Welcome to the inaugural BriefingsDirect SOA Insights Edition, a weekly discussion and dissection of Service Oriented Architecture (SOA) and related news and events with a panel of independent IT industry analysts and guests.
Yours truly, Dana Gardner, principal analyst at Interarbor Solutions, is your host and moderator. The idea with SOA Insights is to bring to the thought leadership market a more focused IT podcast. The market has some great IT podcasts, and we've got a lot of discussion of Web 2.0 and social networking trends, but as far as I know, we don't have any SOA-focused weekly recurring shows.
More and more SOA news, as well as trends, are emerging. This is not just a blip on the hype curve, although there is certainly some hype involved. This is a long-term (perhaps decades long) trend in IT. So, we're going to create our own weekly podcast show on SOA, SOA governance, and related subjects. We're going to have industry analysts joining the show as our panel contributors, and we're going to invite some guests; people that are newsmakers as well as other leaders in this emerging SOA space.
This week's panel consists of Steve Garone, a research partner affiliated with Ideas International, and a former program vice president at IDC. Also joining is Neil Macehiter, research director at Macehiter Ward-Dutton, formerly of Ovum.
This week's topics are BEA's announcements from its recent conference, a slew of IBM SOA-related announcements, including the WebSphere Service Registry and Repository, and Borland's Lifecycle Quality Management (LQM) solution.
Here are some excerpts from the show:
Will it be better to be best-of-breed or will people gravitate to particular dominant platforms? I think the jury is still out on that. It entirely depends on the scope of an organization's SOA ambitions. For example, if they're looking to address some particular project-level server requirements, all departmental solutions and departmental requirements, and they are adopting a SOA approach, they might as well turn to a best-of-breed vendor.
Vendors like Sonic or SOA Software have been quite successful. When organizations start to consider more enterprise initiatives, who are they going to turn to? I think their natural inclination is going to be to gravitate to the big players and then look for the gaps. There are undoubtedly gaps; for example, products of the IBM announcement. Organizations would be looking for a combination of IBM and a Systinet (Now Mercury and HP) to provide registry and repository. Similarly, they will be looking at the BEA offering and looking for gaps there. How well do those BEA [products manage] identity and security? Well, only to a certain extent, so again they're going have to look for best-of-breed capabilities there.
The enterprises will gravitate to the big platform providers and then plug the gaps with best-of-breed capabilities. We will not see this sort of wholesale method of the best-of-breed solutions to implement, to support our SOA initiatives. If you are doing it at enterprise scale, potentially there are quite significant risks. The organizations will fight, plus many organizations are grappling with their sourcing strategies and looking to consolidate suppliers. If you have a massive, significant investment in SAP or ERP backbone, it is likely that you're going to at least consider SAP NetWeaver as part of the solution, and look at plugging the gaps that exist in NetWeaver around things like governance and identity and security.
organizations need to be thinking about governance more broadly than many of the vendors have been promoting. [Governance has been] primarily focused on the development ... side, and to some extent paying lip service to the fact that those governance assets -- the assets you have in your registry-repository -- need to be exploited throughout the entire lifecycle.
This is why the acquisition of Systinet by Mercury and then subsequently by HP is quite interesting. And thinking about the gap that needs to be plugged there. For example, how does SOA registry-repository compare to a configuration management database (CMDB), which is at the heart of their management and monitoring capabilities? I absolutely agree with the analysis in terms of interoperability that is needed from the development side, but I think you equally need interoperability at the management and runtime because it goes back to this point: A service is something you experience. You don't actually experience the service when it's being developed and deployed. You experience it when it's being run and operated.
That is the other angle of governance that we need to highlight. It was interesting, for example, when IBM announced the WebSphere Service Registry and Repository product that they did not talk too much about the integration, and how activity assets, activity complexity, and application manager are going to be exploited.