What are new features offered in SQL Server 2008 ?

SQL Server 2008, code name Katmai, will be officially released on 27 February 2008. There are a number of changes that will impact on database design and some performance gained with new feature. Some aspects will need you consideration below :

1.       Consider to use latest hardware, such as dual core or quad core processor, and 64 bit machine that has maximum memory 1 TB instead of 32 bit that has maximum 4 GB memory. And use 64 bit SQL Server 2008

2.       Consider to use 64 bit latest server operating system (windows 2003 server SP 2 or windows 2003 R2). Those OSes will support SQL Server in high load utilization. If you install Scalable Network Pack for TCP Offload engine, it will free up CPU cycle. If you use Windows Server 2008, it will have capability to send dynamic frame sizing on TCP/IP low level.

3.       Database mirroring on SQL Server 2008 will be much better compared to SQL Server 2005. It will be easy to implement and will compress log stream over network, resulting efficient bandwidth utilization, that will impact on system performance overall.

4.       SQL Server 2008 will use data compression on LOB data type such as IMAGE and TEXT, less physical I/O has to be done, improving performance.

5.       Data compression will also be applied to backup so it will reduce storage requirement.

6.       Regard to store LOB data into database, SQL Server 2008 will introduce new data type FILESTREAM to address performance overhead imposed on store LOB into database. LOB will be keep externally and FILESTREAM will give high performance accessing the LOB outside database because it integrates will Windows file system whilst still maintaining database integrity.

7.       SQL Server 2008 introduce some datetime datatype, namely DATE, TIME, DATETIMEOFFSET (datetime datatype with timezone)  and DATETIME2 (larger year and fractional second range).  It will make datetime related data migration easier from other database platform such as Oracle and DB2.

8.       If you write some queries that involve some grouping with refers to same table, you can just write only one query with GROUPING SETS new TSQL feature of SQL Server 2008. It is simpler and fewer scan on base table, resulting better performance..