Speaker
Description
1. Short overview
R-GMA permits users to define their own data structures along with the fine grained authorization rules specifying who can write and read the data. They can then publish data via a producer API without knowledge of potential consumers. A consumer API is used to retrieve the permitted view of information published by the producers. Previous releases of the server have been patches to the original prototype. The new code has been re–engineered for robustness and scalability.
Provide a set of generic keywords that define your contribution (e.g. Data Management, Workflows, High Energy Physics)
monitoring, information, message, relational
3. Impact
The latest version of the R-GMA server improves on the current implementation. It is based around a new design that was derived from our experiences from the initial prototype and subsequent patches to it. Robustness and scalability are at the centre of the new design. Single points of failure have been removed and the servers have been made as autonomous as possible. Reliance on the delivery of individual control messages has also been removed. From a user’s perspective, the improvements in functionality are the introduction of fine grained authorization and virtual databases (VDB). The authorization is done using SQL views of tables constructed dynamically from user defined rules and VOMS attributes. VDBs allow for the partitioning of data. We envisage that each VO would have one or more VDBs.
URL for further information:
http://hepunx.rl.ac.uk/egee/jra1-uk/index.html
4. Conclusions / Future plans
From the existing deployment we learned not to rely upon any single message being transmitted successfully. For the new deployment there will no longer be a central registry and schema service. Instead there will be several registry replicas per VDB. For the schema there will be a replica for each VDB at each site supporting that VDB with one defined as the master schema. The design permits alternative databases to be used. Currently we only support MySQL but Oracle will be added in the future.
