- Ubiquitous Computing
- Web Services
- The Semantic Web
- Spaces Computing
- Peer-to-Peer Computing
- Collaborative Computing
- Dependable Systems
- Security
- Languages
- Pervasive Computing
- Cluster Concepts
- Distributed Agents
- Distributed Algorithms
- Distributed Databases
- Distributed Filesystems
- Distributed Media
- Distributed Storage
- Grid Computing
- Massively Parallel Systems
- Middleware
- Mobile and Wireless Computing
- Network Protocols
- Operating Systems
- Real-Time and Embedded Systems
- Commentary
- Endnotes
Distributed Agents
An agent can be many things. The broadest accepted definition is that of an entity that acts on, or has the power or authority to act on, behalf of another. An agent can be thought of as a means by which something is accomplished or is caused.
In any agent-based model, a human being (or even a non-human agent) may delegate some authority to the agent, which may be "intelligent," mobile, or both. Given the false start that AI appears to have suffered, it may not be realistic to speak of intelligent agents in NDC. We can say, however, that an intelligent agent may be able to make rule-based inferences and conduct probabilistic decision analyses or learning on behalf of its client.
In the context of NDC, a mobile agent may move between nodes to accomplish assigned tasks; this vision is of particular interest to mobile users and mobile communications. Such a view naturally raises many questions with respect to security and trust-based models, not to mention assumptions regarding code viability on a given node. Recall Deutsch's Eight Fallacies; many of those fallacies must either be ignored or addressed by substantial work at an industrywide level if a standard agent-based model for NDC is to become reality. And given the presumed relationship between the concept of distributed agents and so many other fitscapes of NDC, it is difficult to imagine a scenario that proffers the ultimate achievment of ubiquitous computing (which I consider to be the teleological vector of NDC, if one exists) without such a standard framework.
Not all nodes must sing the same songquite the contrary. But "agentness" must be well formalized if a semblance of the intelligence needed to traverse an arbitrary network is ever to be realized.
Within NDC, agents and the characteristics of distributed agents were first considered in the problem space of systems and network management. The standardization of distributed agents is now being explored in the context of the Semantic Web as well as the DARPA Agent Markup Language (DAML).