Guardians

Project Overview


Guardians logo

The GUARDIANS project identified the technology for the next generation of Information Management tools and
specified how the required infrastructure allowed distributed components to interwork. Information Management in this
context is taken to have a wider meaning than just the content that is of interest to users. It encompasses:
1. Rich, diverse and interactive content drawn from both web resources intended to be delivered over the Internet and
interactive digital broadcast content;
2. Metadata describing these diverse resources;
3. User profiles describing user preferences, both on content and access configuration;
4. Advanced search and notification services ( traditionally called Selective Dissemination of Information SDI);
5. Information Object Agent technology capable of comparing resources and their metadata descriptions to identify
related materials;
6. Distributed software technology for dynamic management and update of metadata, with specific reference to
problems related to availability and accessibility of multimedia resources.
The GUARDIANS project  built upon the results generated by the GESTALT project in the 4th Framework ACTS

Programme. GESTALT defined a flexible, component-based architecture for the discovery and delivery of modular
training and education services. GUARDIANS  focussed on the following areas to extend this architecture utilising the
emerging technologies identified in the text below. The focus was expanded from a purely learning/training domain
to encompass more generic Information Management issues.

Mediation Service
The GESTALT Resource Discovery Service (RDS) based on the GAIA Architecture (rfc 2552) and implemented in a
CORBA environment, will be integrated with next generation search services (e.g. Mozilla browser services) which
support the W3C Resource Description Format (RDF). The project will explore collection level description services,
allowing the RDS to identify and locate repositories holding similar or related materials. It is anticipated that Dublin
Core v2.0 will offer a minimum level of cross-domain search support via the DC core set of elements but that this can
be extended across specific groups of repositories. The RDS will access user profile information to enhance search
services offered to individual users. By empowering the RDS with the ability to target repositories based on their
content profile and detailed knowledge of users via their user profile, a powerful Mediation Service will be created,
better able to serve and anticipate user needs. The Mediation Service will be enhanced to make use of Smart Content (in
a sense making information objects self-aware) that encapsulates the appropriate metadata with a hierarchy of methods
that allow them to interact with other smart content objects and services. This will allow the automatic selection of
smart information objects based upon information stored in a users profile. This matching will not just be a 1 to 1
element match but will use techniques such as Natural Language Processing, rules sets, etc.

Distributed User Profile Services

Central to the GUARDIANS approach is the provision of a distributed User Profile Service using LDAP v3.0 (rfc 2251)
to support an extensible set of online applications.

Reusable Object Technology

In better supporting user needs and providing precise information at the time it is required, there is a clear drive to encapsulate this in finer grained information objects which have a focused scope and which can readily be reconfigured and reused.

Customised Cross-Platform Delivery

GUARDIANS explored delivery of information objects across the Internet and Digital Broadcast. Whilst these
access technologies are very diverse, they are each capable of offering a web-based presentation to the user (see W3C
TV-Web Interest Group).

Distributed Management of Metadata

Another issue addressed by GUARDIANS was the ability to maintain metadata consistency with the current status of the
related content.

The Project ran from 2000 to 2002

Partner List:

Fretwell-Downing Education
UK

British Telecommunications plc
UK

NetG Incorporated
USA

Project Achievements


1. Validated the extended CMI Data Model for reliable and consistent user tracking across Terrestrial Networks and
Digital Broadcast;
2. Constructed advanced search capabilities within the Mediation Service utilising IEEE Learning Objects Metadata
(LOM) to describe information objects and IEEE PAPI to describe individuals and their preferences;
3. Evaluated proposed mechanisms for dynamic personalisation of services according to the delivery platform selected.
Note that this will require the reconciliation of PAPI ‘application’ preferences with the CC/PP device capabilities
and ‘browser’ preferences;
4. Experimented with how the GESTALT QoS metadata can be used (and possibly extended) to support operation
across these technologies and negotiate appropriate levels of QoS;
5. Proved the proposed integration strategy for the back-end management and tracking of information services, thus
demonstrating the feasibility of common services accessible from the office, the factory and the home.
6. Proposed a scheme for distributed, on-line synchronisation of metadata to allow the inclusion of dynamic
information such as quality of service parameters, cost, and availability of resources.

3
adapting the evolving CMI Data Model for the description of discrete, fine-grained information objects, which will
emerge from the convergence of training with information services.
Customised Cross-Platform Delivery
GUARDIANS will explore delivery of information objects across the Internet and Digital Broadcast. Whilst these
access technologies are very diverse, they are each capable of offering a web-based presentation to the user (see W3C
TV-Web Interest Group). Over time and dependent upon physical location, users may access services via either of these
technologies. Thus dynamic personalisation of online services will be vital to presenting information in an appropriate
manner for the user. In particular, there are two proposals currently under development within W3C that are pertinent to
this issue and worthy of further study and experimentation:
1) Personalised Information Description Language (PIDL) – personalisation of online information by providing
enhanced interoperability between personalisation applications;
2) Composite Capability/Preference Profiles (CC/PP) – is a user side framework based on RDF for negotiating user
preferences and device capabilities.
Distributed Management of Metadata
Another issue addressed by GUARDIANS is the ability to maintain metadata consistency with the current status of the
related content. In GESTALT, metadata was used to maintain information related not only to the semantic content of a
resource but also to provide data linked to the delivery process of the content itself. In particular, GESTALT has
defined Quality of Service (QoS) related metadata and successfully demonstrated how this metadata can be used by
applications to demand performance guarantees in the delivery of multimedia content over QoS enabled network
architectures. In the GESTALT implementation the parameters considered (e.g. transmission peak and average rate)
were only related to requirements in terms of network resources considered necessary for achieving adequate quality in
the delivery of the content. However, in a truly distributed scenario, other information should be considered such as the
actual load of the servers from where the content will be delivered, since these also influence the capability of assuring
quality in the delivery process. While semantic information is relatively static over time, QoS metadata as described
here is inherently dynamic as it relates to parameters describing the current state of networks or systems. The inclusion
in metadata of this time and state dependent information requires mechanisms for updating the existing metadata
replicas referring to a specific resource. GUARDIANS will deal with the definition of software and protocol based
mechanisms for the on-line management and synchronisation of metadata.
It is a specific objective of the project (for terrestrial networks) to continue the QoS metadata work started in
GESTALT, by which information objects can be characterised as to their network bandwidth requirements as the
service is launched. Current operation uses RAPI and SCRAPI as the application interface by which QoS is negotiated
using RSVP. Since quality of service is a function of a number of parameters related to multiple components of a
distributed information system (e.g. network architectures, server and resource repositories, client terminals), metadata
will be extended to include such diverse aspects. However, since most of this information can vary in time, according to
the current status of the corresponding component, these extensions will require the design of proper mechanisms for
their on-line updating. The Project will also demonstrate a prototype implementation of the proposed mechanisms.
Despite the diversity of the selected access technologies, there is still a need for a unified model for capturing user
responses in terms of managing these services and tracking user understanding of the information objects presented.
Based on the reusable object technology developed in the project around the CMI Data Model and the related Content
API, the project will produce demonstrable information object content and appropriate user environments for each of
the delivery modes.
Objectives
1. Validate the extended CMI Data Model for reliable and consistent user tracking across Terrestrial Networks and
Digital Broadcast;
2. Construct advanced search capabilities within the Mediation Service utilising IEEE Learning Objects Metadata
(LOM) to describe information objects and IEEE PAPI to describe individuals and their preferences;
3. Evaluate proposed mechanisms for dynamic personalisation of services according to the delivery platform selected.
Note that this will require the reconciliation of PAPI ‘application’ preferences with the CC/PP device capabilities
and ‘browser’ preferences;
4. Experiment with how the GESTALT QoS metadata can be used (and possibly extended) to support operation
across these technologies and negotiate appropriate levels of QoS;
5. Prove the proposed integration strategy for the back-end management and tracking of information services, thus
demonstrating the feasibility of common services accessible from the office, the factory and the home.
6. Propose a scheme for distributed, on-line synchronisation of metadata to allow the inclusion of dynamic
information such as quality of service parameters, cost, and availability of resources.

Funding


EU FP4.