Archive for the ‘WebSphere’ Category

WSTC Day 5

Friday, April 27th, 2007

I’m sat at McCarran International airport now and can’t wait to get home. I’m so tired now that I may even sleep on the plane (something that I can’t usually manage). There were two slots this morning and I went along to hear what the Distinguished Engineers had to say on the subjects of SOA and ESB. Rob High (SOA Foundation Chief Architect) was presenting on SOA Advanced Technology – Today and Tomorrow and Kyle Brown (SOA and Emerging Technologies) covered ESB Best Practices. The latter provided some excellent advice on getting started, designing and implementing an ESB. Hopefully this will make it out as a whitepaper or developerWorks article.

WSTC Day 4

Thursday, April 26th, 2007

I was up early again this morning to sit the SOA fundamentals test before breakfast. This I duly passed along with the Architectural Design of SOA Solutions test later in the day. So, as with Andy, I can now claim to be an IBM Certified SOA Associate and SOA Solution Designer. It does, however, prove that you don’t need to read or even buy Sandy Carter’s book. In fact, knowing that services should be stateless and that the one answer containing an IBM product is likely to be the right one will get you a long way!
(more…)

WSTC Day 3

Wednesday, April 25th, 2007

A 7 o’clock business breakfast meant that much of the rest of today was spent wandering around in a daze. My first presentation of the day was by Marc-Thomas Schmidt who is the architect responsible for ESB (in all its guises) and Service Registry and Repository. This was followed by an excellent session given by Mike Capern and Ryan Zombo. This covered a selection of use cases from WebSphere Service Registry and Repository Proof of Concepts. It demonstrated that, although still a young product, its numerous customizable aspects means that it can meet most requirements.
(more…)

WSTC Day 2

Tuesday, April 24th, 2007

So, I’m going to use the title whether Andy likes it or not! My day started with a focus on the JMS and MQ bindings in WebSphere Process Server and ESB. Corville Allen from development gave an excellent session on some of the intricacies. I picked up a few facts. For example, the ability to set the SCA_JMS_SEND_DB_ERROR system property so that a ServiceRuntimeException is sent back if parsing fails in the data binding for a response. Also, I was aware that the SCA system bus is now optimized out of a flow using a JMS export and import but apparently if the request message is rolled back then, on retry, the system bus is used so that the failed event manager can kick in.
(more…)

WSTC Day 1

Tuesday, April 24th, 2007

My creative juices aren’t flowing today hence the unimaginative title for this posting (and if I couldn’t come up with a good title today, there was no chance I would be able to be inventive by day 5). Sadly, as an internal conference with significant IBM confidential content I can’t cover everything here but I’ll cover what I can. As Andy has already blogged, someone had the good sense to realise that we are a technical audience and the keynote speech was given by Jerry Cuomo, CTO for WebSphere. I particularly liked his demo around Real Time Java which showed garbage collection leading to breaks in a piece of music played with with a Sun JVM compared to the flawless playback when using Metronome GC in an IBM J9 JVM.
(more…)

Transactional simplicity

Tuesday, April 10th, 2007

It’s good to read articles like this once in a while to remind oneself how much simpler J2EE made transactions.

WebSphere Application Server trace

Tuesday, April 10th, 2007

Daniel Julin is also co-author of another developerWorks article looking at WebSphere Application Server trace. The article describes what trace is, how to turn it on (including the appropriate wsadmin incantations) and how to interpret it. Note that, as the article demonstrates, trace can be very useful in diagnosing some issues, it is really intended for use by IBM service personnel who also have access to the corresponding source code, so don’t expect the trace messages always to be self explanatory. The article also has a plug for Trace Analyzer for WebSphere Application Server, a great tool for viewing and analyzing trace logs (and large ones at that) written by two of my one-time colleagues in the transactions team.

Co-locating applications

Tuesday, April 10th, 2007

There was a time when customers would only ever have a single application running on any given application server instance. In the drive to cut costs, this level of application isolation isn’t one that many customers can still afford. Daniel Julin and Stacy Joines have written a developerWorks article that covers some best practices and techniques for troubleshooting co-located applications.