Grid Middleware in the NC BioGrid
Perhaps the biggest challenge of the NC BioGrid project has been to identify the appropriate grid middleware to build our infrastructure around. This has led us to evaluate numerous grid platforms (see our Grid Platforms page for some of these). As might be expected, we have not found one solution that is a clear winner in terms meeting all of our requirements and objectives. We therefore decided that a hybrid solution of multiple grid middleware platforms would best serve our needs until the time that the OGSA initiative is sufficiently mature enough to establish itself as the true industry standard. Until then, we have sought to work with technologies that have a relationship with or roadmap to OGSA. Currently, we are working with the following grid platforms:
- Avaki Data Grid 3.0 -- to provide a globally available filesystem using a uniform name space (similar in function to how local universities use AFS, but with lighter footprint and simpler implementation model).
- Avaki Compute Grid 2.6 -- to provide a meta-scheduler capable of distributing jobs across a heterogeneous set of computers at multiple, separately administered sites; simplifies running applications on the grid without modifying code.
- Globus Toolkit 2.2.2 -- to enable the grid for those who are developing Globus-aware applications.
In addition, we are working with the following supporting technologies:
- Platform LSF -- for job scheduling on clusters and large SMP servers.
- Sun Grid Engine -- for job scheduling on clusters.
- Sun ONE Directory Server -- LDAP infrastructure for managing user accounts.
- Sun ONE Certificate Server -- X.509 Certificate Authority (CA) infrastructure managing certificates for use with Globus.
- MPICH-G2 -- Globus-aware MPI implementation.
- Globus CoG Kits -- "Community Grid Kits" that provide API-level access to Globus.
- xCAT -- for cluster configuration management and administration.