Summer has been hard on my blogging. The sun seems to draw me away from my laptop and onto my mountain bike. Breaking my finger seems to have given me a boost back into the blogging chair (slower typing though) and away from the bike. I have a suggestion to other mountain bikers - don't fall down...it tends to hurt.

Today has been an interesting news day in the OSS world. I have long said that a key weakness of the OSS model in commercial deployment is founded in one of the fundamental elements of the OSS dev model - loosely coupled components. In an earlier blog entry I posited that the tester becomes a celebrity as OSS continues to be commercialized. SpikeSource is my favorite player in this space. The SpikeSource presentations I have sat through are focused on the difficulty in integration of loosely coupled components and the testing rigor required to bring OSS stack solutions up to par for commercial deployment. Of course, if you are to deploy a well tested solution from them or Red Hat or any other OSS player you will notice that the binary licenses and/or support contracts will specify that there shall be no modification of source code (or else the testing is not worth much).  hmmm...

Back to the news of the day, check out this artcile on CNET about the partnership between ServiceMix, Apache Synapse, and Celtix (links in the article). 

         "The goal of the planned alliance, the sources said, is to create a more cohesive integration offering and attract software developers in the increasingly cluttered field of open source..."

If OSS projects are going to be competitive with offerings from companies like IBM, BEA, Oracle and WebMethods (as stated in this particular article), they are going to have to solve the integration problem or get killed on the long-term cost comparisons with their competitors. Religion of licensing, purity of coding practices, rebeling against Redmond all fall to the wayside when a customer starts asking the hard questions about controlling costs. Software licensing is traditionally only 3%-5% of the overall solution costs, so where are the big dollars spent? Integration.  Microsoft has been working on transferring the integration workload onto the technology's shoulders away from manual integration for years. This was not done by accident - it has been done to offer value to our solutions.  I'm not saying we are there yet, but this is the right place to be making technology investments. 

In other news, Sybase CEO John Chen sat down for an interview with CNET and stated someting interesting. They are providing an OSS version of their DB. He acknowledges the fact that MySQL and CA's Ingres were influential in this decision. From what I can tell in my limited view of this is that they are stepping solidly into the "throw-it-over-the-wall" camp vs. the collaborative camp. The idea is to create market momentum in the small/mid-market space in order to drive adoption and then work on a conversion tactic to get some percentage of the free users to move to paying customers.  I think this shows the genius of Mr. Mykos and company and the effect they are having in their space. I have long been a fan of MySQL and their model - although Martin and I disagree on the benefits of his being OSS. I will absolutely say though that the fact that they are OSS is a real driver for them - how it helps them is open to debate.

Mr. Chen made an interesting comment as well. He asserted that the pressures of OSS will drive more value into the DB itself. In fact, that many middleware elements will get pushed down into the DB and that this will drive opportunity for Sybase and thus benefit his shareholders. It strikes me, that if the middleware components get opened and thus commoditized and then moved down into the DB to solve the integration problems they will become less open, not more. Testing will be king in that scenario as layers of abstraction get removed. Given my lack of deep tech skill in the DB world I could be missing something in this. I welcome comments.

 

 

This posting is provided "AS IS" with no warranties, and confers no rights.