PaaS LiveMigration 1,a) 1,b) 1,c) (OS) Live Migration Live Migration (VM) VM Live Migration CPU. Live Migration Live Migration PaaS Live Migration Live Migration CPU CPU 26.3%, 10.0% 27.8% Live Migration, Page Sharing 1. (OS) OS (VM) OS Amazon Web Services [1] Windows Azure [2] OS VM Platform as a Service(PaaS) OS Live Migration [3] Live Migration VM VM Live Migration 1 Presently with Keio Uniersity a) uhoi@sslab.ics.keio.ac.jp b) kono@ics.keio.ac.jp c) koto@sslab.ics.keio.ac.jp VM Live Migration Live Migration CPU VM CPU PaaS VM ( ) Live Migration PaaS Live Migration PaaS VM ( ) Live Migration CPU. XEN 4.3.3-rc1, Ubuntu14.04 LTS c 2015 Information Processing Society of Japan 1
VM Live Migration 3. 1 Live Migration CPU [4] CPU 26.3%, 10.0% 27.8% 2 Live Migration 3 4 5 6 2. Live Migration PaaS Live Migration CPU PaaS Live Migration 1 Live Migration VM CPU (a) Source host (b) Target host Migrating VM VM Colocating VM host VM Live Migration Migration VM 1(a) Migration VM ( ) CPU Live Migration 50% 40% Live Migration Migration VM Live Migration Colocating VM 1(a) Colocating VM CPU Migrating VM 40% 1(b) Colocating VM CPU Live Migration 20-50% Migrating VM Live Migration Live Migration Colocating VM Live Migration VM 3.1 PaaS Live Migration VM Live Migration CPU 3.2 PaaS PaaS OS VM VM PaaS PaaS VM VM 3.3 2 XEN4.3.3-rc1 Migration Domain-0 CBPS [5] c 2015 Information Processing Society of Japan 2
3 2 CBPS XEN CBPS Migration VM Migration Pre-copy [3] SHA-1 SHA-1(160bit) 48 nines 0.9999... 9 9 48 TCP 8 9) nines [6] 4. 4.1 3 Source (Destination) XEN 4.3.3-rc1 Domain-0 Ubuntu14.04 LTS Intel Xeon CPU X5650, 16GB Memory, Gigbit Ethernet Migration, Migration VM NFS NFS Ubuntu 14.04 LTS, Intel Xeon CPU X3480, 16GB Memory, Gigbit Ethernet VM Ubuntu 14.04.1 LTS, 512 MB Memory VM VM(Source VM), VM(Migration VM), VM(Destination VM) 4 Micro Benchmark Source VM, Migration VM, Destination VM CBPS Page Sharing Migration VM Destination Live Migration CPU Micro benchmark mmap VM HTTPERF HTTP [7] MEMASLAP memcached [8] KERNEL BUILD kernel 4.2 Micro Benchmark 4 X VM ( ) 0% Page Sharing Y Live Migration First Iterate c 2015 Information Processing Society of Japan 3
1 Micro Benchmark (%) 0.0 25.9 44.4 64.7 82.8 135158 101172 76960 51269 25920 5 Migration Domain-0 CPU (Source) 6 Migration Domain-0 CPU (Destination) Iterate, First Iterate 50 44.5% 1 Micro Benchmark 80.8% Satori [9] I/O HTTPERF 63 % 60% 5 6 Live Migration domain-0 CPU 50 X Source Desitination CPU CPU CPU Micro Benchmark CPU 4.3 HTTPERF, MEMASLAP, KERNEL BUILD 4.3.1 4.3.2 Live Migration VM 4.3.1 8% MEMASLAP 8.5% PaaS OS 8% HTTPERF 27.8%, KERNEL BUILD 11.4%, MEMASLAP 8.4% HTTPERF c 2015 Information Processing Society of Japan 4
(a) HTTPERF (b) MEMASLAP (c) KERNEL BUILD 7 Live Migration Domain-0 CPU HTTPERF 10.0%, KER- NEL BUILD 5.9%, MEMASLAP 8.5% Satori [9] CBPS I/O HTTPERF Satori KERNEL BUILD NFS NFS NFS Migration Live Migration Live Migration Domain-0 CPU 7 Domain-0 CPU USAGE CPU CPU HTTPERF 26.3% CPU Live Migration CPU 4.3.2 VM 4.3.1 Live Migration VM 8 HTTPERF Source VM 7.2% ( ) Live Migration Migration VM Source VM Live Migration 8 Source VM [httpef] VM 2 MEMASLAP Live Migration CPU USAGE 30 Domain-0 CPU Destination VM CPU Domain-0 CPU Destination VM CPU VM Domain-0, Source VM, Migration VM CPU USAGE CPU. Micro Benchmark CPU USAGE VM Live Migration 5. Live Migration Miyakodori[10] Live Migration VM, Live Migration c 2015 Information Processing Society of Japan 5
2 Memaslap Live Migration CPU Average CPU USAGE AVERAGE (%) [MEMASLAP] (%) Domain-0(Destination) Destination VM 0.0 56.0 32.0 8.3 55.0 36.0 VM Sonic Migration [11] Live Migration PaaS VM Page Sharing Disco[12] VM VMM TPS(Transparent Page Sharing) VM Ware [5] CBPS Satori [9] VM VM Memory Buddies [13] OS GUI ASLR [14] Memory Buddies PaaS Live Migration Page Sharing Page Sharing Live Migration 6. PaaS VM Live Migration Live Migration CPU Live Migration VM Live Migration PaaS PaaS Live Migration CPU 26.3%, 10.0% 27.8% VM Live Migration [1] Amazon: Amazon Web Services. http://aws.amazon. com/jp/. [2] Microsoft: Azure. http://azure.microsoft.com/ ja-jp/. [3] Clark, C., Fraser, K., Hand, S., Hansen, J. G., Jul, E., Limpach, C., Pratt, I. and Warfield, A.: Live Migration of Virtual Machines, Proceedings of the 2Nd Conference on Symposium on Networked Systems Design & Implementation - Volume 2, NSDI 05, Berkeley, CA, USA, USENIX Association, pp. 273 286 (online), available from http://dl.acm.org/citation.cfm?id=1251203.1251223 (2005). [4] Koto, A., Kono, K. and Yamada, H.: A Guideline for Selecting Live Migration Policies and Implementations in Clouds, Cloud Computing Technology and Science (CloudCom), 2014 IEEE 6th International Conference on, pp. 226 233 (online), DOI: 10.1109/Cloud- Com.2014.36 (2014). [5] Waldspurger, C. A.: Memory Resource Management in VMware ESX Server, SIGOPS Oper. Syst. Rev., Vol. 36, pp. 181 194 (online), available from http://doi.acm.org/10.1145/844128.844146 (2002). [6] Henson, V.: Guidelines for Using Compare-by-hash (2005). [7] Hewlett-Packard Development Company, L.: httperf. http://www.hpl.hp.com/research/linux/httperf/. [8] Zhuang, M.: memaslap. http://docs.libmemcached. org/bin/memaslap.html. [9] Mi lós, G., Murray, D. G., Hand, S. and Fetterman, M. A.: Satori: Enlightened Page Sharing, Proceedings of the 2009 Conference on USENIX Annual Technical Conference, USENIX 09, Berkeley, CA, USA, USENIX Association, pp. 1 1 (online), available from http://dl.acm.org/citation.cfm?id=1855807.1855808 (2009). [10] Akiyama, S., Hirofuchi, T., Takano, R. and Honiden, S.: MiyakoDori: A Memory Reusing Mechanism for Dynamic VM Consolidation, Proceedings of the 2012 IEEE Fifth International Conference on Cloud Computing, CLOUD 12, Washington, DC, USA, IEEE Computer Society, pp. 606 613 (online), DOI: c 2015 Information Processing Society of Japan 6
10.1109/CLOUD.2012.56 (2012). [11] Koto, A., Yamada, H., Ohmura, K. and Kono, K.: Towards Unobtrusive VM Live Migration for Cloud Computing Platforms, Proceedings of the Asia- Pacific Workshop on Systems, APSYS 12, New York, NY, USA, ACM, pp. 7:1 7:6 (online), DOI: 10.1145/2349896.2349903 (2012). [12] Bugnion, E., Devine, S., Govil, K. and Rosenblum, M.: Disco: Running Commodity Operating Systems on Scalable Multiprocessors, ACM Trans. Comput. Syst., Vol. 15, No. 4, pp. 412 447 (online), DOI: 10.1145/265924.265930 (1997). [13] Wood, T., Tarasuk-Levin, G., Shenoy, P., Desnoyers, P., Cecchet, E. and Corner, M. D.: Memory Buddies: Exploiting Page Sharing for Smart Colocation in Virtualized Data Centers, Proceedings of the 2009 ACM SIGPLAN/SIGOPS International Conference on Virtual Execution Environments, VEE 09, New York, NY, USA, ACM, pp. 31 40 (online), DOI: 10.1145/1508293.1508299 (2009). [14] Barker, S., Wood, T., Shenoy, P. and Sitaraman, R.: An Empirical Study of Memory Sharing in Virtual Machines, Proceedings of the 2012 USENIX Conference on Annual Technical Conference, USENIX ATC 12, Berkeley, CA, USA, USENIX Association, pp. 25 25 (online), available from http://dl.acm.org/citation.cfm?id=2342821.2342846 (2012). c 2015 Information Processing Society of Japan 7