In its initial phases, MySpace operated with two Web servers communicating with one database server and a Microsoft SQL Server database. The site continued adding Web servers to handle increased user requests. After the number of accounts exceeded 500,000 the site added more SQL Server databases: one served as a master database, the others focused on retrieving data for user page requests. After two million accounts were activated, MySpace switched to a vertical partitioning model in which separate databases supported distinct functions of the Web site. After three million accounts, the site scaled out by adding many cheaper servers to share the database workload.
It eventually switched to a virtualized storage architecture in which databases write data to any available disk, thus eliminating the possibility of an application’s dedicated disk becoming overloaded. MySpace later installed a layer of servers between the database servers and the Web servers to store and serve copies of frequently accessed data objects so that the site’s Web servers wouldn’t have to query the database servers with lookups as frequently