SQL Server 2008 (aka Katmai) allows seamless transition between relational and nonrelational data, enabling users to access documents as data, encode complex XML hierarchies within XML, and query across relational and text data. So how is this achieved?
There is a new FileStream data type which allows large binary data to be stored in the file system and yet remain an integral part of the database with transactional consistency. And yes, you read correct: with transactional consistency!
This enables scale-out of large binary data to be managed by the database and yet stored on cheaper storage and file systems.
Applications can use a new storage type “VARBINARY(MAX) FILESTREAM” to define table columns and store large binary data as files in the file system instead of as a SQL BLOB. Applications can use T-SQL SELECT/INSERT/UPDATE/DELETE statements to query and modify FILESTREAM data.
The FILESTREAM type leverages Win32 rich set of streaming APIs which is proven to be more efficient and faster the larger the data gets.
More on this and other new features of SQL Server 2008 later!
In the meantime, if you are thirsty to see more, then why not downloading the first Community Technology Preview (CTP) of SQL Server 2008 from here or here?
Oh, and if you haven’t guessed by now, Katmai is the code name for SQL Server 2008.