The absolute difference in performance between CFMX Enterprise and CFMX for J2EE/Jrun isn't that great. It's more pronounced on WebSphere because of their VM. We're looking at publishing a brief on Jrun, but haven't had the resources to get it done yet.
That said, there are definitely scenarios where the fact that you're deploying as a standard Java archive can get you advantages that aren't available in the standalone product. For instance, all JVMs I'm aware of do garbage collection on a periodic basis (as opposed to incrementally). Typically, when the VM does garbage collection, it halts all processing. As a result, what you'll sometimes see in CFMX is a temporary drop in performance, followed by a return to the previous level of performance. However, using Jrun (or any other supported app server) you can run multiple instances of CFMX on the same server, each of which has its own JVM. If you then balance the load between them through clustering, you'll see a higher more consistent level of performance, since the VMs manage their garbage collection independently. On top of this advantage, you get the additional advantage of total freedom to develop parts of your application in Java, which might be faster than CF for some types of operations. As to your other question, have you seen the architecture white paper we published previously? It goes into some detail about the architecture of CFMX for J2EE. http://www.macromedia.com/software/coldfusion/whitepapers/pdf/cfmx_j2ee_wp.pdf If there are questions in there that aren't answered, I'd be interested in more detail. Phil -----Original Message----- From: Dave Carabetta [mailto:[EMAIL PROTECTED]] Sent: Thursday, December 12, 2002 1:18 PM To: CF-Talk Subject: Re: CFMX Enterprise vs. CFMX for J2EE >As Phil Costa says: > >"-- Improved performance over CFMX Enterprise because we did additional >optimization to take advantage of the J2EE architecture" > >So, yes, there is a performance difference between CFMX Enterprise and >CFMX for J2EE on JRun 4, and this is more noticeable on multi-processor >systems. > Yeah, I caught that too. Based on both of the above statements, is there a performance brief available that compares MX Enterprise vs. MX for JRun? If not, will one be put together? Is there any sort of document that outlines the differences, optimization or otherwise, between the two engines? It sounds as if the brief comparing MX Enterprise to MX for WebSphere helps you "put your best foot forward" in terms of marketing to potential clients (assuming that WebSphere is where you're seeing the greatest performance for MX for J2EE). But it seems like it would be a good marketing tool to show benchmarks as to how well your own products work together (JRun 4 and MX for JRun) instead of relying on a third-party J2EE vendor. I could see this being especially useful for smaller companies that are interested in integrating MX and J2EE, but don't have the budget to spring for a major vendor's license. But, again, I'm relatively new to J2EE servers, so there could be a very simple reason for doing what you've done. Perhaps the two engines are so close that a comparison between MX Enterprise and MX for JRun would be fruitless? Either way, I'd love to hear your thoughts. Regards, Dave. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Archives: http://www.houseoffusion.com/cf_lists/index.cfm?forumid=4 Subscription: http://www.houseoffusion.com/cf_lists/index.cfm?method=subscribe&forumid=4 FAQ: http://www.thenetprofits.co.uk/coldfusion/faq This list and all House of Fusion resources hosted by CFHosting.com. The place for dependable ColdFusion Hosting.
