A network model for wide area access to structural information

Structural information in medicine is information about the physical body. Recent advances in medical imaging and biotechnology have greatly increased the amount and importance of structural information, and advances in networking envisioned by the High Performance Computing and Communication Initiative (HPCC) will allow this kind of information to be delivered to remote clients over wide area networks. One of the most important factors determining the usability of such a client-server configuration is the time delay between the request for information from the server, and its presentation to the user at the client. In this paper we present a model for predicting the performance of a structural information client based on the ping time, a simple, unobtrusive network measurement. Preliminary results suggest that the relationship between ping time and transfer time for large files is linear, which if borne out by more data, will allow the performance of structural information clients at remote sites to be predicted without the expense of installing them first. At the same time, such a model will be useful for planning improvements to the network in those sites which could most benefit by wide area access to structural information.