Aims
The objective of this work package is to develop and enhance the core service infrastructure that will form the foundations of the DMI platform, which will be deployed to this infrastructure and delivered to the consortium and beyond by WP3. The architecture effort of WP2 determines the components to be developed and included as part of the DMI platform and the modelling and language research undertaken in WP1 will form the basis of the DMI model for the platform. Where requirements are identified for enhancing the capabilities of the infrastructure for wider applicability across a number of tools from WP5 and services from this work package, respectively, the infrastructure will be enhanced accordingly.
The core service infrastructure is built upon the Universal Service Management Technologies (USMT). The USMT will provide a set of common service management capabilities that are required for monitoring and management of services hosted by the DMI platform. The USMT, provided by Fujitsu Labs of Europe at the start of the project, will be enhanced and expanded as required by the work of the project to support the needs of Data Mining and Integration.
Data Mining and Integration processes are implemented as a set of individual processing steps, implemented using Grid services, which are combined to form the end to end process. These processes can be complex workflows involving several sub-processes such as data selection, data filtering, data transformation, data integration, data modelling (applying a data mining algorithm), and post-processing the mining results (e.g. visualisation). The Data Mining and Integration processes are often large in terms of the number of sub-processes in the workflow; in terms of the total execution time of the process; and in terms of the data volume involved.
In this work package, the USMT will be modified to facilitate the management and monitoring of complex DMI oriented workflows, by integrating a number of technologies. These include the OGSA-DAI platform, an enhanced monitoring framework for DMI oriented processes, the OGSA-DQP platform, a semantic registry and semantic provisioning services for providing advanced service and resource publishing and discovery mechanisms.
The USMT will give rise to a layered Data Mining and Integration environment that will allow services to plug into lower service layers with the USMT as the foundation layer.