Sun has announced a new SOA effort named Project Kitty Hawk. It was breezed over pretty quickly in the keynote and press briefing, so some are confused as to what the new project actually entails.

To begin with, Sun has a new SOA Readiness Assessment as part of professional services.

As part of Kitty Hawk, Java Enterprise System is intended to simplify administration, management, security, and provisioning of services in an SOA. For example, a Java Enterprise System registry will provide centralized control of services, versioning, metadata management, services registration, and lookup, according to Sun. Featured components in Java Enterprise System are identity services, Web and application services, portal, communication and collaboration, availability services, and security.

Business integration infrastructure will be featured in Kitty Hawk through Java Business Integration, which is technology based on Java Specification Request 208, for providing standardized integration capabilities.

This announcement comes a day after The Middleware Company announced SOA Blueprints. Sun is participating in this effort and has committed to creating an implementation of SOA Blueprints. Sun was asked about the relationship between SOA Blueprints and Project Kitty Hawk in a press session; Sun responded that The Middleware Company's SOA Blueprints will fit closely with Sun's Project Kitty Hawk, with more details to follow.

Read more: Sun flying SOA effort with Project Kitty Hawk

This strikes me as profit driven, perhaps with an emphasis on professional services for sale by Sun. Sun seems to believe that a knowledge vacuum is emerging around SOA, and that expert consulting could be lucrative. How disappointing that SOA would be generally perceived as difficult. Microsoft isn't claiming that SOA is hard. Nor is Microsoft exploiting SOA as a consulting treasure. Is Sun fumbling SOA?

SOA will be huge revenue generator, not because of technical difficulties, but because it will change business processes and practices. Various workflow orchestration engines will be next milk cow for IT industry.

I generally believe SOA is hard on a process and interface level much more than on a technology level. So I would wonder what would Sun/Mircosoft/IBM etc. add to it? Probably only there products/technology that would ironically often provide more of a lock in than an enabler. Much the same with "BPM" engines. Once actual SOA initiatives manage to get the basic contracts/interfaces right (utterly business driven, I am afraid) and cater for a couple of standard "internal services" like security, auditing etc. the main work is done. I honestly do not believe "blueprints" get you anywhere, or put differently J2EE-blueprints already did more bad than good...