ActorFx Updates – New features, improved performance and a new plugin to ease deployment to Windows Azure

  • Article

From the ActorFx team:

Claudio Caldato, Principal Program Manager Lead, Microsoft Open Technologies, Inc.
Brian Grunkemeyer, Senior Software Engineer, Microsoft Open Technologies Hub

Today Microsoft Open Technologies, Inc, would like to share the news of a new version of ActorFx, V0.7.

ActorFx provides an open source, non-prescriptive, language-independent model of dynamic distributed objects for building highly available objects, including data structure objects, via a standardized framework and infrastructure. ActorFx is based on the idea of the mathematical Actor Model for cloud computing.

With this release, developers now have a more stable code base (based on the Windows Azure SDK 2.0, a new prerequisite) onto which we will build new features that will make ActorFX a more solid framework for building highly distributed applications.

The new release is based on an improved version of the Actor Runtime’s replication support, and contains improvements in several areas:

  • A new AddOnce method for our List actor.
  • Embraced async methods more fully.  Look at GetActorProxyAsync for a good example.
  • Delay-loaded our JavaScript and IronPython language packs.

We also improved reliability and fixed some issues within the networking layer:

  • Continuations on IActorPromises work better.
  • Significantly better error handling for network problems, including adjusted timeouts & retry logic.
  • Better error handling on the server side for a number of scenarios.

On the deployment side, we now have a way to deploy an ActorFX app on Windows Azure using a new plug-in to streamline the deployment process. The instructions are here.

Note that the 0.7 release also requires that the Windows Azure SDK 2.0 be installed before you download the code from the CodePlex site.

Be sure to subscribe to the CodePlex feed and our Blog to be kept current on the latest developments.

If you are using ActorFx in implementations, we’d like to hear more about how you’re using it, how it’s useful to you and how it can be improved. We are looking forward to your comments/suggestions, and stay tuned for even more cool stuff coming in our next release!