A former Facebook engineer who led the first designs of the social networking giant’s Open Compute Project recently launched a startup aimed at reducing the costs associated with operating data centers — using a method called “community-based analytics.”
According to a February 25 PC World article, Coolan is a service that uses data collected from its customers’ servers to predict failures and prevent outages, allowing companies to save money on data center management costs.
Coolan’s service is still in the beta stages, meaning it may be some time before widespread implementation takes place. However, ex-Facebook engineer and cofounder Amir Michael said he wants Coolan’s service to offer its customers more control over their server equipment — a very similar mission to that of Facebook’s Open Compute Project.
What’s interesting about Coolan is that it aggregates data from all its customers, and applies the techniques it learns from looking at one customer’s servers and data to all its customers’ server issues. If one of its customers experiences an outage, Coolan will be able to pinpoint the issue with its historical data and fix the issue almost immediately. All of this is possible due to an increased openness toward sharing data at many data centers.
“People are starting to feel more comfortable about sharing information about their infrastructure,” Michael explained.
All the data Coolan aggregates is collected over a secure HTTPS connection, PC World reports. Private information about individual data centers is never released or shared, either.
Each year, the computer server industry generates approximately $14 billion in revenue. Coolan’s pricing model for its service is still being hammered out as several companies test its beta version.
But ultimately, Coolan hopes the savings from limited server downtime and optimized hardware use will more than make up for the cost of its service for the companies that choose to use it.