During the July 1 Oracle-Bea briefing, I was delighted to see the three crown jewels from the Bea(Tuxedo, Weblogic Server and JRockit JVM) take center stage. I then started wondering about what makes great reliable system software. By System Software, I mean software that other applications and service components build upon.
One obvious answer is that such software has great "RASP" features- Reliability, Availability, Scalability and Performance. It is difficult to build reliable software that scales well and delivers high performance. It takes great engineering and persistence to reach that level.
I am amused-and somewhat miffed- when I see many Silicon Valley companies which are barely few years old, release "Version 10.0" of their software, while world class products like Weblogic Server take 10 years to reach version 9.0.
Long back, someone with whom I was in a relationship impressed upon me the importance of maintaining equnamity in face of the mundane realities of life. The little problems of life, she insisted are no reason to affect the "Zen state of mind"
As I understood the philosophy: Bad traffic, Drivers who cut in front of you, Smelly subway riders should not affect the "Zen state of mind". Niether should one allow "Product Managers" who are quite at sea, "technical leadership" that lacks the crispness of a Martinized suit or Project Managers with warped world view ruffle one's feathers.
But even the most patient mind must sometimes ever so slightly get miffed- I was once testing an agent software. Someone implied that a particular bug was not in the Agent software under test, but caused by "starting and stopping" the Weblogic cluster a few times. I took umbrage, my feathers were ruffled and I lost my "Zen state of mind". I quickly wrote a program, that started and stopped Weblogic cluster repeatedly, without the Agent software. I was able to restart the cluster more than 40 times, without any crash of any kind. On the other hand, with the agent software, after about 2 restarts the cluster would crash. (I now think that the bug in the cluster restart after 40 times was caused by an obscure bug, that my code triggered. )
I think many blogs in SOA world are far too serious. To lighten up a bit, I intend to pepper these blogs with spiritual references, rather like those used in the "Jeeves/Bertram Wooster " stories by P G Wodehouse, to offer perspective and humor.
While conceding that the problem was unlikely to be in Weblogic Server would have shown humility, insisting that the problem was in a proven reliable piece of software showed "Superbia" (depicted on the left) which we all (should) know is a sin.I have unflinching admiration the way Oracle embraces and builds on the best technology from acquired companies- This shows humility, which the scriptures recognize to be a virtue.
While the uninitiated beginner thinks that one can get to Carnegie hall by taking the N, Q, R or W trains, those with vision know that it takes Practice, Practice, Practice.
I will blog on the technical virtues of Weblogic Server, including JDBC connection multi-pools, side-by-side deployment, scripted configuration in a later blog.
PS: I intend to upload a much better engraving of "Superbia" that I obtained from a special exhibition by the Polish National museum at the Zamek Królewski (Royal Castle), during my 2006 visit to Warsaw.
