Google I/O kicked off yesterday in San Francisco, with updates to App Engine, GWT and the company’s Prediction API announced.

Reminding everyone that Google is about more than Android, the company released App Engine 1.5.0. With this update comes support for extended back-end processing of sustained workloads, as well as support for the Google-created Go programming language.

Previously, App Engine was not equipped to run extended processes on the back end, and was instead focused on offering scalable front-end experiences. With the release of App Engine 1.5.0, developers can now access pull queues and improvements to task queues. Previously, App Engine had processes that were limited in resources and runtimes. Google has now figured out how to have long-running, scalable back ends that are useful for compute-intensive applications or applications with persistent states.

App Engine also took on limited map/reduce functionality for in-memory use, with full, platform-wide map/reduce to be offered later this year. Last year, Google announced App Engine for Business, which was a set of additional services offered for enterprises using App Engine. Yesterday, the company announced the termination of that service due to pricing and payment conflicts with the way businesses expected to pay for cloud services. Instead, Google will offer the services designed for App Engine for Business to all App Engine users, instead of reserving them for a specific subset of paying customers.

And now all customers can pay for App Engine hosting. Google announced the availability of service-level agreements and support for App Engine, with a guarantee of 99.95% uptime.

App Engine also added support for its third programming language: Go. This language has special capabilities within App Engine: Developers can write Go code and upload it to the Google cloud, where it is compiled and hosted without the need for compiler or environmental configuration of any kind.

The Google Eclipse plug-in was also updated to make it easier for developers to pull in Google’s APIs, and to use code hosted in Google’s SCM system, Google Code. The update to the Eclipse plug-in also made it easy for developers to build applications that are presented on Android, with back ends hosted in App Engine. App Engine now can host multiple back-end application infrastructures that can be interfaced through a single front end.

Google’s Developer Storage system was also unveiled. This data storage platform gives developers access to the same infrastructure Google uses to host its own information. The service is now open to the public for the first time, and it supports OAuth 2.0 for identity management and authentication.

For the Prediction API, the most significant addition is a gallery of predictive models. These models will offer capabilities ranging from checking e-mails to make sure they’re going to the proper recipient, to analyzing the sentiment of feedback to determine customer satisfaction.

Google Web Toolkit was also updated, though the 2.4 version that was offered online during the show was still in beta and will remain so for some time. With this version, Google has expanded the capabilities of client-side caching, and it has added further compiler optimizations. When compiling GWT code, the compiler will now cache the portions of the application that have not changed, thus speeding up the iterative process by eliminating the need to recompile portions of an application that have not changed during a development cycle.