BEA Systems defended its decision to OEM AmberPoint’s web services management product rather than building or acquiring one of its own, though it conceded that services management has become a key activity in any services oriented architecture, SOA.
Speaking to Computer Business Review, BEA director of product marketing Mark Prichard said, This is a technology we are very comfortable with, having worked alongside AmberPoint for some time. It’s a very good fit with what we do and we think our differentiator is the completeness of our offering, and that includes things that are OEMed as well as our own AquaLogic offerings.
Prichard insisted that despite having to pay AmberPoint royalties as part of its OEM deal with them, the product is by no means a loss leader that simply enables BEA to make revenue from other elements of the AquaLogic suite: I don’t know the exact details [of the OEM financial arrangements] but absolutely the product is viable in its own right, it’s not a loss leader, he said.
The day before yesterday BEA announced that it would OEM a rebranded version of AmberPoint’s services management and governance product under the name AquaLogic SOA Management (ALSM), as well as providing first-level support.
BEA has had arms length ties for some time with AmberPoint, as well as with AmberPoint rival SOA Software. Back in November, SOA Software added the capability to anticipate service policies on BEA’s AquaLogic Service Bus. But now AmberPoint is clearly the preferred offering now that it has won the OEM and re-badging deal.
Prichard said ALSM will be available from early May, and that the only technical integration required is some QA work.
At first glance it seems somewhat surprising that BEA chose to OEM a re-badged AmberPoint product rather than building or acquiring services management capabilities of its own, which historically has been its preferred approach.
But there are few independent services management and governance vendors left – Systinet, Actional and Infravio have all been snapped up (by Mercury/HP, Progress and webMethods respectively).
With those that remain independent being increasingly important to the vendors that resell or OEM them – AmberPoint counts SAP, Tibco, Microsoft, Iona, Software AG, iWay, F5 and Reactivity among its list of resellers, OEMs or integration partners – BEA may have shied away from getting into a bidding war for either AmberPoint or SOA Software. It would not want to be left out in the cold if it were to have lost such a war to a close competitor.
However we believe that the days of the likes of AmberPoint and SOA Software remaining independent are numbered. Services governance and run-time management is only a small part of the broader services lifecycle, which is why most vendors in the space have already been taken out.
But for now, we think that they are playing hard to get with the belief that, at this stage of market development, time is still on the side. That is, they’ll command a much better price once the market matures and more customers decide that they have proliferated enough web services that they have to start thinking about managing them.
Significantly, this is a gap that the IBMs and HPs of the world are only starting to think about. When it comes to run time governance, you’d think that this would be a data center play, which is the home base of the Tivolis and whatever BMC, CA, and HP are calling their systems management products these days. But the challenge is that services dictate higher level views than the infrastructure management folks have been able to deliver. Server availability may be part of the story of managing SLAs (service level agreements) for web services, but not the only part.
As IBM’s VP SOA and WebSphere strategy channels and marketing, Sandy Carter, told us last June as IBM readied its own services management and registry/repository product, We went out into the market to see if there were other repositories out there that suited our needs but there was nothing available we would want to buy. We decided we needed to build it ourselves.
BEA, or another suitor will need to pay a premium though if they do decide to acquire in this space, precisely because of the fact these firms have become so important to a wide range of larger players: Mercury acquired Systinet for $105m in cash in January 2006 though it had just 170 customers and would not reveal its revenues or whether it was profitable. Suffice to say Mercury conceded that Systinet would be dilutive to its earnings in 2006.
For now, OEMing in this space may be a good option, but it seems the larger players are looking to see who blinks first before they swoop on the remaining independent ISVs in the space.