During MicroStrategy's annual European user conference in Barcelona, Guy Bayes, head of Enterprise BI at Facebook, spoke on the topic of Big Data at Facebook. He discussed the technical challenge caused by the massive scale of the data generated by its 1.1+ billion members. To interactively analyze this data from any dimension, Facebook engaged MicroStrategy to create and test a new massively-parallel in-memory analytic technology which MicroStrategy calls "EMMA" (for Extended MPP Memory Architecture). The project has been underway for one year and the current prototype configuration is successfully analyzing just over half of the Facebook dataset and is achieving an average response time of under 5 seconds. At full configuration, the system will run in a cluster of hundreds of commodity servers containing over 10TB of data in memory.
