We are pleased to announce the release of AHE 2.0, a lightweight middleware tool designed to facilitate scientific investigations using grid resources. AHE provides scientists with application specific services to utilize grid resources in a quick, transparent manner with the scientific objective as the main driver of the activity.

Version 2.0 of the AHE brings us a step closer to making use of HPC grids fully transparent to the end user, by building upon the original AHE services and integrating them with other useful middleware tools, with the focus constantly on simplifying grid and HPC use. Support is included for creating and launching advanced co-reservations of time on machines using HARC, facilitating the launch of computational steering jobs using the RealityGrid steering library, and launching cross-site simulations using MPIg. AHE can launch applications on grid resources running a variety of back end middleware, including Globus GRAM 4 and GridSAM, as well as the Unicore 6 OGSA BES interface. This means that the AHE provides a single interface to a wide range of federated resources provided by different grid projects.

Both client and server can be obtained from the AHE download page. The server is supplied as a preconfigured VirtualBox virtual machine image. For more information and to download the software, visit: http://www.realitygrid.org/AHE

The key aspect of AHE V2.0 is the adoption of a novel approach to simplifying the use of grid computing: grid application virtualisation. The motivation behind this approach is to simplify grid use: a layer of user friendly Web services have been placed between the user and the grid, which hides much of its complexity and provides a virtual, ‘abstracted’ interface for any given scientific application in deployment. In practical terms this means that AHE services take care of the actual process of launching an application, manages the staging of input and output files associated with a simulation, sends files needed to run a simulation to their required destination, and retrieves simulation output to locations easily accessible by the user, for further interpretation.

AHE allows for scientific applications to be easily exposed as Web services, and run on a variety of different resources, from high performance grid machines in different continents, to local clusters and even desktop workstations hosted within a single administrative domain. The AHE does not mandate any required features for the applications it hosts, meaning that legacy community applications, often with many years of development investment, can be easily deployed on a grid and hosted in AHE without further modification.

For full information see: http://ccs.chem.ucl.ac.uk/projects/AHE2_0_full.pdf