CRI: CRD: A Web-Accessible Grid-Computing Resource for the Telecommunications Research Community

点击翻译按钮获取中文摘要

This project will provide researchers in the fields of information theory, communication theory, and fundamentals of wireless networking with a grid-powered web service via a web portal interface. The web service will enable rapid execution of computationally demanding computer simulations that support the research of those working in the field. It is necessary because compute-intensive simulations are essential to progress in the field, yet the associated computational requirements often exceed the available resources of all but the largest institutions. The extant grid supplying power to the initiative leverages the idle capacity of participating computers. Rather than deploying a dedicated infrastructure, it uses existing computers in the libraries, student labs, and staff offices of the participating institution and other participating institutions that choose to donate idle resources.The software to be run on the grid leverages an extensive free Matlab-based library developed by the Principal Investigator and used by an existing base of over 400 scientists and engineers. To simplify the user experience and obviate the need for users to purchase Matlab licenses, a web portal interface will be developed. Researchers from other institutions will be given access to the service and will be encouraged to donate their own idle computer capacity to the initiative.Intellectual Merit: Research in the fields which this grid will serve highly mathematical. The underlying mathematics often involves multidimensional integrals that can only be solved via Monte Carlo integration or simulation. Examples include the design of efficient error correction codes, the calculation of channel capacity under modulation constraints, and the design and analysis of wireless ad hoc networking protocols. The grid-based infrastructure that lies behind the proposed web service will allow rapid parallel execution of Monte Carlo simulations. The consequence is that researchers will obtain results faster (by orders of magnitude) than could be obtained when running on a single processor. Their results will also benefit from dedicated tools for their work and the collaborations hosted by the infrastructure.Broader Impacts: There are both research and educational broader impacts. The research impact is that the proposed web-based simulation service and underlying grid-based infrastructure will provide world-class computing power to groups that might not otherwise have access. The proposed account creation and management policies will allow for underrepresented or disadvantaged groups or institutions to have a larger share of computing resources. Two notable forms of dissemination will be accomplished by continuing to publish the underlying software under an open source license and by presenting tutorials on the software and the infrastructure at conferences.