jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From François Cassistat...@maya-systems.com>
Subject Re: JCR-RMI, SPI, WEBDAV, and Sling's REST
Date Tue, 24 Nov 2009 19:46:37 GMT
What's seems to make the JackRabbit RMI layer a little slow is that all jackrabbit operation
on session, node, etc. are passing by RMI and wait for the client to tell the next operation
to do. (This may be acceptable in certain cases, I think it is a great option since it is
really easy to set up.)

But since JackRabbit 2.0 does not support the RMI layer (yet), I am working on my own RMI
interface to discuss with my JCR utility functions I use in my project. I expect it should
give better performances since all nodes/session/properties operations and lazy iterations
will be made directly by the server.

If you succeed to separate the JCR operations from the of your project to add your own RMI
interface, it will make your code more modular and more JCR-independent.


Le 2009-11-24 à 1:53 PM, ChadDavis a écrit :

> On Tue, Nov 24, 2009 at 1:05 AM, Alexander Klimetschek <aklimets@day.com> wrote:
>> On Mon, Nov 23, 2009 at 22:53, ChadDavis <chadmichaeldavis@gmail.com> wrote:
>>> On Mon, Nov 23, 2009 at 2:49 PM, Dave Brosius <dbrosius@mebigfatguy.com>
>>>> Perhaps you are interested in this?
>>>> http://jackrabbit.apache.org/jackrabbit-jcr-rmi.html
>>> I am interested in that.  However, I've read that it's performance is
>>> horrible.  Are you familiar with the performance?
>> It is not horrible by default. Depending on your use cases it might be
>> sufficient.
> I think I'll use the JCR-RMI remoting for my initial proof of concept,
> architectural testing.
>>> Also, in the example code it doesn't show how to properly create a
>>> repository.  I did find some slightly more complete code on the WIKI,
>>> but it uses the TransientRepository, which seems like a tutorial only
>>> type thing.
>> The TransientRepository is not a tutorial-type thing, it's to be used
>> when you want to run Jackrabbit in your standalone application.
> It says that it destroys the repository when the last session expires,
> then recreates the repo when another new session arises . . . this
> seems kind of like a huge performance hit.  Maybe I'm
> misunderstanding?
>> For webapp scenarios there is the jackrabbit-webbapplication
>> deployment which includes a servlet that starts up the repository and
>> registers it under JNDI etc. See
>> http://jackrabbit.apache.org/jackrabbit-web-application.html
>> To play around, the standalone server is a good start. See
>> http://jackrabbit.apache.org/standalone-server.html
>> See also http://jackrabbit.apache.org/first-hops.html and
>> http://jackrabbit.apache.org/deployment-models.html
>> Regards,
>> Alex
>> --
>> Alexander Klimetschek
>> alexander.klimetschek@day.com

View raw message