logo

Intelligent Interactive Distributed Systems


Project: Security & Trust
Home  People  Publications  

Introduction:

One of the aspects of managing large scale distributed systems is security. Trust is the basis for many of the solutions proposed. Large scale distributed systems in which code, to be more precise software agents, migrate between systems, place additional requirements on the design of such systems. Not only do the systems need to be protected from malicious agents, the agents themselves need to be protected from malicious systems. In large scale systems, performance is also of importance.

In open Internet scale environments, guaranteed secure hardware is often not an option. Encryption based technologies (e.g. use of encryption of serialised agent using public/private keys or the use of certificates) provide partial solutions. Complete solutions have not yet been found. Trust plays an important role. Certification based solutions, for example, require recognised certification provider/authorisation services.

Challenges:

The AgentScape Operating System is being designed to support protection of all parties within the limits of resources available. Look up services (see directory services) such as name services, location services and yellow pages, also need to be secure/trust-worthy. The role of trust in human interaction is currently one of the areas explored (see distributed design). The role of identity management/integrity is another (see ALIAS)



login
For comments on this web-site, please contact: Michel Oey