The latest version of Google's SDK (software development kit) for its cloud platform App Engine includes the High-Replication Datastore, which has been generally available, the company said in a blog post on Tuesday.

An experimental version of the High Replication Datastore was introduced in January, following a six-month period during which there were reliability issues with the standard storage option, which Google wanted to solve more fundamentally, it said at the time.

When using the High-Replication Datastore, data is replicated across multiple data centers. This provides the highest level of availability, but comes at the cost of higher latency due to the propagation of data, according to Google.

One of the most significant benefits of the High-Replication Datastore is that applications remain fully available during planned maintenance periods, as well as during unforeseen problems such as power outages, Google said.

For applications that crave more CPU power or memory to serve requests, Google has also introduced two larger front-end instance classes, or virtual servers.

Three classes are now available - with increasing memory, CPU limits, and associated cost. By default, applications use the most basic instance, which includes 128MB memory and 600MHz CPU.

Google has also added new experimental APIs. The Conversion API allows users to convert between PDF, HTML, text and images, while the Logs Reader API lets users build reports, gather statistics, and analyse requests, Google said.

The new SDK is available for Java and Python, and both are designated version 1.6.1. They can be downloaded from Google's website.