Kien T. Le
Department Computer Science
110 Frelinghuysen Road
Piscataway, NJ 08854-8019
732 445 6450x4634
Core 340 - Panic Lab
Parasol and GreenSwitch: Managing Datacenters Powered by Renewable EnergyTo appear In Proceedings of ASPLOS 2013.
Ph.D. Thesis, January 2013.
Optimizing the Use of Request Distribution and Stored Energy for Cost Reduction in Multi-Site Internet ServicesIn Proceedings of SustainIT 2012.
GreenHadoop: Leveraging Green Energy in Data-Processing FrameworksIn Proceedings of Eurosys 2012.
GreenSlot: Scheduling Energy Consumption in Green DatacentersIn Proceedings of Super Computing (SC11), November 2011.
In Proceedings of Super Computing (SC11), November 2011.
In Proceedings of the 31st Int'l Conference on Distributed Computing Systems (ICDCS 2011), June 2011.
In Proceedings of the First International Green Computing Conference (IGCC'10), August 2010.Winner of the Best Paper Award.
- Poster abstract, In Proceeding of the ACM SIGMETRICS, June 2010.
- In Proceeding of the Workshop on Power Aware Computing and Systems (HotPower), October 2009.
- Technical Report DCS--TR--639, Dept. of Computer Science, Rutgers University, July 2008.
- In Proceedings of the 8th International Middleware Conference (Middleware 2007), November 2007.
- Rutgers Mini Green Datacenter with Ricardo
Bianchini and Thu Nguyen
Exploring scheduling and energy management infrastructure in a green datacenter.
- Building a free-air cooled datacenter comprising a cluster powered by solar and wind energy.
- Designing models for predicting the varying availability of green energy.
- Exploring scheduling paradigms (e.g., pre-computation) to maximize usage of green energy.
- Multi-Datacenter Internet Services with Ricardo
Bianchini, Thu Nguyen, and Margaret Martonosi
Built a framework to distribute load across geographically distributed data centers that minimizes energy costs while satisfying service level agreements.
- Accounted for server and cooling energy usage and datacenter peak power demand.
- Leveraged differences in energy costs arising from different timezones, variable electricity prices, and outside temperatures.
- Explored several optimization and heuristic-based approaches.
- Explored the implications of caps on brown energy consumption.
- WanStore with Ricardo Bianchini and Thu Nguyen
Built a framework to distribute requests across multiple back-end services that minimizes cost while satisfying service level agreements.
- Considered back-end services with distinct performance vs. cost characteristics.
- Explored two optimization approaches and showed that a hybrid strategy was the best.
- Built a file system that optimizes over multiple third-party block
storage Web services.
- Wayfinder Federated File System with Thu Nguyen and
Built semantic directories, which are searches persistently embedded in a file system namespace.
- Explored machine learning techniques for improving relevancy of matches.
- Participated in evaluation of different distributed search algorithms.
- Built general content parsing infrastructure.
- Resource profiling in JikesRvm with Thu Nguyen and Barbara Ryder
Built a fine-grain profiling framework that uses the processor hardware counters.
- Ported the JikesRvm hardware monitoring package to the Intel platform.
- Improved the package API to allow finer grain measurements.
- Officially contributed package back to the project (http://jikesrvm.org/Acknowledgments).