Business Data Warehouse Appliance

Microsoft and HP announced the release of their latest Data Warehousing focused appliance last week, the Business Data Warehouse Appliance (BDW):

Not to be confused with the BDA (the Business Decision Appliance aka the “PowerPivot Appliance”) this latest appliance is targeted at data warehouse workloads and follows the Fast Track Data Warehouse principals.

This is great move, and I think they have the sizing just right;  Approx. 5TB-8TB of compressed user data (depending on your achieved compression ratio) will cater for a decent proportion of the warehouses and data marts in operation today.

The announcement appears to focus on the fast deployment (they claim it can be installed and configured in around 10 minutes), and that’s pretty impressive but I’d like to know what other appliance specific value they have added to the overall package.  After all, installation and configuration of the appliance is just the tip of the iceberg.


I’ve been lucky enough to be involved in a Fast Track Data Warehouse implementation so I have a couple of ideas (some of which we’ve implemented) that I’d like to see baked in to a Data Warehouse Appliance offering:

  • Operational Reporting (Some of which are already achievable through the SQL Server Management Data Warehouse)
    • Disk and File group usage reports – Help with on-going capacity planning by detailing growth and trends over time for the disk as a whole and file groups associated with each database.
    • Fragmentation Reports – [The methodology surrounding Fast Track has a high emphasis on avoiding and managing fragmentation]  Reports that detail levels of fragmentation at both the physical and logical level would potentially pre-empt any fragmentation related issues.  A kind of combination of WinDirStat and Internals Viewer would be a great graphical representation of the fragmentation at both those levels.
  • Database Administration/Developer Accelerators
    • Database/object Creation functions – I want to be able to create a database and or a file group of a specific size and let the appliance do the work of creating the physical files on each of the mount points for me.  PDW (SQL Server Parallel Data Warehouse Edition) already does something like this that’s baked into the product.
    • Simplified partition management – For example, I’m likely to want to age my data over time and merge smaller partitions into larger partitions (certainly until the maximum number of partitions limit is raised in ‘Denali’), or I might want to remove all the data from a specific partition in preparation for a reload.  Make it easy and handle all the swapping out, multi merging etc for me.
    • Fragmentation Management – For example, I might want to select 1 or more tables and have them completely rebuilt to remove any extent fragmentation.
    • Resize management – My file group is approaching full, I need a bigger one, I want a function to perform that resize in a ‘Fast Track approved’ manner.


    There’s a whole host of other ‘value-adds’ e.g. a ‘Best Practice Analyser’ that could be included as part of a Data Warehouse appliance, and it will certainly be interesting to see how the appliance develops following adoption over the next few appliance updates/revisions.