| Enterprise Content
ManagementEnterprise
Application IntegrationEnterprise Portals Web ServicesApplication DevelopmentApplication ManagementApplication Servers |
| |
|
Consolidation RulesThe market for Java-based app servers has consolidated dramatically over the last two years--Novell bought Silverstream, Macromedia bought Allaire, HP bought and then decommisioned Bluestone, while Sun bought Kiva and NetDynamics. The leading J2EE players of the last two years or so have been BEA WebLogic and IBM WebSphere. Their dominance is set to continue although Oracle 9iAS is fast emerging as serious competition to these two. Oracle has some natural advantages on its side, in terms of the size of its installed base, but also its ecosystem of partners. JBoss meanwhile is a wild card--a very powerful and elegant open source J2EE server. ![]() This work is licensed under Creative Commons License. |
Application ServersThe market for application servers is heating up, while software license charges for this critical infrastucture component are coming down. The next major force in this market will be the introduction of Microsoft's .NET Server. The Java world is well entrenched though and Microsoft's attempts to force its way into this market should spur further innovation from both the vendor and open source Java communities. Users are set to benefit from this competition in terms of both pricing and functionality. Application servers provide services for managing middle tier transactions, such as database access, security, scalability, and failover. They offer a runtime environment for developers to write to. Traditionally many of these functions were built directly into applications by ISVs and enterprise organizations. The problem, however, is that roll-your-own approaches tend to be far more labor intensive--not only do you have to build these functions but you also have to support them going forward, train people in their specifics, and so on. Java standardization is one element of a broad trend towards infrastructure standardization to mitigate such effects; other examples would be Apache for web page serving and Linux for operating systems. Because Java is a standard, and one that developers really like to use, there are many skilled Java developers out there--around 3m or so worldwide according to most estimates. Of course not all of these programmers are capable of building large scale transactional applications, but that is a still a ready supply of warm bodies for your IT architects to marshall. This broad skills base drives a virtuous circle as more J2EE adoption means more Java programmers and administrators available and more Java-based solutions which means more developers and so on... There is far more to Java application servers than J2EE however. J2EE is pretty heavyweight, and not every app needs its functionality. It is complex and somewhat cumbersome to deploy. That is one reason many organizations deploy applications on servers running Java 2 Standard Edition, especially in the shape of Apache. Apache may have started life as a limited engine for deploying web pages but it has grown into a full service app server platform; it is appropriate when scale up scalability is not as important. It works very well in a load balanced scale out infrastructure such as that deployed by Bear Sterns, the investment bank. There are also some other interesting approaches, such as Improv Software's Cirquet system, which provides a distributed management and development framework based on J2SE. But for large scale heavyweight transactional applications J2EE is still the most likely deployment choice. Microsoft is a little late to the party, and still has some major challenges ahead in convincing organizations that its own app server platform is as scalable as J2EE. Most large scale distributed computing shops grew up on Unix, and tend to dismiss Windows solutions as too lightweight. Whether or not this is assumption is justified, Microsoft needs to change the perception that it can't compete in scalability terms. Achieving some successes in TPC-C performance won't be enough. Microsoft needs to establish scalability narratives for communities to buy into in order to underpin success in the space. Microsoft's massive clout with ISVs means that Windows .NET server will definitely be a success going forward, whether or not Microsoft can appeal to Unix-oriented shops in the short term. There is still a lot of jockeying for position and surprising strategic shifts in the app server market. The announcement that webMethods is to bundle the JBoss open source application server is a clear example of this trend in action. webMethods is not just trying to keep costs down with the deal---it is also gearing up for further competition with BEA as the app server specialist begins to more closely address the integration space, which is webmethods home ground. webMethods in bed with a free open source app server--whatever next? |
| |
|