Enhanced biomedical image scanning technology and growing network accessibility have created a need for faster and more efficient data exchange over the Internet and in closed networks. Researchers and biologists are conducting experiments on large biomedical data sets and want to share information with other scientists to progress in their research. But the information they are dealing with are large-scale data sets that occupy gigabytes of space, which makes the storing of these data sets onto one's local hard drive very difficult. The size of these data sets also makes them difficult to transmit over currently existing network links. To overcome these difficulties, those data sets are stored in large-scale data repositories, from which they can be retrieved upon user's request. The challenge is to make the data accessible within a reasonable amount of time at a reasonable quality without losing detail or image resolution. We describe a hierarchical storage scheme based on...