Could not post the details yesterday coz I was tired attending many sessions. Here are the major points from the sessions I attended

SQL Server 2008 -Beyond Relational Data by Praveen Srivatsa, Asthrasoft Consulting

XML Feature
We deal with most of the data that is Non Relational in nature

SQL Server 2008 has support for non relational data and it provides various data types

You can use XQuery with Xml type. This allows you to specify a schema as well so you have some control over the type being stored. In general the data can be stored in XML type and retrieved as a relational model or vice versa

Mapping of new date/time data types to xsd types

CLR Feature
Lots of queries around this feature from many people. And I got scared about this atleast from the Support aspect of it . Developers wanted to put many things as a CLR object. And below is my thought about this

Only 13 assemblies are allowed to be used called 'Blessed Assemblies'. UNSAFE load of an assembly can cause SQL Server memory corruption since it loads it inside MTL which is used for by sql server for external hosting of COM,CLR and out process providers

Filestreams
Benefits of filestream against BLOB types explained. Books OnLine(BOL) has more details on this.

ADO.NET Data Services Futures by Harish Ranganathan, Developer Evangelist

This talk was about ADO.Net Data Services offline still under development. Harish showed a smart client which can work offline. This is pretty pre-beta state. There were a lot of queries but most of them I believe would be clear if one goes through the ADO.Net synchronization services like one of the query was how does this handle a conflict of column  here is the MSDN page for this http://msdn.microsoft.com/en-us/library/bb726002.aspx 

One of the interesting questions asked in the session was :-

Can we keep version based copies of the modified records ? As of now I don’t this is thought about , but you can give your feedbacks on the Microsoft Connect site .

Modern Data Access Patterns and Implementations by Don Smith, Program Manager,P&P

This talk was an architect level discussion around the new patterns that should be incorporated for data access in particular with the DDD(Domain Driven Design). There is a lot being discussed about this in the P&P group and they are working on few patterns for data access, the code for which will be released once the VS 2010 beta is available. The pattern thats being explored are

• Unit of work
• ActiveRecords
• Repository pattern
  

Check the codeplex site for more details http://www.codeplex.com/dataguidance

Building scalable applications using Velocity by Abid Khan, Program Manager, Velocity

Velocity is the new kid on the block was distributed caching. Many cool features in particular the load balancing. One of the internals is that this was purely build using .NET FX and leveraging the WCF as the backend for query over NamedPipes channel. Pretty much utilizes most of the concepts from clustering world and automatically detects any offline and online of hosted velocity server. Key takeaways

• As of now it does not invalidate the cache and refresh itself from the database. Going forward this will be thought about
  
• Please use a dedicated servers for velocity rather than clubbing it with the SQL Server or Web Farm
  
• Velocity can be enabled for High Availability where in it will replicate the data across different hosted servers and thus available any time
  
• Can be used in any .net client

Vision/Road Map

• Being able to provide LINQ interfaces to query instead of the traditional get/put methods
  
• Support for non .NET clients like PHP
  
• Availability for HPC (High Performance Computing)

More details here at MSDN site http://msdn.microsoft.com/en-au/data/cc655792.aspx

until next post :)