Project of the Month - April 2008

SugarCRM for WebSphere sMash

Description of project

This project demonstrates how Sugar Community Edition can run on IBM WebSphere sMash, offering users new opportunities to integrate SugarCRM with Web 2.0 and Java technologies.

IBM WebSphere sMash is a new platform based on the Project Zero technology incubator. WebSphere sMash simplifies the deployment of dynamic Web applications using RESTful web service and the scripting languages Groovy and PHP. You can find out more at www.projectzero.org.

The www.projectzero.org site community site will incubate the WebSphere sMash technology functioning as a continual Beta.

Support for the PHP language in WebSphere sMash is provided by a Java-based runtime package from IBM rather than the php.net distribution. This runtime supports almost all the language features of PHP 5.2 and a growing subset of the functions provided by PHP extensions, including all those needed for SugarCRM Community Edition 5.0.

The end result? Customers can now run SugarCRM on IBM's best in class JVM technology. This way developers can easily combine the best components from the Java and PHP worlds by leveraging convenient same process integration between PHP and enterprise Java applications. What is more, developers can tailor and extend SugarCRM more easily with WebSphere sMash visual tooling and conventions rather than writing lots of code.

Trove Info

What inspired you to create this project?

We built the WebSphere sMash product to create an environment where programmers could quickly and simply create agile applications, services and widgets using web 2.0 technologies such as REST, Atom, and the scripting languages PHP and Groovy. We think that running SugarCRM on WebSphere sMash brings some interesting opportunities and we'd like to explore them with the Sugar community.

What business pain points were you trying to solve specifically?

WebSphere sMash is about time to value. Businesses continue to expand their ecosystems & partnerships. These partnerships cause integration work items for IT. There is a significant advantage to the business that can address these work items rapidly. With sMash we can leverage PHP, Java, Groovy and visual composition tools to quickly solve these problems.

What technical pain points were you trying to solve specifically?

Businesses have recognised the benefits of moving to a Service Oriented Architecture but it has simply been too challenging to quickly create applications which consume those services. IBM is systematically Web 2.0 enabling our product line with REST and ATOM. This liberates these products and the content they represent to the web. This combined with WebSphere sMash enables developers to quickly write new applications that interact with enterprise systems. By running SugarCRM on WebSphere sMash we thing that developers will find it easy to integrate Sugar with these services.

What do you want to build next?

To a great extent that depends on the feedback we get and that's a big part of the reason to create the project on SugarForge. We think that the tight integration with Java that you get from WebSphere sMash unlocks a lot of possibilities, especially for businesses with a java background but we'd like to know what you think.

We're also pretty excited about the web based tooling in WebSphere sMash and we'll continue to incubate it on ProjectZero.org. Our goal is to build a capable browser based IDE that can be used to create script assets either visually using tooling like the [flow and feed assembly tools]. We think this fits well with SugarCRM. See the video

We built our PHP runtime from nothing in about a year and we are not done. Running SugarCRM Community Edition and the other PHP applications we already support required us to implement a long list of PHP extensions but there are some SugarCRM Plugins and extensions that still will not run. As we move forward we want to get to the point where they all run and where we can run SugarCRM Professional and Enterprise.

As we built our PHP runtime we knew it would be critical to focus on compatibility with PHP.net ahead of working on performance so although we could have started with a PHP compiler we decided to build an interpreter that would be easy to change and evolve and then transition to a compiler once we knew we had it right. As I write this what we have in ProjectZero and WebSphere sMash is a pretty complete PHP interpreter. In our research labs we are working on a PHP to java bytecode compiler. I'm excited about the outcome when the JVM Just In Time (JIT) compiler can get its teeth into PHP; We have some pretty exciting microbenchmark results in the lab.

As we move forward with the WebSphere sMash project we want to add support for advanced deployment options such as clustering and caching based on our experience with WebSphere XD.