Upstream: Exposing Performance Information from Cloud Providers to Tenants

• Adriano Lange Universidade Federal do Paraná
• Marcos Sunye Universidade Federal do Parana
• Luis Carlos Bona Universidade Federal do Parana

DOI: https://doi.org/10.5753/wscad.2019.8673

Resumo

Infrastructure-as-a-Service (IaaS) is a widely adopted cloud computing paradigm due to its flexibility and competitive prices. To improve resource efficiency, most IaaS providers consolidate several tenants in the same virtualization server, which usually incurs variable performance experiences. In this paper, we evaluate the CPU time received by tenants’ virtual machines (VMs). We present a model that estimates the probability of a VM to receive, at least, a determined fraction of CPU time using limited information about the host and VMs running on it. We constructed this model using a series of experiments with different numbers of CPU cores and co-located VMs.

Referências

[1] SPEC - Standard Performance Evaluation Corporation, . URL http://www.spec.org/.

[2] Stress-ng, . URL http://kernel.ubuntu.com/{˜}cking/stress-ng.

[3] G. Aceto, A. Botta, W. de Donato, and A. Pescapè. Cloud monitoring: A survey. Computer Networks, 57(9):2093–2115, 2013. doi: 10.1016/j.comnet.2013.04.001.

[4] M. Ahmad, S. Duan, A. Aboulnaga, and S. Babu. Predicting completion times of batch query workloads using interaction-aware models and simulation. In International Conference on Extending Database Technology (EDBT/ICDT), page 449, Uppsala, Sweden, 2011. ACM Press. doi: 10.1145/1951365.1951419.

[5] M. Armbrust. Above the Clouds: A Berkeley View of Cloud Computing. Technical Report UCB/EECS-2009-28, EECS Department, University of California, Berkeley, 2009.

[6] G. Costa. KVM-specific MSRs. URL https://github.com/torvalds/linux/blob/master/Documentation/virtual/kvm/msr.txt.

[7] C. B. Hauser. Kvmtop. URL https://github.com/cha87de/kvmtop.

[8] C. B. Hauser and S. Wesner. Reviewing Cloud Monitoring: Towards Cloud Resource Profiling. In IEEE 11th International Conference on Cloud Computing (CLOUD), pages 678–685, 2018. doi: 10.1109/CLOUD.2018.00093.

[9] C. B. Hauser, J. Domaschka, and S. Wesner. Predictability of Resource Intensive Big Data and HPC Jobs in Cloud Data Centres. In IEEE International Conference on Software Quality, Reliability and Security Companion (QRS-C), pages 358–365, 2018. doi: 10.1109/QRS-C.2018.00069.

[10] A. Lange. Upstream Repository. URL https://github.com/alange0001/upstream.

[11] F. Licht, B. Schulze, L. C. E. Bona, and A. R. Mury. Analysis of parallelized libraries and interference effects in concurrent environments. Computers & Electrical Engineering, 66:435–453, 2018. doi: 10.1016/j.compeleceng.2017.08.028.

[12] NASA. NAS Parallel Benchmarks. URL https://www.nas.nasa.gov/publications/npb.html.

[13] J. Schad, J. Dittrich, and J.-A. Quiané-Ruiz. Runtime Measurements in the Cloud: Observing, Analyzing, and Reducing Variance. Proc. VLDB Endow., 3(1-2):460–471, 2010. doi: 10.14778/1920841.1920902.

[14] I. Stoica, H. Abdel-wahab, and K. Jeffay. On the Duality between Resource Reservation and Proportional Share Resource Allocation. In In Proc. of Multimedia Computing and Networking, pages 207–214, 1997.

[15] F. Wu, Q. Wu, and Y. Tan. Workflow scheduling in cloud: a survey. The Journal of Supercomputing, 71(9):3373–3418, 2015. doi: 10.1007/s11227-015-1438-4.

[16] W. Wu, Y. Chi, S. Zhu, J. Tatemura, H. Hacigümüs, and J. F. Naughton. Predicting Query Execution Time: Are Optimizer Cost Models Really Unusable? In International Conference on Data Engineering (ICDE), pages 1081–1092, 2013. doi: 10.1109/ICDE.2013.6544899.

[17] W. Wu, X. Wu, H. Hacigümüs, and J. F. Naughton. Uncertainty Aware Query Execution Time Prediction. Proc. VLDB Endow., 7(14):1857–1868, 2014. doi:10.14778/2733085.2733092.