| Summary: | The article discusses scalable SQL, focusing on the need to scale up relational database systems to internet size. Online social networks are cited as an example. Topics discussed include user load, data load, distributed processing and partitioning. The CAP (consistency, availability, partition tolerance) theory propounded by Eric Brewer is discussed, which argues that no distributed web service can guarantee all three of the CAP variables will be optimally met. The implications for scalable architecture and message delivery are analyzed in terms of global transaction models.
|
|---|