Repository synchronization in the OAI framework

The Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH) began as an alternative to distributed searching of scholarly e-print repositories. The model embraced by the OAI-PMH is that of metadata harvesting, where value-added services (by a "service provider") are constructed on cached copies of the metadata extracted from the repositories of the harvester\´s choosing. While this model dispenses with the well-known problems of distributed searching, it introduces the problem of synchronization. Stated simply, this problem arises when the service provider\´s copy of the metadata does not match the metadata currently at the constituent repositories. We define some metrics for describing the synchronization problem in the OAI-PMH. Based on these metrics, we study the synchronization problem of the OAI-PMH framework and propose several approaches for harvesters to implement better synchronization. In particular, if a repository knows its update frequency, it can publish it in an OAI-PMH identify response using an optional about container that borrows from RDF Site Syndication (RSS) Format.