Difference between revisions of "IS-Collector"

Revision as of 17:18, 21 August 2009

The IS-Collector is the gCube Service in charge of collecting WS-ResourceProperties (including the Resource Profiles from the IS-Registry) published through producers’ components (e.g. IS-Publisher) and to make them available for consumption.

It is designed as an aggregator service[1], able to create Aggregator Sinks that query remote Aggregator Sources (in particular QueryAggregatorSources, as those created by the IS-Publisher) to harvest resource properties. The collected information is stored in an embedded instance of an XML DB (eXist XML database[2]). This allows persisting the information as well as to design the IS-IC query interface by relying on the XML Query query (XQuery[3]).

Major features are:

• storage, indexing, and management of GCUBEResource profiles
• storage, indexing, and management of RunningInstances' state
• full XQuery support over the handled information
• remote management of the XMLStorage

Architecture

From an architectural point of view, the service is composed by several portTypes which globally implement the service functionalities.

The following picture shows the actual architecture of the Collector service:

Figure 1. IS-Collector Reference Architecture

where:

• the orange boxes are the classes implementing old and deprecated portTypes needed to maintain backwards compatibility with the previous implementation of the service
• the blue boxes are the classes providing the new functionalities
• the grey boxes are the Aggregator Framework classes or interfaces extended/implemented within the service.