Studies started in year:
 Projects and Studies
P845
Agent Technologi in perspective
closed
For further information please contact:

Juan A. Siles
Project Supervisor
EURESCOM
Wieblinger Weg 19/4
69123 Heidelberg, Germany

Project Information
What is this Project about

Although there exists many implementations of agents, the technology is generally not considered mature. A mature agent frameworks is still a long term goal. However, agent technology is appearing in many different disguises, from the General Magic Telescript mobile agent through to complex artificial life form agents exhibiting complex, perhaps intelligent behaviour. Equally diverse applications of the technology are now appearing, ranging from children’s games such as ‘Creatures’ from Cyber Life to the control and management of complex systems such as communications networks and fighter aircraft. Major steps forward are being made in the development and understanding of such agents and increasingly ideas are drawn from a wide range of technical disciplines, most notably artificial intelligence, complexity theory and distributed computing.

Much of the focus of this Study will be on autonomous mobile agents as they are likely to be highly relevant to TelCos networks and systems. Mobile agents present a dynamic way to optimise the computational and communication resource usage of future distributed application components. Earlier distribution technologies based on messaging, like OMG CORBA, are currently unable to exchange code. Java overcomes this and provides distributed applications with the capability to exchange code in addition to usual messaging and object state information. The significance of mobile agents lies in the dynamic distribution of application components into suitable locations for effective and economical computing and transport resource usage. To facilitate this, the agent software components are given the freedom to traverse the network to find hosts best able to execute their code and handle their communications. There is, however, a price to be paid for freedom of movement, including computational and code complexity overheads and the stringent security requirements that stem from the autonomous nature of the Mobile Agents.

A key issue is whether there exist a relationship between network service cost/performance and the usefulness of the mobile agent technology in certain application areas. Its not certain whether a fast, reliable and cheap transport service could eliminate the need for mobile agents or whether they have advantages which will make them a de facto part of tomorrows network.

Objectives

The main objectives of this Study are to:

  • Clarify the key concepts of software agent technology, evaluate their maturity and indicate their evolutionary potential.
  • Identify application areas for customers and TelCos where agent technology has practical value. Identify, compare and contrast alternative technologies to meet the demands in these areas. In particular evaluate their usefulness as perceived by the customers as well as by the TelCos.
  • Identify the market opportunities and possible improvements that agent technology could provide, and assess whether it could create unexpected market or regularity events. Also assess legal and commercial issues and comment. Identify what the impacts will be on TelCos of the intensive/ extensive use of agents and the products and services they enable.
  • Review possible application/middleware level infrastructure that would support and exploit agent technology and comment.
Project Results
Deliverable Title  No Planned issue date
The impacts of Agent Technology on the way TelCos are providing services and approaching new markets D1 published

 

Project Information
Project Leader and Contractor Code: Alexander Grosse, DT
Project Supervisor: Juan A. Siles
Project Participants: BT, DT, ST
Project Start Date: 1 October 1998
Project Completion Date: 9 July 1999