The Infiniflow Enterprise Service Bus (ESB) Runtime Service leverages the popular Mule open source ESB, enabling ESB messaging and connectivity across the Infiniflow Service Fabric and providing enhanced scale and resilience to Mule ESB installations.

The ESB is a key element in any service-oriented architecture (SOA), but Forrester and other industry experts caution that an ESB is only one piece of an enterprise-wide SOA strategy. The ESB can be thought of as a messaging bus which becomes part of the platform and provides access to different services, but most companies are aiming to build a flexible SOA infrastructure and it is here that the combination of Infiniflow and Mule ESB offers rewards.

Enhancing Existing Mule Implementations

Infiniflow provides an enterprise class, model-driven, distributed, dynamic, resilient and scalable runtime platform. As a result of OSGi™-enabling Mule ESB, it can now be deployed on the Infiniflow Service Fabric, allowing your Mule ESB messaging implementations to automatically benefit from added scalability and resilience.

Providing Off-Fabric Connectivity to Existing Applications

The Mule ESB can also be used for Infiniflow Service Fabric implementations to provide access to off-Fabric applications required for an end-to-end system. Using Mule ESB in this way takes advantage of its extensive support of transports and integration protocols.

Infiniflow ESB Foundations

The Infiniflow ESB Runtime Service provides a supported implementation of the open source Mule4Newton project that is freely available for download and evaluation from the www.codecauldron.org community site.

Mule4Newton

Mule ESB is not fundamentally designed to run on a multi-node distributed environment, so the Mule4Newton project is an excellent illustration of how to migrate from single node or ‘cluster' thinking, into the far more exciting and beneficial world of resilient, scalable and adaptive distributed systems.

The guides and downloadable demos on the Mule4Newton website use Mule and Newton to demonstrate:

• Deploying OSGi bundles into arbitrary topologies and managing them at runtime,
• The basics of the configuration mechanism and deployment units,
• Building a simple adaptive system which scales as more servers are supplied to it,
• Running Mule as a stable n-tier application using Newton .

The demos assume you have a good knowledge of Maven and are well acquainted with Mule, but are very straight-forward and easy to follow, using pre-built projects. Step by step, the demonstrations take you through:

• Downloading and building Newton ,
• Downloading and installing the demo application,
• Producing OSGi bundles and the dependencies for a Mule project,
• How to incorporate Eclipse, Intelij 7 and Felix where appropriate,
• Implementing a Lifecycle Component Listener,
• Deploying single and multiple instances to Newton ,
• Producing n-tier and other topologies,
• Configuring the distributed system to automatically deal with node failure, increased loading, the addition of resources etc.

As a result of working through the demo, you will be able to produce n–tier (and other) topologies for your Mule based applications, run multiple Mule applications on the Newton ‘fabric' without fear of classloading issues thanks to OSGi, and deploy updates to multiple nodes with a single command.

What's next for Mule4Newton?

The current guides and demos just scratch the surface of the potential of Newton as a basis for enterprise deployments. Future guides and demos will look at how to create dynamic topologies that change with loading, how to do efficient load-balanced messaging across an auto-scaling deployment environment, and adding support for referencing Newton services in Mule (SCA integration).

On July 2nd 2008, Mule4Newton project lead, Neil Ellis, will present a webinar exploring the advantages of building an open source ESB-based messaging solution on the Newton/Infiniflow distributed OSGi runtime. Click here to register for the webinar.

Back to newsletter