There are a whole bunch of new white papers detailing new features in SQL2008 (you can see the complete list here) but I thought I’d highlight the BI paper:
It’s the first time I’ve seen a fairly detailed discussion in public of what’s going to be new in SQL2008. Here’s a section on what’s new in the AS engine:
The premise of Online Analytical Processing (OLAP) is that instant access to accurate information enbles end users to answer even the most complex questions at the speed of thought. Thus, the aim to continuously excel in providing even faster query times and data refresh rates is a priority during the development process of any Microsoft SQL Server Analysis Services release, an aim that naturally also has been driving the release of SQL Server 2008 Analysis Services.
SQL Server 2008 includes Analysis Services that enable you to drive broader analysis with enhanced capabilities, including complex computations and aggregations. Analysis Service provides enterprise-scale performance through:
· A flexible caching model. With Analysis Services, you can control how data and aggregations are cached to optimize query performance while maintaining an acceptable level of latency between the cache and its underlying data store.
· Declarative attribute relationships. In an Analysis Services dimension, you can explicitly declare relationships between attributes in a hierarchy. This enables Analysis Services to pre-generate aggregations when a cube or dimension is processed, which improves runtime query performance.
· Block computation. Block computation eliminates unnecessary aggregation calculations (for example, when the values to be aggregated are NULL) and provides a significant improvement in analysis cube performance, which enables users to increase the depth of their hierarchies and complexity of computations.
· Writeback to MOLAP. Analysis Services 2008 removes the requirement to query ROLAP partitions when performing writebacks, which results in huge performance gains.
· Scale-out Analysis Services. A single read-only copy of an Analysis Services database can be shared between many Analysis Servers through a virtual IP address. This creates a highly scalable deployment option for an Analysis Services solution.
The two points here are the biggies. Block computation should significantly increase the performance of many calculations that don’t perform well on SQL2005; writeback to MOLAP should finally solve the biggest weakness of AS for financial applications, namely that writeback performance at the moment is rubbish (maybe PerformancePoint will consider it good enough to use now!).