[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Ozone in real applications ?
On Mon, 23 Oct 2000, Maurizio De Cecco wrote:
> Falko Braeutigam <falko@softwarebuero.de> writes:
>
> > It's a very simple solution to provide a "transactional environment for
> > persistent Java objects". (if simplicity is a win for you)
> >
> > ozone's single instance architecture should be a win. (otherwise you shouldn't
> > use it).
>
> OK, this is clear to me; i found the architecture clean and clearly a win for
> some kind of application (including, probably, mine :-).
>
> > > Any benchmark in real situations ?
> > What is a real situation?
>
> This is a good question ;->
>
> For example, in the backend of a web application with a real-life workload.
> (well, still vague, but you get the idea).
>
> Or may be better, benchmarks that show where are the limits of the system
> in terms of size and "transaction" per seconds; ok, again, very vague, but
> anything that can help find out if ozone can fit a given planned work-load.
An implementation of the OO1 benchmark is already part of the package. Code for
OO7 is in the CVS. Both can be used to bench performance of ozone and compare
ozone to other products. However, are OO1 or OO7 real-worls applications? No.
They don't define multi-user scenario, which, for real-world things, is very
important. How does it scale? Under which circumstances? Which JDK? Which kind
of hardware? What kind of fron-end system? etc. etc...
IMHO the only way to find out if ozone meets your needs is to write a prototype
of the system using ozone. Very, very simple of course but reflecting your
particular environment.
And, don't forget to keep us posted ;)
Falko
--
______________________________________________________________________
Falko Braeutigam mailto:falko@smb-tec.com
SMB GmbH http://www.smb-tec.com