While I need to get back and read the rest, and to do so more thoroughly, I just did a quick scan through the Tompkins piece on BEA and wanted to share a couple of thoughts.
1. The application server market is extremely new, particularly with respect to any volume deployment. Moreover, it is one that is changing very rapidly, e.g., the big hiccup in products caused by J2EE and which probably will be caused by Java.next, when that gets defined. This makes me inclined to be extremely cautious about annointing anyone as the likely winner since each year's new sales are likely to total several times the total installed base at that start of that year. More than once already within application servers, not to mention other areas, we have seen bright promissing leaders fall by the wayside and virtually disappear ... anyone remember Apptivity?
2. I am still looking for what it is that they have that no one else does. For an enabling software gorilla, it seems to me that there needs to be some compelling advantage which that gorilla has and others don't and which they can't easily acquire through normal means. I made the case with Actuate, for example, that their clean, componential, object oriented technology meant that they were in a position to rapidly evolve their product while potential competitors had large, monolithic, non-OO designs which made rapid evolution of the product very difficult. With appservers, however, it seems that we have a common standard in J2EE of what it is that they need to do and that the appserver itself is simply a question of delivering on that specification with good performance. One company may be a bit ahead of the other simply by virtue of having the more recent release.
One area I see for possible differentiation is in the area of whole solution. In this respect, all BEA really has is Tuxedo, which while well known strikes me as yesterday's solution and a very narrow addition. IBM and Sun/iPlanet certainly have much broader offerings, iPlanet in particular. This may or may not be as big a factor as it has been in other markets, however, since there is a strong tradition in the Java market of components, i.e., buying the best piece from a wide range of vendors. I'm not making predictions yet. |
