I missed it when it was actually announced, but VMWare's acquisition of B-Hive is very big news, IMHO, for the enterprise cloud/utility computing market. (Thanks to GridToday for the heads up.) I'm not a huge fan of VMWare's automation technology to date (DRS/HA scalability is a joke ), but this acquisition may change that perception over time. If nothing else, it signals VMWare's entrance into the "serious" cloud/utility computing infrastructure market. (Still looking for that term...)
Taking a very Cassatt-like approach to policy-based automation (even using the term Service Level Automation throughout its marketing literature), B-Hive Conductor claims to be an all-virtualization approach to the problem. Agent-less from a software standpoint, the B-Hive "controller" is a virtual appliance that runs in your VMWare infrastructure and manages the creation, destruction and replacement of VMs on a policy-based basis. (Disclaimer: I am a Cassatt alumni.)
The result is that VMWare now has the pieces/parts for a serious "fog computing" offering (the term being a tongue-firmly-in-cheek reference to cloud computing behind the firewall). With features like dependency mapping between servers; monitoring and reporting; and SLAuto, B-Hive is for real and should push VMWare into the production data center in a big way. Granted, they will still take the Microsoft modeled approach of an all proprietary suite, but--hey--people still buy Microsoft for production data centers as well, so there is no accounting for taste.
What is VMWare's key advantage? VirtualCenter. Take the following quotes from Forrester's James Staten in GridToday's feature story:
"Forrester’s Staten agrees that the ability of B-hive's tool to tell the load balancer what to do is “very valuable,” but notes that automation on this level is far beyond what most virtualization users would be able to digest at this point. “What we see in virtual environments is they're going up the maturity curve, where we have a large number of customers who are starting to make the move from tactically trying it out to strategically implementing it,” he explained. “And they’re not yet, even in the strategically implementing it phase, ready to start automating it.”This is an excellent strategy for introducing SLAuto into data centers everywhere. The big problem that Cassatt and others have had in speeding up the sales cycle is that SLAuto touches everything, and as such needs to be sold to damn near everyone in IT before a project can be successful. By allowing the end contributors to dabble with the automation on their own, before committing to a production project, there is a much higher likelihood that (someday) automation will be adopted. It doesn't bring the revenue any earlier, but I would think it would significantly bring down the cost of sales.
The key, Staten believes, is for VMware to make its automation tools ready before users get to that point so they can start learning about and trusting the tools. B-hive was experiencing some early success, he said, but “putting this into [VMware] VirtualCenter will make it possible for a lot more customers to start learning about and getting familiar with automation and actually using it.” In addition, he says, the promise of added support and certification, as well as the presence of VMware's large R&D team, will add a level of comfort for prospective users."<
So, is VMWare going to pull away from the cloud/utility infrastructure pack? Perhaps, but not today, or even tomorrow. Give it months, in which time some of its biggest rivals in this space will advance their own technologies. If nothing else, I hope this acquisition will light a fire under the butts of those who figured VMWare would be relegated to "just a feature". Or do the same for those enterprises that fear letting go of minute by minute control to an automation platform.