Abstract
We introduce the FORSETI system, which is a principled approach for holistic memory management. It permits a sysadmin to specify the total physical memory resource that may be shared between all concurrent virtual machines on a physical node. FORSETI models the heap size versus application throughput for each virtual machine, and seeks to maximize the combined throughput of the set of VMs based on concepts from economic utility theory. We evaluate the FORSETI system using a standard Java managed runtime, i.e. OpenJDK. Our results demonstrate that FORSETI enables dramatic reductions (up to 5x) in heap footprint without compromising application execution times.
- Amazon elastic mapreduce. Checked 10 Feb 2015, http://docs.aws.amazon.com/ElasticMapReduce/ latest/DeveloperGuide/MemoryTuning.html.Google Scholar
- The neo4j manual. Checked 25 Nov 2014, http://neo4j. com/docs/stable/configuration-jvm.html.Google Scholar
- R. Alonso and A. W. Appel. An advisor for flexible working sets. In Proceedings of the ACM SIGMETRICS Conference on Measurement and Modeling of Computer Systems, pages 153–162, 1990. Google Scholar
Digital Library
- S. L. Bird and B. J. Smith. Pacora: Performance aware convex optimization for resource allocation. In Proceedings of the 3rd USENIX Workshop on Hot Topics in Parallelism, 2011.Google Scholar
- S. M. Blackburn, P. Cheng, and K. S. McKinley. Myths and realities: The performance impact of garbage collection. ACM SIGMETRICS Performance Evaluation Review, 32(1):25–36, 2004. Google Scholar
Digital Library
- S. M. Blackburn, R. Garner, C. Hoffman, A. M. Khan, K. S. McKinley, R. Bentzur, A. Diwan, D. Feinberg, D. Frampton, S. Z. Guyer, M. Hirzel, A. Hosking, M. Jump, H. Lee, J. E. B. Moss, A. Phansalkar, D. Stefanovi´c, T. VanDrunen, D. von Dincklage, and B. Wiedermann. The DaCapo benchmarks: Java benchmarking development and analysis. In Proceedings of the ACM SIGPLAN Conference on Object-Oriented Programming Systems, Languages, and Applications, pages 169–190, 2006. Google Scholar
Digital Library
- R. H. Byrd, P. Lu, J. Nocedal, and C. Zhu. A limited memory algorithm for bound constrained optimization. SIAM Journal on Scientific and Statistical Computing, 16(5):1190–1208, 1995. Google Scholar
Digital Library
- C. Cameron and J. Singer. We are all economists now: Economic utility for multiple heap sizing. In Proceedings of the 9th Implementation, Compilation, Optimization of OO Languages, Programs and Systems workshop, 2014. Google Scholar
Digital Library
- C. W. Cobb and P. H. Douglas. A theory of production. American Economic Review, 18(1):139–165, 1928.Google Scholar
- D. Detlefs, C. Flood, S. Heller, and T. Printezis. Garbage-first garbage collection. In Proceedings of the ACM SIGPLAN International Symposium on Memory Management, pages 37– 48, 2004. Google Scholar
Digital Library
- C. Grzegorczyk, S. Soman, C. Krintz, and R. Wolski. Isla Vista heap sizing: Using feedback to avoid paging. In Code Generation and Optimization, 2007. CGO ’07. International Symposium on, pages 325–340, 2007. Google Scholar
Digital Library
- M. Hertz, S. Kane, E. Keudel, T. Bai, C. Ding, X. Gu, and J. E. Bard. Waste not, want not: resource-based garbage collection in a shared environment. In Proceedings of the ACM SIGPLAN International Symposium on Memory Management, pages 65– 76, 2011. Google Scholar
Digital Library
- M. R. Hines, A. Gordon, M. Silva, D. Da Silva, K. D. Ryu, and M. Ben-Yehuda. Applications know best: Performancedriven memory overcommit with Ginkgo. In IEEE Third International Conference on Cloud Computing Technology and Science, pages 130–137, 2011. Google Scholar
Digital Library
- S. Kim, H. Kim, J. Lee, and J. Jeong. Group-based memory oversubscription for virtualized clouds. Journal of Parallel and Distributed Computing, 74(4):2241–2256, 2014. Google Scholar
Digital Library
- P. Lengauer and H. Mössenböck. The taming of the shrew: Increasing performance by automatic parameter tuning for Java garbage collectors. In Proceedings of the 5th ACM/SPEC International Conference on Performance Engineering, pages 111–122, 2014. Google Scholar
Digital Library
- P. J. Lloyd. The origins of the von Thunen-Mill-Pareto-Wicksell-Cobb-Douglas function. History of Political Economy, 33(1):1–19, 2001.Google Scholar
Cross Ref
- M. Maas, K. Asanovic, T. Harris, and J. Kubiatowicz. The case for the holistic language runtime system. In Proceedings of the First International Workshop on Rack-scale Computing, 2014.Google Scholar
- M. Raghavachari, D. Reimer, and R. D. Johnson. The deployer’s problem: Configuring application servers for performance and reliability. In Proceedings of the 25th International Conference on Software Engineering, pages 484–489, 2003. Google Scholar
Digital Library
- T.-I. Salomie, G. Alonso, T. Roscoe, and K. Elphinstone. Application level ballooning for efficient server consolidation. In Proceedings of the 8th ACM European Conference on Computer Systems, pages 337–350, 2013. Google Scholar
Digital Library
- Y. C. Tay, X. Zong, and X. He. An equation-based heap sizing rule. Performance Evaluation, 70(11):948–964, 2013. Google Scholar
Digital Library
- D. Vengerov. A reinforcement learning approach to dynamic resource allocation. Engineering Applications of Artificial Intelligence, 20(3):383–390, 2007. Google Scholar
Digital Library
- D. Vengerov. Modeling, analysis and throughput optimization of a generational garbage collector. In Proceedings of the ACM SIGPLAN International Symposium on Memory Management, SIGPLAN International Symposium on Memory Management, pages 1–9, 2009. Google Scholar
Digital Library
- D. R. White, J. Singer, J. M. Aitken, and R. E. Jones. Control theory for principled heap sizing. In Proceedings of the ACM SIGPLAN International Symposium on Memory Management, pages 27–38, 2013. Google Scholar
Digital Library
- T. Yang, M. Hertz, E. D. Berger, S. F. Kaplan, and J. E. B. Moss. Automatic heap sizing: Taking real memory into account. In Proceedings of the ACM SIGPLAN International Symposium on Memory Management, pages 61–72, 2004. Google Scholar
Digital Library
- T. Yang, E. D. Berger, S. F. Kaplan, and J. E. B. Moss. Cramm: Virtual memory support for garbage-collected applications. In Proceedings of the 7th Symposium on Operating Systems Design and Implementation, pages 103–116, 2006. Google Scholar
Digital Library
- C. Zhang, K. Kelsey, X. Shen, C. Ding, M. Hertz, and M. Ogihara. Program-level adaptive memory management. In Proceedings of the 5th International Symposium on Memory Management, pages 174–183, 2006. Google Scholar
Digital Library
- W. Zhao, Z. Wang, and Y. Luo. Dynamic memory balancing for virtual machines. SIGOPS Oper. Syst. Rev., 43(3):37–47, 2009. Google Scholar
Digital Library
- C. Zhu, R. Byrd, J. Nocedal, and J. L. Morales. Software for large-scale bound-constrained optimization, 2011. http://users.iems.northwestern.edu/~nocedal/ lbfgsb.html. Introduction Current Practice Forseti System Contributions Motivation Microeconomic Theory Forseti System Forseti Daemon Individual Throughput Functions Making Recommendations VM Lifecycle Modified VMs Calculating Throughput Proxy Library HotSpot JVM Evaluation Experimental Platform Selecting Experimental Workloads Selecting Memory Target Sizes Illustrative Examples Experimental Methodology Results Overhead Related Work Future Work ConclusionsGoogle Scholar
Index Terms
The judgment of forseti: economic utility for dynamic heap sizing of multiple runtimes
Recommendations
The judgment of forseti: economic utility for dynamic heap sizing of multiple runtimes
ISMM '15: Proceedings of the 2015 International Symposium on Memory ManagementWe introduce the FORSETI system, which is a principled approach for holistic memory management. It permits a sysadmin to specify the total physical memory resource that may be shared between all concurrent virtual machines on a physical node. FORSETI ...
Control theory for principled heap sizing
ISMM '13: Proceedings of the 2013 international symposium on memory managementWe propose a new, principled approach to adaptive heap sizing based on control theory. We review current state-of-the-art heap sizing mechanisms, as deployed in Jikes RVM and HotSpot. We then formulate heap sizing as a control problem, apply and tune a ...
Inter-rack live migration of multiple virtual machines
VTDC '12: Proceedings of the 6th international workshop on Virtualization Technologies in Distributed Computing DateWithin datacenters, often multiple virtual machines (VMs) need to be live migrated simultaneously for various reasons such as maintenance, power savings, and load balancing. Such mass simultaneous live migration of multiple VMs can trigger large data ...






Comments