Thanks for your help. That has helped to clear things up for me. What I really need to find out is what sort of latency there is when using RMI to make calls to the JCR API.
Thanks again, Kris Alexander Klimetschek wrote: > > Don't know about performance differences between of a pure RMI lookup > vs. a JNDI lookup (although JNDI is recommended, because it is the > standard in J2EE systems), but anyway it will only affect the *lookup* > of the Repository interface. Once you have done that in your startup > code, you'll never need to do a lookup again in your application. If > you do remoting, you will be using RMI starting from there, for > getting Sessions and doing all the JCR API calls. > > Regards, > Alex > > On Tue, Jun 17, 2008 at 1:58 PM, krisNog <[EMAIL PROTECTED]> wrote: >> >> Hi, >> >> Thanks for the reply. I understand the fundamental difference between the >> two but I'm more concerned with performance. If you look at the >> jackrabbit-webapp the local.jsp/remote.jsp describe how performance is >> "much >> better" using a local lookup such as within container context or through >> local JNDI where-as using RMI to look up is recommended as a last resort. >> My >> question is what does "significantly better" mean? I understand that RMI >> has >> capabilities that JNDI alone does not but what penalty would I incur if I >> chose to use RMI even if I didn't absolutely require the remoting >> capability. Keeping in mind that everything I will be accessing will be >> from >> within the context of web-containers. >> >> >> >> Alexander Klimetschek wrote: >>> >>> Hi, >>> >>> you misunderstand JNDI. It is just a directory service for looking up >>> objects, not a remoting protocol. RMI however is both a remoting for >>> Java (what you mean) and offers a simple way to address these remote >>> objects. JNDI is supposed to be a superior directory API, eg. it can >>> also address underlying RMI objects among other things. But only RMI >>> is a remoting interface which is available in Jackrabbit. Where did >>> you read about a performance difference? >>> >>> Regards, >>> Alex >>> >>> On Tue, Jun 17, 2008 at 4:08 AM, krisNog <[EMAIL PROTECTED]> >>> wrote: >>>> >>>> I have a question regarding best practices. According to the Jackrabbit >>>> documentation JNDI is preferred over RMI for performance reasons. >>>> However, >>>> I've noticed many Jackrabbit users still opt to use RMI. I understand >>>> RMI >>>> can be used in remote situations where JNDI cannot be used but other >>>> then >>>> that what is the tradeoff if any between RMI and JNDI? Historically, I >>>> understand RMI is slower then JNDI but does that paradigm still stand >>>> or >>>> has >>>> RMI performance increased enough to be comparable to JNDI? specifically >>>> when >>>> dealing with large Jackrabbit repositories? >>>> >>>> Thanks for any insight >>>> >>>> Kris >>>> -- >>>> View this message in context: >>>> http://www.nabble.com/RMI-vs.-JNDI-tp17876722p17876722.html >>>> Sent from the Jackrabbit - Dev mailing list archive at Nabble.com. >>>> >>>> >>> >>> -- >>> Alexander Klimetschek >>> [EMAIL PROTECTED] >>> >>> >> >> -- >> View this message in context: >> http://www.nabble.com/RMI-vs.-JNDI-tp17876722p17891000.html >> Sent from the Jackrabbit - Dev mailing list archive at Nabble.com. >> >> > > > > -- > Alexander Klimetschek > [EMAIL PROTECTED] > > -- View this message in context: http://www.nabble.com/RMI-vs.-JNDI-tp17876722p17898756.html Sent from the Jackrabbit - Dev mailing list archive at Nabble.com.
