This research concerns the problem of discovering, identifying and locating resources in a nationwide network. The research is based on two fundamental tenets. The first is that users should be able to identify a resource with a set of attributes or properties that describe the resource. For example, rather than having to know the name or address of a processor, the user should be able to state that he or she wants a processor that has a particular architecture, runs a certain operating system, supports a given software package, and is reachable via a specific communication protocol. It should be the naming system's responsibility to locate a resource that matches such a description. The second tenet is that the naming system for a nationwide network will be highly decentralized. The primary reason for this decentralization is not the problem of scale so much as the problem of autonomy. Such a network is expected to consist of a collection of autonomous systems, each of which is responsible for managing and naming its own resources. For example, the National Research and Education Network spans thousands of autonomous organizations. The naming system for such a network probably will be decentralized proportionally.The researchers are investigating methods for identifying resources with a set of attributes, and exploring techniques for searching a large, high decentralized name space. This effort is based on their experience designing and using two experimental naming systems: Profile and Univers.
