Page 66 - profile2014.indd

研究員
吳真貞 Jan-Jan Wu

Research Fellow
Ph.D., Computer Science, Yale University
Tel: +886-2-2788-3799 ext. 1610 Fax: +886-2-2782-4814
Email: wuj@iis.sinica.edu.tw
http://www.iis.sinica.edu.tw/pages/wuj

● Research Fellow, IIS, Academia Research Description
Sinica (2011-)
● Associate Research Fellow, IIS, My research interests include cloud computing, and virtualization on multicores. My re-
Academia Sinica (2001-2011) search in cloud computing focuses on two areas: cloud OS (Infrastructure as a Service) and
● Assistant Research Fellow, IIS, cloud databases. For cloud OS, we jointly developed a cloud OS prototype, Roystonea, with
Academia Sinica (1996-2001) National Taiwan University. Roystonea provides dynamic provision, scheduling, and man-
agement of virtual machines, automatic scaling of system resources for application service
● Associate Software Engineer, requirements, and dynamic resource management for performance/energy tradeo .

Institute of Information Industry
(1987-1989) With the growth of data size in many application domains, the need for data management
● Ph.D., Computer Science, Yale systems is continuously increasing. We have developed a data management system for
University (1995) cloud, named SQLMR. SQLMR compiles SQL queries to a sequence of MapReduce jobs. Ex-
isting SQL-based applications are seamlessly compatible with SQLMR, and users can man-
● M.S., CSIE, National Taiwan Uni-

versity (1987) age SQL queries at the Tera to PataByte scale. Our future e orts will focus on cloud database

virtualization. The goal is to enable applications with SQL queries to run on di erent NoSQL
● B.S., CSIE, National Taiwan Univer- platforms transparently and e ciently.

sity (1985)
In another line of work, we focus on the core technologies that support virtualization, par-
ticularly dynamic binary translation (DBT) and binary optimization on multicore systems.
At present, QEMU is the most popular re-targetable DBT system that supports virtualization for multiple guest and host ISAs and OS’s.
QEMU requires 10X or even more runtime overhead in practice. It also requires sophisticated and elaborate binary optimization schemes
to cut down the overhead. Through our current work, we have developed a multi-threaded DBT prototype, called HQEMU, using QEMU

and LLVM (Low Level Virtual Machine) as our building blocks. This is the rst e ort to successfully combine QEMU and LLVM, and to sig-

ni cantly improve the performance of QEMU. We are extending our current work to develop binary manipulation technologies that are

crucial to support system-level virtualization.
Publications
1. Ding-Yong Hong, Jan-Jan Wu, Pen-Chung Yew, Wei- 6. Chun-Chen Hsu, Pangfeng Liu, Chien-Min Wang, Jan-Jan
Chung Hsu, Pangfeng Liu, Chien-Min Wang, Yeh-Ching Wu, Ding-Yong Hong, Pen-Chung Yew and Wei-Chung Hsu,
Chung,”Efficient and Retargetable Dynamic Binary Transla- “LnQ: Building High Performance Dynamic Binary Transla-

tion on Multicores”, IEEE Transactions on Parallel and Dis- tors with Existing Compiler Backends,” Proceedings of the
tributed Systems, vol.25,no.3, Feb. 2014 40th International Conference on Parallel Processing (ICPP),
September 2011.
2. Yi-Hong Lyu, Ding-Yong Hong, Tai-Yi Wu, Jan-Jan Wu,
Pangfeng Liu, Wei-Chung Hsu,”DBILL: An Efficient and Re- 7. Chao-En Yen, Jyun-Shiung Yang, Pangfeng Liu and Jan-Jan

targetable Dynamic Binary Instrumentation Framework using Wu, “Roystonea: A Cloud Computing System with Pluggable
LLVM Backend,” ACM International Conference on Virtual Component Architecture,” International Conference on Par-
Execution Environments(VEE), March 2014. allel and Distributed Systems (ICPADS), December 2011.
3. Li-Yung Ho, Tsung-Han Li, Jan-Jan Wu, Pangfeng Liu, “Ky- 8. Meng-Ju Hsieh, Chao-Rui Chang, Li-Yung Ho, Jan-Jan Wu,
lin: An Efficient and Scalable Graph Data Processing Sys- “SQLMR: A scalable database management system for cloud

tem”, IEEE International Conference on Big Data, Oct. 2013 computing,” the 40 International Conference on Parallel
th
Processing (ICPP), pages 315-324, September 2011.
4. Chao-Jui Chang, Jan-Jan Wu, Wei-Chung Hsu, Pangfeng Liu,
“Efficient Memory Virtualization for Cross-ISA System Mode 9. Li-Yung Ho, Jan-Jan Wu, P. Liu, “Optimal Algorithms for

Emulation”, ACM International Conference on Virtual Execu- Cross-Rack Communication Optimization in MapReduce
tion Environments(VEE), March 2014. Framework,” IEEE International Conference on Cloud Com-
puting (IEEE CLOUD), Washington DC, USA, July 2011.
5. Ding-Yong Hong, Chun-Chen Hsu, Pen-Chung Yew, Jan-Jan
Wu, Wei-Chung Hsu, Pangfeng Liu, Chien-Min Wang, Yeh- 10. Ching-Chi Lin, P. Liu, Jan-Jan Wu, “Energy-Aware Virtual
Ching Chung, “HQEMU: A Multithreaded and Retargetable Machine Dynamic Provisioning and Scheduling for Cloud
Dynamic Binary Translator on Multicores,” International Computing,” IEEE International Conference on Cloud Com-
Symposium on Code Generation and Optimization (CGO), puting (IEEE CLOUD), pages 461-462, Washington DC,
March 2012. USA, July 2011.

66 研究人員 Research Faculty