Serpens suite for Kepler is a set of actors, composites and workflows, developed by Poznan Supercomputing and Networking Center (PSNC), that provide support for different grid middleware stacks - (e.g. gLite and UNICORE). The extension covers the standard activities related to remote job management.

Serpens contains three modules additional to regular Kepler release.

1) The first module is responsible for cooperation with gLite middleware, the widely adopted grid middleware. Using a set of actors and ready template workflows, you can build solutions ranging from single jobs to numerous range-scan parametric jobs. The actors available in the module lets you:
* List available services for given grid virtual organisation.
* Upload and download many files in parallel.
* Prepare, submit and resubmit jobs.
* Manage jobs.


2) The second module is a bridge to UNICORE grid middleware. A set of Kepler actors perform basic operations using UNICORE services. Their combination gives the possibility to build complex and featureful workflows.

Namely, the actors in the module allow to:
* Query UNICORE registry and obtain information about available resources.
* Upload/download data to/from UNICORE storage.
* Submit jobs into UNICORE sites.
* Check statuses of the running jobs.
* Retrieve Standard Output and Standard Error of UNICORE job.
* Retrieve UNICORE job output generated by application started at remote machine.


3) The last module contains actors useful in communication with Vine::Toolkit services. Vine::Toolkit gathers several grid middleware access libraries under a uniform set of operations.

Again, actors available range from information getters to action executors.


Serpens and its modules are released officially as a 2.2.1 version. Recently, UNICORE module was released independetly as a minor version upgrade - 2.3.0. This new release allows to communicate with UNICORE server on a BES (Basic Execution Services) interface and enhances security.

The module comes with documentation of every actor and examples ranging from basic operations to a full-featured workflow: from input preparation, job submission and monitoring to the final step of output retrieval.