Saturday, August 29, 2015

A multi-agent approach for autonomous digital preservation

A multi-agent approach for autonomous digital preservation. J. Pellegrino,  M. Maggiora, W. Allasia. Multimedia & Expo Workshops, 2015 IEEE International Conference. June 29 2015-July 3 2015. 
     Digital obsolescence is caused by the ongoing development of new software and new formats, so the risk of obsolescence can be estimated from a global environment.The model described in the article
presents two main strategies to cope with digital obsolescence: migration and emulation, and then focuses on migration. Migration consists in converting digital objects into a new and more recent format and is the process we are going to focus on in this work.

An agent can be defined as a computer system that is capable of autonomous actions to meet its  objectives. The environment and all the agents that interact together to share information constitute a multi-agent system. The agents interact together in a variety of ways to meet the objectives. An agent-based model could be set up to deal with digital preservation issues; they can acquire, evaluate and share information in order to understand obsolescence risks and the best preservation action to perform. This will require a trust relationship between archive entities. There are some tools that perform obsolescence identification and metadata extraction, such as AONS (Automated Obsolescence Notification System), DROID (Digital Record and Object Identification), and JHOVE (JSTOR/Harvard Object Validation Environment).

Preservation processes require institutions to define a preservation plan. PLANETS (Preservation and Long-term Access through Networked Services) and Scout (developed within the SCAPE Project) are tools that help with identifying preservation issues and managing digital repositories. None of the tools discussed so far use the described model which aims to emulate a distributed environment where  archive entities can share information to find solutions to the digital preservation issues.

The model is self updating from preservation actions taken, and is made up of three main parts:
  1. Global: it includes all those variables accessible to every agent. 
  2. Entities: contains the declarations of all the species of agents that take part in the model. 
  3. Experiment: dedicated to the experimental setup.
The work presented in this paper provides a novel approach to the decision processes concerning one of the most common digital preservation issues such as the migration process. The agents in this migration process have the capability of communicating, cooperating and propagating information about the performed actions in order to help each of the other agents find the best solution to a given preservation issue. The goal is to provide a framework so users can simulate different digital preservation scenarios and approaches. The model has show the stability of the framework under various use cases.

No comments: