I am a Ph.D student in computer science at the Institute of Computing Technology,
Chinese Academy of Sciences since 2010.
Prof. Jianfeng Zhan is now my advisor.
My research focuses on operating system, parallel and distributed computing, Power saving, and Data Center Computing Testbed.
I received my B.S. degree in 2010 from Huazhong University of Science and Technology in China,
in Computer Science.
My hometown, Jingzhou, lays in the south of Hubei Province. It has a long history.
My research interests focus on the following points of Cloud computing and Data center computing:
- Operating systems.
- Data center computing testbed.
- Instrumentation, profiling, and tracing in data centers.
- Power saving in multi-tier systems.
- Heterogeneous multi-kernel Operating System. Most system researches start off with the same observation: that CPU clock rates are stagnating and the way to improve performance is now to scale horizontally by adding more cores to the machine, rather than vertically by increasing the speed of the CPU. Meanwhile, there are heterogeous processing devices that can be used as co-processors or independent processors, like GPGPU, Xeon Phi(MIC), etc. Many challenges emerge in managing heterogeneous processing units in a single computer. Main problems include efficiently control access to shared resources between multiple kernels and processing units, the scalability of system management(data structure and physical resources), the mangement of multiple flexible systems, the power efficience problem in multi-kernel systems, etc. We are currently building new multi-kernel operating system in a single architecture-heterogeous manycore server node.
- HTC-testbed. Our Data center computing testbed HTC-testbed v1.0
has been released in Sep. 2011.
In the area of HTC, Internet enterprises are conservative to share their real data, including the applications and workloads on real services, for reasons of "commercial confidentiality". As a result, the whole research community is facing the trouble of data lock-in. Our goal is to create a public testbed of data center computing so as to solve the data lock-in problem. In the first version of HTC-testbed, we have build basic services both in the front and back ends. The scale of the data has climbed to 5 TB. We are currently making it more applicative for diverse hardware.
- PowerTracer. Saving power has recently become a really hot topic. However, Reachers doesn't pay enough attention to power savings in multi-tier systems. Existing approaches are not satisfactory to save power in data centers, though there are many ways of saving power in an individual node. Our tool PowerTracer focuses on diagnose energy inefficiency in multi-tier services through tracing requests on systems. The tracing data comes from our another tool PreciseTracer. PreciseTracer is under its final testing and modifying.
- PreciseTracer. As more and more multi-tier services are developed from commercial off-the-shelf components or heterogeneous middleware without source code available, both developers and administrators need a request tracing tool to (1) exactly know how a user request of interest travels through services of black boxes and (2) obtain macro-level user request behaviors of services without manually analyzing massive logs. So we developed the tool PreciseTracer to reach these goals. We designed a precise tracing algorithm to deduce causal paths of requests from interaction activities of components of black boxes. Meanwhile, we present two abstractions, component activity graph and dominated causal path pattern, to represent individual causal paths for each request and macro-level user request behaviors, respectively. Based on these abstractions, we proposed a derived performance metric, latency percentages of components, to enable debugging performance-in-thelarge.
- Requirements change management. Requirements changes are difficult to avoid in software development processes, and how to manage requirements change is still an open issue. I propose an innovative algorithm, named LCS-NP++, for comparing different versions of requirement specifications to decide requirement changes. I have integrated this algorithm into a production software process management platform–Qone, which is developed by Institute of Software, Chinese Academy of Sciences. Our practices show it is effective and efficient in comparing requirement items.
PowerTracer: Tracing Requests in Multi-tier Services to Diagnose Energy Inefficiency [pdf].The 9th International Conference on Autonomic Computing (ICAC'12), 2012. San Jose, California. Accepted as short paper
LogMaster: Mining Event Correlations in Logs of Large-scale Cluster Systems [pdf].The 31st IEEE International Symposium on Reliable Distributed Systems (SRDS 2012). 8th-11th October 2012. Irvine, California.
Characterization of Real Workloads of Web Search Engines [pdf].IEEE International Symposium on Workload Characterization (IISWC'11), Nov. 2011. Austin, Texas.
Comparison of Requirement Items Based on the Requirements Change Management System of QONE [pdf].Second WRI World Congress on Software Engineering (WCSE'10), Dec. 2010. Wuhai, China.
PowerTracer: Tracing Requests in Multi-tier Services to Reduce Energy Inefficiency [pdf].IEEE Transaction on Computers. Accepted, March, 2014.
CloudRank-D: Benchmarking and Ranking Cloud Computing Systems for Data Processing Applications [pdf].Frontiers of Computer Science (Front. Comput. Sci.), 2012, 6(4): 347–362.
Precise, Scalable, and Online Request Tracing of Multi-tier Services of Black Boxes [pdf]. PreciseTracer source code. User manual.IEEE Transaction on Parallel and Distributed Systems (TPDS), Aug. 2011
- Outstanding Commy of Graduate University of Chinese Academy of Sciences (GUCAS), 2011
- Outstanding College Graduate of HUST, 2010
- National Fellowship, 2009
- National Encouragement Scholarship, 2008
- Outstanding Student of HUST, 2008
- 2rd Rank in the Calligraphy Competition of Department, 2007
- The National Scholarship, 2007
- Outstanding Individual of the Long March in Morality of HUST, 2007
- Outstanding Student of HUST, 2007
- The People’s Scholarship of Huazhong University of Science and Technology (HUST), 2006-2009.