In the early days of computing, machines were used for scientific applications that required computing power for number crunching. Today the uses of computers have grown exponentially, including business applications such as email, multimedia, data mining, and others which require intensive data retrieval. CPU speed has increased in accordance with Moore's law, or double every 18 months, while at the same time memory access speed has only been improved an average of 9% per year over the past 20 years. Peak CPU speed is no longer matched by the deliverable performance of your application, thus data-access has become the bottleneck of computing, especially for data-centric business applications. CS Professor Xian-He Sun has proposed a novel data access architecture, called server push architecture, to solve the data-access bottleneck. National Science Foundation grant to develop this architecture.
In addition, Prof. Sun is a co-investigator for a five-year project entitled "National Computational Infrastructure for Lattice Gauge Theory," for the Department of Energy, in which he is responsible for the workflow development. These two projects started in August and September of 2006, respectively.
|