Amazon and Microsoft update cloud cache services

Having launched push notification services on the same day during August, Amazon Web Services and Microsoft repeat the feat on September 4th by updating their in-memory caching services.

Developers typically rely on Memacached or Redis to push key/value pairs into the cache and retrieve them when required. Since this data is transient and doesn’t involve disk I/O, frequently accessed data is populated in the cache. In-memory cache is one of the popular database services on cloud along with relational and NoSQL databases. Instead of setting up, configuring and managing the caching server, developers can leverage the managed caching service exposed by some of the mature cloud providers.

Amazon launched the caching service known as ElastiCache two years ago. On the same lines of Amazon RDS, ElastiCache is available in a variety of configurations called node types. Customers can choose from an entry-level node type called Micro (213 MB of memory) to a high-end node called Quadruple Extra Large (68GB of memory). Right from day one, AWS ensured that the ElastiCache service is compatible with the popular open source caching service, Memcached. This makes it simple for developers to switch from existing caching servers to Amazon’s managed caching service. Following the tradition of AWS services like Amazon RDS that started with MySQL but eventually got the support of Oracle and MS SQL, Amazon brought the support for Redis to ElastiCache. Redis is the next popular in-memory database preferred by many web developers. Developers can now choose between Memcached and Redis when launching the ElastiCache nodes.

Microsoft evolved the AppFabric Cache into Windows Azure Cache to offer that as a managed service. This was primarily available as a PaaS offering which was meant for developers using the Cloud Services feature of Windows Azure. Though Windows Azure added additional capabilities including Web Sites, Windows / Linux VMs, the caching service was confined to the original PaaS layer. With the latest announcement, Microsoft brought the cache service to other compute services making it possible for Windows Azure developers to consume it across PaaS and IaaS deployments. Developers consuming this service can allocate up to 150GB of memory to store the in-memory data. The interesting catch with this cache feature is that it is not binary compatible with memcached forcing developers to use the proprietary API. This will be a big let down for LAMP and open source developers hosting applications on Windows Azure as they need to re-write applications.

As the two cloud factories from Seattle keep churning out new features, developers are eating them up!