a design question

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

a design question

mchenini
Hello,

This is a design question: For Lucene to be able to process a million
documents and in the purpose for the search application to be scalable
and still have a good response time do we need to use an EJB container
such as Weblogic or is a Servlet container such as Tomcat sufficient to
do the job? This design should take into consideration remote searching.

Thanks,
Mohamed
====================
This email/fax message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution of this
email/fax is prohibited. If you are not the intended recipient, please
destroy all paper and electronic copies of the original message.
Reply | Threaded
Open this post in threaded view
|

Re: a design question

mark harwood
EJB explicitly precludes you from accessing files, including via third party libraries such as Lucene.

http://java.sun.com/blueprints/qanda/ejb_tier/restrictions.html

In practice you may be able to get away with it but I see no particular reasons why using an EJB server should offer any benefits over a Servlet container.

Cheers
Mark

----- Original Message ----
From: "Chenini, Mohamed " <[hidden email]>
To: [hidden email]
Sent: Thursday, 12 October, 2006 3:25:44 PM
Subject: a design question

Hello,

This is a design question: For Lucene to be able to process a million
documents and in the purpose for the search application to be scalable
and still have a good response time do we need to use an EJB container
such as Weblogic or is a Servlet container such as Tomcat sufficient to
do the job? This design should take into consideration remote searching.

Thanks,
Mohamed
====================
This email/fax message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution of this
email/fax is prohibited. If you are not the intended recipient, please
destroy all paper and electronic copies of the original message.




Send instant messages to your online friends http://uk.messenger.yahoo.com

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: a design question

Bill Taylor-2
IN THEORY, EJB containers are better able than Tomcat to spread
incoming requests over a multitude of servers.  There was considerable
discussion some time ago about index search speed on a single
processor.   I do not remember the details, but there was some
information about how fast a search might be expected to go with one
processor.

If that is fast enough, Tomcat should suffice.

My search application has a few short documents so my experience is not
relevant.

On Oct 12, 2006, at 10:46 AM, mark harwood wrote:

> EJB explicitly precludes you from accessing files, including via third
> party libraries such as Lucene.
>
> http://java.sun.com/blueprints/qanda/ejb_tier/restrictions.html
>
> In practice you may be able to get away with it but I see no
> particular reasons why using an EJB server should offer any benefits
> over a Servlet container.
>
> Cheers
> Mark
>
> ----- Original Message ----
> From: "Chenini, Mohamed " <[hidden email]>
> To: [hidden email]
> Sent: Thursday, 12 October, 2006 3:25:44 PM
> Subject: a design question
>
> Hello,
>
> This is a design question: For Lucene to be able to process a million
> documents and in the purpose for the search application to be scalable
> and still have a good response time do we need to use an EJB container
> such as Weblogic or is a Servlet container such as Tomcat sufficient to
> do the job? This design should take into consideration remote
> searching.
>
> Thanks,
> Mohamed
> ====================
> This email/fax message is for the sole use of the intended
> recipient(s) and may contain confidential and privileged information.
> Any unauthorized review, use, disclosure or distribution of this
> email/fax is prohibited. If you are not the intended recipient, please
> destroy all paper and electronic copies of the original message.
>
>
>
>
> Send instant messages to your online friends
> http://uk.messenger.yahoo.com
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: a design question

chrislusf
I think a standalone J2EE application will be good and better loose
coupling than EJB. You can seperate memory, disk, and CPU resources
from your main application. You can send results back in XML, JSON, or
other formats.

Chris Lu
-------------------------
Instant Full-Text Search On Any Database/Application
site: http://www.dbsight.net
demo: http://search.dbsight.com

On 10/12/06, Bill Taylor <[hidden email]> wrote:

> IN THEORY, EJB containers are better able than Tomcat to spread
> incoming requests over a multitude of servers.  There was considerable
> discussion some time ago about index search speed on a single
> processor.   I do not remember the details, but there was some
> information about how fast a search might be expected to go with one
> processor.
>
> If that is fast enough, Tomcat should suffice.
>
> My search application has a few short documents so my experience is not
> relevant.
>
> On Oct 12, 2006, at 10:46 AM, mark harwood wrote:
>
> > EJB explicitly precludes you from accessing files, including via third
> > party libraries such as Lucene.
> >
> > http://java.sun.com/blueprints/qanda/ejb_tier/restrictions.html
> >
> > In practice you may be able to get away with it but I see no
> > particular reasons why using an EJB server should offer any benefits
> > over a Servlet container.
> >
> > Cheers
> > Mark
> >
> > ----- Original Message ----
> > From: "Chenini, Mohamed " <[hidden email]>
> > To: [hidden email]
> > Sent: Thursday, 12 October, 2006 3:25:44 PM
> > Subject: a design question
> >
> > Hello,
> >
> > This is a design question: For Lucene to be able to process a million
> > documents and in the purpose for the search application to be scalable
> > and still have a good response time do we need to use an EJB container
> > such as Weblogic or is a Servlet container such as Tomcat sufficient to
> > do the job? This design should take into consideration remote
> > searching.
> >
> > Thanks,
> > Mohamed
> > ====================
> > This email/fax message is for the sole use of the intended
> > recipient(s) and may contain confidential and privileged information.
> > Any unauthorized review, use, disclosure or distribution of this
> > email/fax is prohibited. If you are not the intended recipient, please
> > destroy all paper and electronic copies of the original message.
> >
> >
> >
> >
> > Send instant messages to your online friends
> > http://uk.messenger.yahoo.com
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: a design question

Otis Gospodnetic-2
In reply to this post by mchenini
Gecko? ;)
My advice: stay away from EJBs as much as you can.  They are too complicated and too heavy for most systems.  Servlet containers like Jetty, Tomcat, or Resin are often perfectly suitable for the job and a lot simpler.

Otis

----- Original Message ----
From: "Chenini, Mohamed " <[hidden email]>
To: [hidden email]
Sent: Thursday, October 12, 2006 10:25:44 AM
Subject: a design question

Hello,

This is a design question: For Lucene to be able to process a million
documents and in the purpose for the search application to be scalable
and still have a good response time do we need to use an EJB container
such as Weblogic or is a Servlet container such as Tomcat sufficient to
do the job? This design should take into consideration remote searching.

Thanks,
Mohamed
====================
This email/fax message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution of this
email/fax is prohibited. If you are not the intended recipient, please
destroy all paper and electronic copies of the original message.




---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: a design question

Mark Miller-3
An EJB container will generally not give you better performance than a non
EJB container (other than it might be a more efficient...but that will not
be because it is an EJB container). The main difference is that you will be
able to use EJB's and the other Java EE goodies that a J2EE container
provides, like persistence and distributed transaction management. Unless
you are going to go with Glassfish (there may be one or two other obscure
ones) you will not be using EJB 3.0 and it is generally not recommended to
use < EJB 3.0--see Rod Johnson's J2EE without EJB as well as tons of other
books. You are better off going with tomcat and combination of Spring and
Hibernate or something. There is an RMI searcher in Lucene and you should be
able to easily use it with a servlet container. Apaches XML-RPC 3.0 java
extension also has a remote object proxy system and will allow you to pass
serial objects back from remote methods although it does not seem to support
dynamic remote loading of implementations. Jini might also be worth looking
into, although it seems that it is not possible to use Jini in tomcat
securley without modifing both Jini and Tomcat. Glassfish is supposedly
better suited to this.

In the end, a million documents is nothing to lucene if you have even
remotley decent hardware (probably even if you don't). Lucene will spit
around a million documents in her sleep. I'd develop in Tomcat (or similiar)
and check out solutions like Spring and Hibernate as you need J2EE (or JEE)
solutions.

- Mark

On 10/13/06, Otis Gospodnetic <[hidden email]> wrote:

>
> Gecko? ;)
> My advice: stay away from EJBs as much as you can.  They are too
> complicated and too heavy for most systems.  Servlet containers like Jetty,
> Tomcat, or Resin are often perfectly suitable for the job and a lot simpler.
>
> Otis
>
> ----- Original Message ----
> From: "Chenini, Mohamed " <[hidden email]>
> To: [hidden email]
> Sent: Thursday, October 12, 2006 10:25:44 AM
> Subject: a design question
>
> Hello,
>
> This is a design question: For Lucene to be able to process a million
> documents and in the purpose for the search application to be scalable
> and still have a good response time do we need to use an EJB container
> such as Weblogic or is a Servlet container such as Tomcat sufficient to
> do the job? This design should take into consideration remote searching.
>
> Thanks,
> Mohamed
> ====================
> This email/fax message is for the sole use of the intended
> recipient(s) and may contain confidential and privileged information.
> Any unauthorized review, use, disclosure or distribution of this
> email/fax is prohibited. If you are not the intended recipient, please
> destroy all paper and electronic copies of the original message.
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>