At the recent MIX11 conference, Microsoft announced the production release of the Windows Azure AppFabric Caching Service. The service stores data in-memory to improve performance on Windows Azure and SQL Azure. While this feature been available for Windows Server since July of last year, this will be its first foray into the cloud.
Citing its simple configuration and ease of integration into existing applications, Microsoft says that the AppFabric Caching Service is highly scalable, will cache any type of data regardless of size, and is secured via the AppFabric Access Control Service (which is now available for Windows Azure).
Microsoft recommends planning and implementing the AppFabric Caching Service based on three different scenarios involving data type and access patterns. These are:
- Reference Data
Generally read-only; an example would be a product catalog
Should be stored in local cache as much as possible, and refreshed periodically
- Activity Data
Both read and write; examples are a shopping cart or session state
Should be stored in cache tier since it’s frequently updated
- Resource Data
Shared read or exclusive write; an example would be auction bidding
Should be partitioned across several caches for optimal performance
Rather than being a transparent layer that automatically caches frequently-used data, the AppFabric Caching Service gives developers full control over what is being managed. It also has the ability to store data in the local cache (generally on the web server itself). Since the service is hosted as part of the Windows Azure platform, it’s automatically partitioned and optimized for performance.
The Windows Azure AppFabric Caching Service is scheduled for release at the end of April. Some of the existing features of the Windows Server AppFabric Caching Service (such has the High Availability setting) will not be available in the initial version, but Microsoft is planning to integrate those that are relevant to the cloud platform.