The document discusses balancing performance, capacity, and cost for cloud data storage. It notes that most data follows a pattern of occasional reads after initial writes, but some data is frequently read and written. Effective cloud storage needs high capacity and high performance at a lower cost than on-premises storage. While cloud storage was initially just about capacity, it now requires performance for active uses like file synchronization and big data. Performance is needed for cloud computing to be faster and cheaper than alternatives. The document outlines strategies for increasing storage performance like intelligent data placement algorithms and tiering for performance rather than just capacity. This enables cloud providers to reduce costs and increase revenue.