A new Gartner report, Trends in Platform Middleware: Disruption Is in Sight, highlights some disruptive elements that threaten the currently dominant middleware platforms. Amongst those disruptive influences, EDA (Event Driven Architecture) was prominently identified as an alternative approach to more traditional component based architectures. The paper's authors see EDA moving from vertical industry support to general purpose business critical application architectures, thereby driving a new wave of adoption:
EDA support will be integrated into general-purpose platform middleware and will become an essential ingredient in most new enterprise-class, business-critical projects. EDA will foster innovation not only in platform middleware but also in packaged applications, development tools, business process management and monitoring, and other software industry sectors.
Combined with the trend towards SOA, these emerging programming models represent a significant threat to traditional middleware vendors. Even with all of the work from vendors to make the Java EE and .NET platforms support the SOA programming model, the authors see both SOA and EDA as key drivers away from these platforms and towards new programming models:
...mainstream programming models' limited support for increasingly popular SOA and EDA styles is becoming more apparent. Conceived in the late 1990s to support Web applications, these programming models have absorbed basic SOA and event-processing capabilities, but these are bolted-on features that further complicate already-complex programming models.
As alternatives, the Gartner report highlights the Service Compnent Architecture (SCA), Service Data Objects (SDO), EDA, BPEL, and Windows Workflow Foundation (WWF) as emerging programming models that will eclipse the more traditional component based programming models that dominate both Java EE and .NET today.
Another significant theme of the paper was the impact of open source, and especially the Spring Framework, on the middleware space. And far from being a lower-cost, lower-capability alternative, these alternatives are emerging as first class options that sometimes exceed the capabilities of the traditional offerings.
Somewhat surprisingly, XTP (Extreme Transaction Processing) was identified as one of the driving forces behind the adoption of both Spring and OSGi. But while the performance and reliability of Spring certainly doesn't hurt, the primary driving force behind Spring adoption is still the simplified programming model it provides. As Java EE and .NET have both bloated in complexity, Spring provides a critical foundation for simplified enterprise development. And Spring is so effective that it is making some wonder if the traditional application server is even necessary:
Although most established Java EE vendors support Spring, its increasing industry adoption is a potential threat to its now-dominant role as this adoption paves the way for alternative, non-Java-EE-compliant platforms.
The open source application servers were also mentioned. Red Hat's JBoss still leads this space, with both Glassfish and Geronimo growing in adoption. IBM's support for Geronimo and BEA's support for Spring were both seen as confirmations of the validity of open source options for infrastructure middleware. This in turn will increase the trend towards commoditization and put further price pressure on traditional middleware platforms.
Lastly, the increasing relevance and adoption of dynamic languages and RESTful services were recognized as low impact trends. While they may significantly change the way developers build applications on top of these platforms, they will probably have less of an impact on the choices organizations make when selecting the middleware itself. Both of these trends also represent the broader industry trend towards simplification that is also behind the success of Spring and the recent direction of the Java EE spec itself.
The full paper is available to Gartner subscribers or for purchase. What do you think about Gartner's points?