时间:2019-07-13 10:57:04

目:Rethinking Abstractions in Virtualized Systems

报告人:饶嘉 美国德克萨斯大学阿灵顿分校




Abstraction is a powerful technique for hiding the complexity of computer systems. By decoupling high-level, simple interfaces from their low-level, complex implementations, abstraction helps improve programming productivity, ease system management, and enhance reliability. Virtualization, the process of creating a virtual copy of a physical resource, is one such example of abstraction and has been the key technological driver of cloud computing. However, performance, cost-effectiveness, and predictability remain critical challenges in virtualized systems, impeding the adoption of virtualization in many critical domains. While many factors are responsible for the overheads associated with virtualization, a fundamental problem is the semantic gaps due to abstraction.

In this talk, I will first demonstrate that the abstractions in widely adopted virtual systems, such as virtual machines, containers and virtualized networks, have been primarily designed as a minimal interface for correct program execution and fault isolation, but inevitably create semantic gaps that hamper cross-layer optimizations across the software and hardware stacks. This deficiency makes it difficult for cloud systems to meet the diverse needs of individual users while simultaneously maintaining high utilization and efficiency. Next, I will present our recent works on bridging the semantic gaps in different types of virtualized systems. The talk will be concluded with discussions on how to improve the abstractions in virtualized systems to enable a holistic solution.


Dr. Jia Rao is currently an Assistant Professor in the Department of Computer Science and Engineering at the University of Texas at Arlington. His research lies broadly in the area of Operating Systems, Parallel and Distributed Computing, and Cloud Computing. His recent focus is on understanding application behaviors and the interactions between systems and architecture in multi-tenant systems. Highlights of Dr. Rao's research include a National Science Foundation (NSF) CAREER award, best paper awards at APSys (2016), ICAC (2013), and best paper nominations at HPCA (2013) and HPDC (2013). Dr. Jia Rao received his B.S. and M.S. degrees in Computer Science from Wuhan University in 2004 and 2006, respectively, and a Ph.D. degree in Computer Engineering from Wayne State University in 2011.