I knew it was coming, but I am greatly saddened that this time has come.
In an article on Forbes.com Monday, William "Bill" Coleman--former Sun executive, and the "B" in BEA--let it be known that his latest venture, Cassatt, is close to shutting its doors for good. With heavy investment from the likes of Warburg Pincus (reportedly well north of $100 million), Cassatt describes itself as "champion(ing) the vision of managing data centers like a 'compute utility.'"
Cassatt has been building and selling policy-driven infrastructure automation of various sorts for most of six years now. In fact, I would argue that much of the world's marketing material around data center optimization, dynamic scaling of applications, and automated failure recovery owes its roots to a vision that Coleman and Cassatt have been expressing since it was founded.
I was lucky enough to get involved with the company in 2006, when I was hired to run sales engineering for the West Coast. I had just worked on Sun's Web sites, which at the time were operated in a highly manual fashion. Several years prior to that, I had worked for Forte Software, the distributed application development and run-time environment, where I had seen the power of "drag and drop" resource deployment and scaling.
I wasn't sure what to expect when I came in for that first face-to-face interview, but as soon as I saw a demo I knew that what Cassatt was doing would alleviate a lot of pain in both the deployment and ongoing operations of Web applications and other highly scalable distributed applications.
Cassatt's technology was (is) very powerful and feature rich. (Early on the user interface was a little clunky, and you had to drop into the Linux command line much too often, but those problems were addressed in later releases.) So why is the company failing?
The answer is in the pervasiveness to IT of a truly end-to-end real-time infrastructure. The platform that Cassatt built to halve the cost of running your data center also required you to change the way you provisioned, managed, and consumed everything in your data center. Every branch of IT--server, storage, networking, arguably even facilities--was forced to change their processes and skill sets to both embrace and trust policy-based automation and dynamic provisioning.
I noted the disruptive nature of this cultural change two years ago, in a post where I pointed out that those pretty little static network diagrams--in which every server, every IP address, and every network connection were carefully mapped--are history in a real-time infrastructure.
In the end, though the pipelines were always big, the deals dragged on for months and months and often failed to close in the end. Coleman himself acknowledged as much in the Forbes.com article:
"What frustrates me is my own naivete," he says. "I thought I could give companies something radical that had a proven return on investment, and they would be willing to change all their companies' computer policies and procedures to get that. Right now it's hard to get people to get beyond proof of concept tests or a data center energy analysis."
I ended up leaving the company in 2008, however, frustrated with the lack of traction this "go big or go home" approach was getting.
There is a lesson there for those who think they can just waltz on into the enterprise market and take it over. It's not that easy. HP, IBM, and Sun have all been trying to sell "utility computing" for years, with mixed success. If you want to win in private cloud management, give customers a way to start small, and grow into the concept. All or nothing is clearly not the way.
I am very curious as to where the technology ends up, now. There was some very cool research going on, and I heard through sources that a new generation of the platform was in a demonstrable state. The real value of the company is Coleman, Rob Gingell (the CTO and VP of Development), and several of the top engineering, marketing, and field staff, though. These guys know IT utilities, and have some amazing ideas on how to simplify the operation of an enterprise IT infrastructure.
Many thanks to Bill, and a tip of the hat in appreciation for the ambitious vision that Cassatt attempted to achieve.