> What I wanted to bring up was that it should be possible to run both in > the same VM, rather than just going for the separate VM > but-we-can-run-both-VMs-on-the-same-machine approach.
YES !!! If it was your point, I just perfectly agree with you. > Yes, but older versions of WebSphere used to run them in separate > processes. WebSphere 3.5 and 3.0 had a JVM option to allow both containers to run on the same JVM. I don't remember if it was possible on Websphre 2.0. > -----Message d'origine----- > De: Alex Blewitt [mailto:[EMAIL PROTECTED] > Date: jeudi 7 ao�t 2003 15:06 > �: Emmanuel L�charny > Cc: [EMAIL PROTECTED] > Objet: Re: Which EJB container? Which Web container? > > > > IHMO, having a Web container running on the same JVM process than EJB > > container should be an option. It can boost performances, but stucks > > you in > > an architecture that is not very easy to manage. > > I'm not sure that I agree with that by default. Yes, it may be > desirable to have several Web applications talk to the same EJB > components, but they can do that by JNDI anyway. > > In situations I've been in where that approach was chosen, they ran > into problems with versioning of the EJB components. So yes, in theory, > it's nice, but in practice having the EJB components and Web components > in part of the same application guarantees that the versions between > them are compatible. Administrators may shudder with the idea of > updating an EJB component (even with a small change) that many web > applications are depending on it may break with the 'newer' version and > be a lot harder to fix (especially if they'd not saved the older > version :-) > > > I like the idea to have a EJB server on a specific machine, and other > > machines around the (internal) net talking to this EJB server. What if > > you > > change an EJB? You'll have to change all the Web Container (and you > > can have > > many Web container talking to a single EJB container). That means > > possible > > mistakes, that could be avoided by isolating this EJB Container from > > Web > > containers. > > You'd have to rebuild and/or redeploy the applications, yes. But I > propose that this situation (where you have multiple Web apps using the > same EJB apps) happens much less often than having a single Web app > using a single EJB set.) > > Plus, the change of a single EJB may still break other Web apps that > depend on it ... but that is obviously a management issue. > > > Of course, they are different situation where having a Web contaienr > > and an > > EJB container sharing the same JVM process could be the best solution. > > I'd propose that it would certainly be the best/default way (IMHO :-) > for running an EAR that contains both WAR and EJB-JAR files. > > > So : make it optional. > > It could certainly come in useful if you want to architect the system > you describe above, though with the obvious management > advantages/disadvantages/caveat. > > > BTW, IBM allows you to do whatever you want : WEB container on a JVM > > and EJB > > container on another, or both containers on the same JVM, it's just a > > parameter to set. > > Yes, but older versions of WebSphere used to run them in separate > processes. Now they do allow running the EJBs and Wars in one process, > and in fact the current best practices suggest to use them in one > process rather than splitting them. > > So, when an EAR contains both WARs and EJB-Jars, deploy them in the > same VM. But allow for a WAR to be run in a separate VM and EJB-Jar in > a separate VM when deployed separately. > > What I wanted to bring up was that it should be possible to run both in > the same VM, rather than just going for the separate VM > but-we-can-run-both-VMs-on-the-same-machine approach. > > Alex. > >
