a
> tool that could be used
> not for testing and profiling, but also to do capacity planning and
> sizing of applications.
> Another nice feature having such a tool is that those analytical
> models could benefit from
> information gathered when profiling the application, as
> identification of resources
> utilization (CPU, memory, I/O), etc. making the models more accurate
> -- and easier to be
> maintained.
> Regards,
> Alexandre Junqueira