5 ways to reduce advertising network latency

Project Watch: Microsoft 2008 This project I began writing about in late January, the prototyping of a large database project using the latest versions of a Microsoft software stack, has been an unusual exercise.

Unusual because, from the start, both the executive and the technical personnel were fully co-operative throughout the entire undertaking. The starting point was an existing SQL Server 2005 database running on top of Windows Server 2003.

Both the executive and technical sides agreed that SQL Server 2008 had significant advantages for this particular project. Every group will identify its own killer functionality or functionalities, but for us these were the spatial data types and file streaming.

We were aware that other database engines also offer spatial data types - both DB2 and Oracle do - but a change of engine was not an option. The level of disruption that an engine change would entail was deemed unacceptable, and it seemed likely that the pain would outweigh the gain.

Having decided to prototype the move to SQL Server 2008, though, we then resolved to go the whole hog and move to Windows Server 2008 and Visual Studio 2008 at the same time.

What madness was this, moving to products so new they hadn't even launched? Well, we knew Windows Server 2008 was coming shortly, Visual Studio 2008 was already out and SQL Server 2008 was due with Windows Server 2008. SQL Server 2008 was later pushed back to the third quarter of the year.

We knew if we could make this work, though, then the entire set-up should be stable for the next two to three years at least. The development team was keen to go ahead - we all love new toys, after all - and the executive team reasoned the experience gained would benefit the project over the two to three year period.

Now the project is complete, it will - in the fullness of time - go live with about 1TB of data and to several thousand users. It is being used in its prototype form by a small subset of users who report that it is already delivering invaluable insight into the data.

Where was the horror typically associated with migration, where was the pain? I was expecting - not to say looking forward - to writing about the hair-tearing frustrations and incandescent rages attendant upon working with Microsoft's community technology previews (CTPs) and early releases of new software. But it was not to be.

Apart from the software installation the project has been pretty much drama-free: the operating system hasn't crashed, the database hasn't crashed, stuff just works. As a developer, I am delighted the project has gone so well.

There seem to be good indications that Microsoft is beginning to take its CTPs relatively seriously, at least for its high-end server products.

Of course not everything is perfect - there are bugs, some of the functionality is not yet in place in the CTP software - but there's nothing at all out of the ordinary about such imperfections. The news is all good when viewed in a broader context.

This series of Project Watch was intended to let people observe the migration to an all-new software stack from a safe distance, see what worked and what didn't; and ultimately help them to decide whether to follow suit.

The bottom line from our experience is this: if you are happily using SQL Server 2005 (or some other database engine) and can see no killer features in SQL Server 2008, don't rush to change. If you are tempted by SQL Server 2008's new features and potential benefits as both our technical and management teams were, then - unequivocally - the time is right to go ahead and start prototyping.®

5 ways to prepare your advertising infrastructure for disaster