Superstring: A Scalable Service Discovery Protocol for the Wide-Area Pervasive Environment

Robinson and Indulska

distributed hashtables service discovery pervasive computing

@inproceedings{robinson:icon-2003,
  title={{Superstring}: A Scalable Service Discovery Protocol for the
         Wide-Area Pervasive Environment},
  author={Robinson, R. and Indulska, J.},
  booktitle={{IEEE} International Conference on Networks ({ICON})},
  year={2003},
  pages={699--704}
}

Attach:Robinson-ICON2003.pdf

In Superstring~\cite{robinson:icon-2003}, service descriptions take the form of a structured string consisting of a linearly ordered set of labels capturing the topic area and capabilities of the service, similar to standard DNS names. In registering a service, a node takes the top level component of the description and uses the DHT to map that component to a controlling node. That node in turn stores a reference to the service and then repeats the same process on the next component of the description, and so on until referents have been stored at all components of the description. Querying proceeds similarly, but doesn't quite follow a linear chain in the same fashion as there may be unspecified components in the query. Instead, each node along the chain of mapped specified components computes the intersection of the results associated with subcomponents, returning the resulting set of nodes.