Gorillas on the move

Looks like I picked a good week to come back from vacation. HP and BEA are each at last unveiling their SOA offerings this week — HP went first Monday, BEA follows on Thursday.

When we issued our SOA Management 2005 Report in February this year, HP and BEA were two of the three heavyweight vendors we said "will make a major impact when they bring finished products to market." Now we're about to find out the nature of that impact. It's not a given that these big players will grab a big slice of market share as a simple consequence of introducing products. Sometimes, big gorillas turn out to be cripplingly sclerotic. Question marks still hang over the relevance, scope and effectiveness of their products — questions that Loosely Coupled will be analyzing over the coming weeks and months both here on the website and in our paid publications. But even if their own sales are low, the impact on perceptions of the maturity of the market and on what customers look for in SOA management products will significantly affect everybody else's sales and positioning.

I think one of the challenges both vendors will find is that one of the key selling points of SOA is that it's a means of breaking away from dependence on individual IT suppliers. That's not how big IT vendors have tended to work in the past. With no middleware platforms of its own, of course, HP can lay claim to be more open to a multi-vendor strategy than most of its peers, but I find it interesting that it seems to be pitching its SOA Manager product primarily at the IT professional services market. This is another form of supplier relationship where enterprises may feel they want a bit more choice and flexibility than they've had in the past. But if they've relied on an IT integrator to hand-build their SOA infrastructure they risk being just as tied in as if they'd plumped for, say, an all-Oracle middleware setup.

One also has to ask what exactly HP has been doing with the technology over the past two years if it's still better suited to pitching at professional services teams rather than directly to customers. Can it really be so difficult for a company with HP's huge resources to create a packaged SOA management product? I guess it's unfair of me to ask that question rhetorically here without first asking the opinion of my colleague Keith Rodgers, who follows HP more closely than I do, or without putting it directly to HP. But it's a burning question for me, and I will be seeking a convincing answer.

As for BEA, what I'll be watching out for is the extent to which its planned Freeflow technology depends on its WebLogic application server. Jeff Schneider made an interesting post back in February titled, SOA, Just An Application Server?: "... note the 'coupling levels' within the J2EE platform and the fact that it was hard coded to support three configurations: one-tier, two-tier and three-tier (not N-Service) ... The J2EE stack fundamentally wasn't designed to support a distributed service network." All the SOA marketing talk coming out of BEA to date has been all about enabling customers to build SOAs using WebLogic as the foundation, which to me seems to miss the most absolutely crucial point of building an SOA, which is to secure platform independence. OK, there are other pragmatic reasons why it might make sense in the short-term to compromise on that ultimate goal and stick with a single-vendor solution for now. But BEA, along with its customers, needs to be planning a time when it can move away from the WebLogic application server. I believe its application development and service assembly/composition tools will prove to be the keys to its ultimate success or otherwise as an SOA vendor, and the extent to which they feature in tomorrow's announcements will be something else to take careful note of.

posted by Phil Wainewright 10:10 AM (GMT) | comments | link