The institute of parallel and distributed systems (IPADS) conducts research in all aspects of computer systems, with a primary focus on operating systems, system virtualization, programming model and runtime for multicore and distributed systems. Our research also involves other disciplines such as computer architecture, language and compilers. The research themes of IPADS are improving the performance scalability, energy efficiency and dependability of centralized and distributed computer systems.
We have built a number of systems to address the scalability and dependability issues of computer systems. Examples include a trusted cloud platform called CloudVisor that addresses the trustworthiness of multi-tenant cloud, a portable parallel full-system emulator called COREMU that provides scalable and efficient emulation of several hundreds of cores running commodity operating systems, an operating system scaling scheme called Cerberus that mitigates contentions in commodity operating systems on multicore, scalable and efficient graph-processing platforms like PowerLyra, Polymer and Imitator, and dynamic software updating systems POLUS and LOCUS that can seamlessly update commodity operating systems and multithreaded software with new features or to newer versions.
Our current research focus is in providing programming models and system support for large-scale computing systems, scaling existing software stack to existing and future computer platforms with hundreds to thousands of homogeneous and heterogeneous cores, and provide hardware and software solutions to secure and trustworthy computing environments in both mobile and cloud platforms.