Resource Information Schema and Services

In order for users to identify appropriate resources within a Grid infrastructure there must be some form of resource information, provided through standard schema and with standard query mechanisms. There are three different schema used in production grid projects today, the Glue Schema, the ARC schema and the CIM schema with the NAREGI vendor extensions. The Glue schema is used by OSG, EGEE and Teragrid and has been mapped to LDAP, XML and the relational model. The NDGF are using the ARC schema and NAREGI is using the CIM schema with a vendor extension.

Participating Infrastructures

Meeting Summary

Status 06/06/06

1. Still need to agree on a common minimal attribute set (several possibilities have been proposed)- waiting on chair's decision
2. Need to identify participating sites which will agree to advertise these sites in their local schema
3. Need to agree on a site/tech which will scoop up these different schemas and display them
4. Some work is being done on translators

Note: Not all work in done via the gin-info group. There is are many bilateral interoperability activities on going between different grid infrastructures. It is hoped that the gin activity will leverage this work and be a point of contact for everyone.

The GIN-BDII is a top-level BDII which contains information from all Grids in accordance to the Glue schema.
The ARC-BDII is a top-level BDII which contains information from all Grids in accordance to the Nordugrid schema.
The Gin Cell Domain is a Naregi cell domain containing information from from all Grids in accordance to CIM and the Naregi vendor extensions.

The above work has resulted in all information from the grid infrastructure being found in one place. The information for different sites in different grid infrastructures can be of various quality. This is due to a number of reasons. Not all attributes can be found in all grids, this will lead to missing data. The implementation may have errors. Sites may not be publishing the information correctly.

The impact of missing and incorrect data will only be apparent when we try out some use cases on the data. The initial use cases should be what we would like to demonstrate at Super Computing.

Use Cases

1. A GoogleEarth map showing all the sites with the grid infrastructure affiliation
2. Google Earth 2007
3. Service Discovery for CEs which can be used by the GIN VO

Conclusions

1. SiteEntry

Bilateral Activities

• EGEE/OSG
• EGEE/NDGF

Attachments:

OGF18-gin-info.ppt [GINInfoWiki/OGF18-gin-info.ppt]