A location model for a web service intermediary

Recently, Web services have entered the competition for a new type of distributed e-commerce platform. This paper studies the placement of servers of a Web service intermediary. The intermediary serves as a common interface to its clients while obtaining Web services from independent providers. The intermediary locates its servers by minimizing costs where a major component is network latency. We propose and study an integer programming formulation to determine the locations and usage rates of the servers of an intermediary. A greedy heuristic method is used to obtain good solutions to the problem.