Research

My research interests are in the area of computer systems modeling and performance evaluation. One aspect of my research has been development of models and efficient methodologies for evaluation of systems at design time. High level abstractions that allow for evaluation of design choices (rather than prediction of performance numbers of the final product) can provide insight into the system being designed, expose the tradeoffs involved, and lead to significantly better designs of systems. In general, design choices are better studied through analytical models --- at design time it is desirable to obtain results quickly since there is a large number of possible design choices. Higher accuracy which may be obtained through detailed simulations or measurements is often not needed at design time or simply not possible or economical to achieve. Hence, I work on analytical performance evaluation methodologies that lead to (a) reasonably simple abstractions, needed for ease of applicability, and (b) efficient solution techniques, needed for rapid evaluation of design ideas.

Another aspect of my research has been quality-of-service-based design of large-scale continuous media storage systems and their subsequent performance evaluation. Many systems today are moving towards applications whose primary function is to transfer large volumes of data from an I/O subsystem through the network, possibly subject to some form of real-time constraints. Therefore, efficient storage system designs are fundamental to the viability of these applications, and I expect our work to be of use to a broad range of current and future systems and applications.

More recently, I have focused on performance aspects of scalable upload applications over wide-area networks. To the best of our knowledge there is no existing work, except ours, on making upload applications scalable and efficient. This corresponds to an important class of applications, which includes digital government, digital democracy, and Internet-based storage. My current work in this area is on design of an application layer infrastructure that will improve scalability and security of current and future Internet-based services.