DRIVER-II was an EU FP7 “e-Infrastructure - Capacity Programme” co-funded project (January 2008 - December 2010). It involved 12 partners from 10 European countries. DRIVER-II extended geographically the European Repository Infrastructure created by its predecessor project, DRIVER, while putting it into production. Currently, this e-Infrastructure is composed of more than 220 archives coming from 23 EU Countries which contain approximately 1,000,000 documents. From the technological point of view, the DRIVER e-Infrastructure is powered by the D-Net software suite built as part of the two subsequent DRIVER projects. This suite offers advanced integration, retrieval and user-centric services around the repositories it confederates.
The scope of collaboration with DRIVER–II was initially placed on the areas of Interoperability and Service Provision. The investigation of the interoperability solutions between the two infrastructures exists in the D4Science-II workplan. However the ground of activities has expanded to cover technology and best practices exchange among the teams that implement the technologies behind the two projects.
More particularly, a very tight collaboration with the DRIVER-II technical team has been established for investigating a wide range of options for collaboration among the systems developed by the two projects. This is facilitated by the fact that NKUA, CNR and CERN are core technological partners in both projects.
Due to the fact that DRIVER-II project concluded its activities early in the lifecycle of D4Science-II, taking advantage of the observation the a substantial fraction of collaboration was on the RTD fields, the collaboration scope has been revised to embrace the gCube and D-Net Systems infrastructures in a wider manner, beyond the limits of the two projects. This approach has given the opportunity of extending the collaboration also with the OpenAIRE project (see D4Science-II Collaboration: OpenAIRE (Open Access Infrastructure for Research in Europe)).
Several internal meetings and a number of teleconferences have been conducted in order to define the options and the viable interoperability and servicing options among the two infrastructures. The joint activities initially defined a number of cases that where subsequently further elaborated and homogenized with cases stemming from the analysis of other infrastructures. A subset of these cases, considered to be the most viable, feasible and cost-effective, have been promoted to reach the implementation and servicing stage.
In overall the activities performed are as follows:
• DRIVER-D4Science Interoperability cases have been documented on the D4Science-II NA4 Wiki;
• DRIVER Functional requirements have been documented on the D4Science-II NA5 Wiki;
• Interfaces for interoperability among gCube and D-Net system have been implemented in the gCube system;
• A set of gCube services for delivering services to the Ecosystem, under the request of DRIVER-II project, have been designed and implemented;
• A set of gCube services for delivering D-Net services to the Ecosystem, under the request of other communities and infrastructures habe been designed and implemented;
• Advanced interoperability cases have been fully tested some of them are already in production.
During this process, the analysis of services and other technologies revealed some options for further enhancing gCube services, in the area of interoperability, not only through the adoption of standards but also through the reuse of technologies that have been applied in D-Net system. As a result new design and implementation activities that take into account such findings were started.
Outcome and plans for the future
The DRIVER VRE is one of the Virtual Research Environments that have been deployed by D4Science-II. It is a programmatically accessible environment that serves as an enabler of the interoperability features of the two technologies.
Beyond this, as of today, the achieved results of the collaboration among the two projects are:
• gCube-powered VREs can register D-Net-powered repositories as data sources and span searches to both internal (gCube) and external (DNet) engines, homogenize and present the results in a uniform manner.
• gCube-powered VREs can publish their content to be harvested and validated by D-Net-powered repositories
• D-Net repositories can send requests for content transformations to gCube-powered infrastructures, hosting the gCube Data Transformation Service.
• gCube new information retrieval engine adopts the D-Net proposed language for Information Retrieval and Search introspection, based on Contextual Query Language along with a series of extensions.
Ongoing work focuses on the implementation of specific advanced functionalities within D4Science-II for consumption by D-Net. Such functionality covers :
• Calculation of Bibliometric indices,
• parallel indexing
• distributed OCRing
Additionally, the direct integration of D-Net repositories within gCube storage system is under development.
Furthermore, cases under consideration include automatic document categorization and author disambiguation.