ERCIM News No.27 - October 1996

Digital Library Research at ICS-FORTH

by Jakka Sairamesh, Sarantos Kapidakis and Christos Nikolaou

ICS-FORTH is currently involved in R&D activities regarding the archi-tecture (system design and implementation) and economics of Digital Libraries. We are a part of the ERCIM Digital Libraries initiative called SAMOS, which is a G7 attested project. The SAMOS project aims at the development of a networked computer science technical report library in Europe. A digital library architecture will provide Internet access to a distributed, decentralised multi-format collection of documents and a multilingual interface. The SAMOS system will support a distributed digital library by storing and providing access to technical reports from the ERCIM labs.

The SAMOS system will be an open system. This means that other European institutions (university computer science departments, research organizations, industrial R & D departments) could be connected to SAMOS and have access to a large number of European technical report collections. Furthermore, a study concerning the connection of SAMOS with the DIENST system - developed in Cornell University (NCSTRL project), USA - is being conducted. This connection will allow querying from Europe collections of technical reports connected with the Dienst system in the US, and vice versa.

Work at ICS-FORTH addresses various issues concerning the design and development of large decentralized and auto-nomous Digital Library sys-tems. The primary goal is to provide architectures for distributed searching and retrieval based on QoS (quality of service) as preferred by users, and provide mechanisms for efficient usage of resources, such as processing, I/O and memory in the servers, and bandwidth and buffers in the networks. Novel performance-based architectures are also being investigated for optimized retrieval and presentation in large scale digital library systems. For example, users could have preferences on QoS and cost in viewing multimedia objects, and the architecture provides the required services.

Digital Libraries at ICS-FORTH are also being considered from an economic perspective (electronic commerce) so as to understand copyright, payment and legal issues related to information providers and Internet service providers. The goal is to understand issues of accounting, billing and payment in a commercial Internet environment. This leads to an investigation of usage costs of resources ­p; such as storage, network bandwidth and server processing - for retrieval and presentation of multimedia objects.

Currently, DIENST system services are being enhanced in several ways: performance management, load balancing of requests, scalable indexing architectures and QoS in searching and presentation. A testbed of 7 DIENST servers (shown in the figure) are being used for experiments. Some of the ideas from the experiments will be proposed for the operational DIENST system (NCSTRL project). For example, the figure shows the load (requests per hour) on the Dienst servers over half a day. Such performance metrics will be used for load balancing user requests. Statistics from the operational network are currently being gathered for better understanding of user queries and system behavior. Research work is underway to develop a large scale digital library using open standards such as OMG CORBA for promoting interoperability.

Number of requests (y axis) versus the elapsed time in hours (x axis) to each of the 7 DIENST servers.

Please contact:
Christos Nikolaou - ICS-FORTH
Tel: +30 81 391676

return to the contents page