The Windows Azure Platform contains two storage technologies: relational with SQL Azure, and NoSQL with the Windows Azure Storage Service. These are complementary, in that storage service provides cost-effective, scalable storage while SQL Azure provides powerful query capability.
The amount of data stored in SQL Azure can be increased through sharding, in which data is distributed among several SQL Azure databases. This distribution is based on some value in the data. Although improving the scalability of SQL Azure, sharding does not address cost effectiveness.
One technique to addresses this is to scale-out data from SQL Azure into the Windows Azure Blob Service (or Windows Azure Table Service). In this technique, large data objects are stored in the Blob service and a pointer to the data stored, along with other queryable data, in SQL Azure. This effectively marries cost-effective scalable storage with a powerful query capability...