Yang Wang
Abstract
With the development of storage technology and applications, new caching policies are continuously being introduced. It becomes increasingly important for storage systems to be able to select the matched caching policy dynamically under varying workloads. This article proposes SOPA, a cache framework to adaptively select the matched policy and perform policy switches in storage systems. SOPA encapsulates the functions of a caching policy into a module, and enables online policy switching by policy reconstruction. SOPA then selects the policy matched with the workload dynamically by collecting and analyzing access traces. To reduce the decision-making cost, SOPA proposes an asynchronous decision making process. The simulation experiments show that no single caching policy performed well under all of the different workloads. With SOPA, a storage system could select the appropriate policy for different workloads. The real-system evaluation results show that SOPA reduced the average response time by up to 20.3% and 11.9% compared with LRU and ARC, respectively.
- Ari, I., Amer, A., Gramarcy, R., Miller, E., Brandt, S., and Long, D. 2002. ACME: Adaptive caching using multiple experts. In Proceedings of the Workshop on Distributed Data and Structures, Carleton Scientific, 2002.Google Scholar
- Ari, I., Gottwals, M., and Henze, D. 2004. SANBoost: Automated SAN-Level caching in storage area networks, In Proceedings of the International Conference on Autonomic Computing (ICAC). Google ScholarDigital Library
- Bansal, S. and Modha, D. S. 2004. CAR: Clock with adaptive replacement. In Proceedings of the USENIX File and. Storage Technologies Conference (FAST), 142--163. Google ScholarDigital Library
- Cao. P. and Irani, S. 1997. Cost-aware WWW proxy caching algorithms. In Proceedings of the USENIX Symposium on Internet Technologies and Systems. Google ScholarDigital Library
- Chen, Z., Zhou, Y., and Li, K. 2003. Eviction-based cache placement for storage caches. In Proceedings of the USENIX Annual Technical Conference, 269--281.Google Scholar
- Chen, Z., Zhang, Y., Zhou, Y., Scott, H., and Schiefer, B. 2005. Empirical evaluation of multi-level buffer cache collaboration for storage systems. In Proceedings of the ACM SIGMETRICS Conference on Measurement and Modeling of Computer Systems (SIGMETRICS), 2005. Google ScholarDigital Library
- Chou, H. T. and Dewitt, D. J. 1985. An evaluation of buffer management strategies for relational database systems. In Proceedings of the VLDB Conference. Google ScholarDigital Library
- Gill, B. S. and Modha, D. S. 2005. WOW: Wise ordering for writes?combining spatial and temporal locality in non-volatile caches. In Proceedings of the USENIX Conference on File and Storage Technologies (FAST). Google ScholarDigital Library
- HP TRACES. http://tesla.hpl.hp.com/public_software/Google Scholar
- Jiang, S. and Zhang, X. 2002. LIRS: An efficient low inter-reference recency set replacement policy to improve buffer cache performance. In Proceedings of the ACM SIGMETRICS Conference on Measurement and Modeling of Computer Systems. 31--42. Google ScholarDigital Library
- Jiang. S., Ding, X., Chen, F., Tan, E., and Zhang. X. 2005. DULO: An effective buffer cache management scheme to exploit both temporal and spatial localities. In Proceedings of the USENIX Conference on File and Storage Technologies (Fast). Google ScholarDigital Library
- Johnson, T. and Shasha, D. 1994. 2Q: A low overhead high performance buffer management replacement algorithm,” In Proceedings of the VLDB Conference. 297--306. Google ScholarDigital Library
- Lee, D., Choi, J., Kim, J. H., Noh, S. H., Mim, S. L., Cho, Y., and Kim, C. S. 2001. LRFU: A spectrum of policies that subsumes the least recently used and least frequently used policies. IEEE Trans. Comput. 50, 12, 1352--1360. Google ScholarDigital Library
- Li, X., Aboulnaga, A., Sachedina, A., Salem, K., and Gao, S. B. 2005. Second-tier cache management using write hints. In Proceedings of the 4th USENIX Conference on File and Storage Technologies (FAST), 115--128. Google ScholarDigital Library
- Mcvoy, L. and Staelin, C. 1996. lmbench: Portable tools for performance analysis. In Proceedings of the USENIX Technical Conference. 279--295. Google ScholarDigital Library
- Megiddo, N. and Modha, D. S. 2003. ARC: A self-tuning, low overhead replacement cache. In Proceedings of the USENIX File and Storage Technologies Conference (FAST), 115--130. Google ScholarDigital Library
- Menon, J. 1994. Performance of RAID5 disk arrays with read and write caching. Distrib. Parall. Datab. 2, 3, 261--293. Google ScholarDigital Library
- Menon, J. and Hartung, M. 1988. The IBM 3990 disk cache. In Proceedings of the IEEE Computer Society International COMPCON Conference.Google Scholar
- Nelson, M. N., Welch, B. B., and Ousterhout, J. K. 1998. Caching in the Sprite network file system. ACM Trans. Comput. Syst. 6, 1, 134--154. Google ScholarDigital Library
- O'neil, E. J., O'neil, P. E., and Weikum, G. 1993. The LRU-K page replacement algorithm for database disk buffering. In Proceedings of the International Conference on Management of Data, 297--306. Google ScholarDigital Library
- Robinson, J. T. and Devarakonda, M. V. 1990. Data cache management using frequency-based replacement. In Proceedings of the ACM SIGMETRIC Conference on Measuring and Modeling of Computer Systems, 134--142 Google ScholarDigital Library
- Ruemmler, C. and Wilkes, J. 1993. A trace-driven analysis of disk working set sizes. Tech. rep. HPL{OSR{93{23, Hewlett-Packard Laboratories, Palo Alto, CA, USA.Google Scholar
- Salmon, B., Thereska, E., Soules, C. A. N, and Ganger, G. R. 2003. A two-tiered software architecture for automated tuning of disk layouts. In Proceedings of the Workshop on Algorithms and Architectures for Self-Managing Systems.Google Scholar
- SPC-1 TRACES. http://traces.cs.umass.edu/index.php/Storage/StorageGoogle Scholar
- TPC-C TRACE. http://tds.cs.byu.edu/tds/tracelist.jsp?searchby=attribute&type=Disk+I%2FO&length=All&platform=All&cache=All&pageNum=0&searchAction=Go&x=52&y=19Google Scholar
- UNH PROJECT. 2006. http://unh-iscsi.sourceforge.net/Google Scholar
- Yadgar, G. and Factor, M. 2007. Karma: Know-it-all replacement for a multilevel cache. In Proceedings of the USENIX File and Storage Technologies Conference (FAST). Google ScholarDigital Library
- Zhou, Y. and Philbin, J. F. 2001. The multi-queue replacement algorithm for second level buffer caches. In Proceedings of the USENIX Annual Technical Conference. 91--104. Google ScholarDigital Library
Index Terms
- SOPA: Selecting the optimal caching policy adaptively
Recommendations
Caching across heterogeneous information sources: an object-based approach
Information processing and technologyInformation exchange has been increased drastically and rapidly due to the major large scale Internet expansion. The structure of information sources worldwide has been altered and both structured and semi-structured data are stored in various ...
Streaming Machine Learning for Supporting Data Prefetching in Modern Data Storage Systems
AI4Sys '23: Proceedings of the First Workshop on AI for SystemsModern data storage systems optimize data access by distributing data across multiple storage tiers and caches, based on numerous tiering and caching policies. The policies' decisions, and in particular the ones related to data prefetching, can severely ...
Performance study of a collaborative method for hierarchical caching in proxy servers
WWW7: Proceedings of the seventh international conference on World Wide Web 7
Reviews
Seetharami R Seelam
Wang et al. describe a framework for adaptively selecting input/output (I/O) caching policy when given a set of policies. The main idea is to capture a set of policies and the data required, in the form of a module, and select a policy that is appropriate for a given workload, based on the current policy cache hit rate. If the hit rate falls below a user-definable threshold, other policies are evaluated on the metadata captured from the workload. It is well known that there is no "silver bullet"-that is, there is no single caching policy that will result in optimal caching performance. That being said, this paper provides a sound framework for realizing multiple policies and allowing for optimal cache performance for applications. Online Computing Reviews Service
Access critical reviews of Computing literature here
Become a reviewer for Computing Reviews.
Comments