ApacheCon NA 2010 Session

High performance cloud-enabled SCA runtimes

An overview of the Tuscany native SCA sub-project, building high performance cloud-enabled SCA runtimes:

The Tuscany SCA C++ runtime is an implementation of a useful subset of SCA (Service Component Architecture) in C++, supporting application components written in C, C++ and Python. The runtime is built on Apache HTTPD as a set of HTTPD modules, and is optimized for clustering, virtual hosting, multi-tenancy and instant configuration and application deployment on cloud infrastructures such as EC2, Rackspace etc.

The SCA Python runtime is a pure Python implementation of the same subset of SCA, optimized to run and scale Python components on the Google AppEngine Python platform.

The project is also implementing a collection of cloud-friendly runtime components, useful to build scalable applications in cloud environments, such as proxying (HTTPD), caching (Memcached), queuing (Qpid/C), Web services (Axis2/C), JSON-RPC and REST/Atom components for example.

We will present the design of these runtimes, and will demo how to develop, deploy and run composite applications distributed across multiple clouds (EC2 and Rackspace using Apache Libcloud, and Google AppEngine). We will also discuss our experience, challenges and lessons learned with these clouds, and the project roadmap.