Seminar: Characterizing and Optimizing Virtualized Network Performance in the Cloud

Kun Suo headshot













Kun Suo

The University of Texas at Arlington

January 11, 2019

3 pm

Patrick F. Taylor Hall, Room 3107


To leverage the elastic resource allocation of cloud computing and enhance the service availability and productivity, numerous applications and businesses have been moved into the cloud during the past ten years. Besides the high resource utilization, flexible resource management, virtualization in the cloud also incurs additional overhead, scheduling delays as well as semantic gaps among hardware, operating system and applications, especially for the I/O‐intensive services.

This talk covers two of my recent projects on characterizing and optimizing efficiency of virtualized networks. I will first introduce Time Capsule, an in‐band packet profiler which traced packet level granularity latency across different boundaries in virtualized systems with negligible overhead. With Time Capsule, I discovered multiple bugs in the Xen’s credit scheduler which caused the long tail latency of I/O workloads. Next, I will present xBalloon, a lightweight approach to preserving static and dynamic priorities between I/O‐bound and compute‐bound tasks and boosting the I/O performance under discontinuous time. Experiments show that xBalloon can improve throughput and tail latency by up to an order of magnitude. The talk will conclude with a discussion of new challenges and opportunities in the future cloud infrastructure.


Mr. Kun Suo is a Ph.D. Candidate from Department of Computer Science and Engineering at The University of Texas at Arlington. He received his B.E. from Nanjing University in 2012 and has been working with Prof. Jia Rao on computer systems since 2013 . His research experience spans across multiple areas in cloud computing, virtualization, operating systems, containers and software defined network. He has published in top conferences including APSys, SoCC, EuroSys, Middleware, INFOCOM, HotCloud, etc. He received the best paper award from APSys 2016 and best poster award from SoCC 2017.