OS enhancements for data-intensive server systems
Introduction:
Recent studies on operating system support for concurrent server systems mostly target CPU-intensive workloads with light disk I/O activities. However, an important class of server systems that access a large amount of  disk-resident data, such as the index searching server of large-scale Web search engines, has received limited attention. In this work, we will examine operating system techniques to improve the performance of data-intensive server systems under concurrent execution. We look at OS enhancements in three aspects of the operating system: file system prefetching, memory management, and disk I/O management. We have proposed a competitive  prefetching strategy that can balance the overhead of disk I/O switching and the wasted I/O bandwidth of prefetching unnecessary data. We have explored new memory management approaches for prefetched data, in order to reduce prefetching-incurred page thrashing at high concurrency levels.
Publications:
  • Chuanpeng Li, Kai Shen, and Athanasios E. Papathanasiou, "Competitive Prefetching for Concurrent Sequential I/O". in Proc. of the 2nd ACM EuroSys Conference, Lisbon, Portugal, March 2007. PDF version.
  • Kai Shen, Ming Zhong, and Chuanpeng Li, "I/O System Performance Debugging Using Model-driven Anomaly Characterization". In Proc. of the 4th USENIX Conference on File and Storage Technologies (FAST'05), San Francisco, CA, December 2005.Abstract. Postscript version. PDF version. Kernel patch code release for bug fix #1 in the paper.
  • Chuanpeng Li and Kai Shen, "Managing Prefetch Memory for Data-Intensive Online Servers". In Proc. of the 4th USENIX Conference on File and Storage Technologies (FAST'05), San Francisco, CA, December 2005.Abstract. Postscript version. PDF version. Kernel patch code release for the Coldest prefetch memory policy in the paper.
  • Chuanpeng Li, Athanasios E. Papathanasiou, and Kai Shen, "Competitive Prefetching for Data-Intensive Online Servers". In the First Workshop on Operating System and Architectural Support for the on demand IT InfraStructure (OASIS'04), Held along with ASPLOS-XI, Boston MA, October 2004.Abstract. Postscript version. PDF version.
Support: This project is supported by the National Science Foundation CAREER award CCF-0448413.
Note: Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.