a workflow system for data processing on virtual resources

Extended work on our Job Submission Service has been published in the International Journal on Advances in Software. It can be downloaded from: http://www.iariajournals.org/software/soft_v2_n23_2009_paged.pdf. The paper is titled 'A Workflow System for Data Processing on Virtual Resources' and gives a more in-depth view of the surrounding software architecture.

a case for scala

For quite some time now I have been putting off a rather simple implementation task for exposing HTML representations of entities based on an RDF/XML model in my JAX-RS application. Mainly because it either requires an XSL transformation or parsing the whole graph into an object model and using templates to create the HTML. Both cases are usually time consuming and somewhat unexciting. I have learned to like the idea of being able to use XML as a first-class citizen in the progamming language when I did some work with ActionScript 3 recently. So I decided to finally give Scala a chance and it turned out to integrate really well. The code snippet below actually shows a Scala class extending a Java class and overriding a Java method, both tagged with JAX-RS annotations. Path and resource names have been changed.

@Path("/Resource")
class HtmlResourceController extends AbstractResourceController {
   @GET
   @Path("/{id}")
   @Produces(Array(MediaType.TEXT_HTML))
   override def findResourceById(@PathParam("id") id:String) : Response = {
      val rdfXmlRes = super.findResourceById(id);    
      ...
   }
}

The Array declaration within @Produces is only needed for the scala compiler. Within this method I now get the full and yet easy to use XML processing power of Scala, as described here: Scala and XML.

The drawbacks are the additional dependency to Scala and having to maintain two programming languages within the same project but I can live with that for now, especially as it gives me access to a whole set of interesting language features like mixins, closures, currying, and so forth.

concurrentlinkedhashmap

I would like to call your attention to the following project: http://code.google.com/p/concurrentlinkedhashmap/. It is a version of java.util.LinkedHashMap optimized for use in concurrent environments. It's perfectly suited for the implementation of caches. It supports different eviction policies.

The following discussion thread http://groups.google.com/group/concurrentlinkedhashmap/browse_thread/thread/1b6501f0dede6ff6 is particularly focused on the implementation of the LRU EvictionPolicy. This is also a call for feedback and suggestions for implementation.

a service for data-intensive computations on virtual clusters

We have recently finished a paper that is to appear in the Proceedings of The First International Conference on Intensive Applications and Services (INTENSIVE 2009). Any feedback is highly appreciated.

Here is the preliminary link: A Service for Data-Intensive Computations on Virtual Clusters