CloudPhysics CEO and co-founder John Blumenthal will probably cringe at the above headline. His data center analytics platform CloudPhysics, which launched this week after raising $2.5 million from the Mayfield Fund and angels, isn’t exactly a social network. But it’s heavily inspired by the consumer Web. Subscriptions to the beta platform open up next week and offer IT system administrators analyses and predictive modeling around virtualized data centers.
Former VMware executives, Blumenthal and his co-founder Irfan Ahmad watched data centers shift to the cloud without any best practices on accurately configuring these new, more complicated systems.
“Right now changes to a virtual machine occur in the realm of religion — you’re praying what you’re changing will work the way you think it will,” says Blumenthal. “The complexity involved is beyond human understanding.”
CloudPhysics found its inspiration in consumer Web platform companies, in particular Facebook, Amazon, and Google. While consumer Web companies learn and predict behavior based on click streams and the data we publicly share online, CloudPhysics’s algorithm will predict the behavoir of a virtualized data center, starting with VMware.
Inspired again by online social behavoirs, CloudPhysics created clever ways to gain the trust of system administrators who must authorize access to its virtual network. When a data center manager downloads the virtual appliance that provides data to CloudPhysics, an instant analysis of the infrastructure comes on-screen. Think of it like the data center getting a profile page in CloudPhysics’s platform. This profile is made of schemes and diagrams that break down its assembled parts and can be shared with “friends” also on CloudPhysics, a feature still in the works.
Another way CloudPhysics’s is gaining collective intelligence is through crowdsourcing. Launched on Monday, the company invited the data center community to suggest what it calls “scorecards,” basically small problem sets about everyday challenges in virtual data center environments. CloudPhysics already has 641 contributed scorecards.
Naturally all this data belongs to CloudPhysics, which is how it will make money. Blumenthal says the information pulled from actual data centers is often proprietary and will be scrubbed prior to entering CloudPhysics’s platform. Subscribers will get a 30-day free trial to the platform and its scorecards, followed by an option of a basic or premium subscription. The pricing is yet to be determined.
CloudPhysics will use the seed funding to further build out its user interface, which we can guess will again be inspired by the consumer Web.