We present SGuard, a new fault-tolerance technique for distributed stream processing engines (SPEs) running in clusters of commodity servers. SGuard is less disruptive to normal s...
YongChul Kwon, Magdalena Balazinska, Albert G. Gre...
Applications structured as parallel task graphs exhibit both data and task parallelism, and arise in many domains. Scheduling these applications on parallel platforms has been a lo...
Abstract. A new approach for acquiring knowledge of parallel applications regarding resource usage and for searching similarity on workload traces is presented. The main goal is to...
This paper investigates what it entails to provide I/O service diļ¬erentiation and performance isolation for virtual machines on individual multicore nodes in cloud platforms. Sh...
Advances in multiprocessor interconnect technologyare leading to high performance networks. However, software overheadsassociated with message passing are limiting the processors ...
Debashis Basak, Dhabaleswar K. Panda, Mohammad Ban...