I’ve recently returned from IBM Connect 2014 in Orlando, Florida where I presented a two hour masterclass session on XPages Scalability (JMP401: Masterclass: XPages Scalability) along with my colleague and XPages Architect, Martin Donnelly. This session was designed to be a natural follow-on from the previous XPages Performance Masterclass series already available for you to watch here on this blog.
The first half of JMP401 describes the distinct differences betweeen Performance and Scalability – highlighting the fact that they are two separate entities and should be dealt with in that manner. Essentially, efforts to improve one can conflict with the other and vice versa. A brief recap of the Performance area then closes out the first half so as to ensure you are primed and ready to deep dive into the second half – afterall maximum Scalability cannot be achieved without maximum Performance!
The second half of JMP401 then nose dives directly into the mechanics of the XPages Runtime that deal directly with, and have the biggest impact on Scalability. It then shows how to decipher the internal memory footprint of an XPages server – from the applications and user sessions within it, going right down to the individual component tree memory structures for each XPage. Finally, details about the range of settings you can leverage at both server and application level are explained so you know exactly how to fine tune and optimize your XPages box! The session closes out with a test case scenario based upon my XPages Insights Into Big Data application, where two different versions of this application are profiled. One is highly scalable, whilst the other is not! Neatly demonstrating a case and point in how XPages developers / customers fall into common Scalability pitfalls everyday and how to avoid them!