An Empirical Analysis of Amazon EC2 Spot Instance Features Affecting Cost-effective Resource Procurement

Many cost-conscious public cloud workloads ("tenants") are turning to Amazon EC2's spot instances because, on average, these instances offer significantly lower prices (up to 10 times lower) than on-demand and reserved instances of comparable advertized resource capacities. To use spot instances effectively, a tenant must carefully weigh the lower costs of these instances against their poorer availability. Towards this, we empirically study four features of EC2 spot instance operation that a cost-conscious tenant may find useful to model. Using extensive evaluation based on both historical and current spot instance data, we show shortcomings in the state-of-the-art modeling of these features that we overcome. Our analysis reveals many novel properties of spot instance operation some of which offer predictive value while others do not. Using these insights, we design predictors for our features that offer a balance between computational efficiency (allowing for online resource procurement) and cost-efficacy. We explore "case studies" wherein we implement prototypes of dynamic spot instance procurement advised by our predictors for two types of workloads. Compared to the state-of-the-art, our approach achieves (i) comparable cost but much better performance (fewer bid failures) for a latency-sensitive in-memory Memcached cache, and (ii) an additional 18% cost-savings with comparable (if not better than) performance for a delay-tolerant batch workload.

[1]  Muli Ben-Yehuda,et al.  Deconstructing Amazon EC2 Spot Instance Pricing , 2011, 2011 IEEE Third International Conference on Cloud Computing Technology and Science.

[2]  David E. Irwin,et al.  Cloud Spot Markets are Not Sustainable: The Case for Transient Guarantees , 2016, HotCloud.

[3]  Yang Song,et al.  Optimal bidding in spot instance market , 2012, 2012 Proceedings IEEE INFOCOM.

[4]  Volodymyr Turchenko,et al.  Applications of neural-based spot market prediction for cloud computing , 2013, 2013 IEEE 7th International Conference on Intelligent Data Acquisition and Advanced Computing Systems (IDAACS).

[5]  Ethan Katz-Bassett,et al.  SPANStore: cost-effective geo-replicated storage spanning multiple cloud services , 2013, SOSP.

[6]  Guillaume Pierre,et al.  Wikipedia workload analysis for decentralized hosting , 2009, Comput. Networks.

[7]  Miao Pan,et al.  Optimal Resource Rental Planning for Elastic Applications in Cloud Market , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium.

[8]  Yonggang Wen,et al.  On the Cost–QoE Tradeoff for Cloud-Based Video Streaming Under Amazon EC2's Pricing Models , 2014, IEEE Transactions on Circuits and Systems for Video Technology.

[9]  Ming Mao,et al.  A Performance Study on the VM Startup Time in the Cloud , 2012, 2012 IEEE Fifth International Conference on Cloud Computing.

[10]  Rajkumar Buyya,et al.  Statistical Modeling of Spot Instance Prices in Public Cloud Environments , 2011, 2011 Fourth IEEE International Conference on Utility and Cloud Computing.

[11]  Ohad Shamir,et al.  On-demand, Spot, or Both: Dynamic Resource Allocation for Executing Batch Jobs in the Cloud , 2014, ICAC.

[12]  Bingsheng He,et al.  Monetary cost optimizations for MPI-based HPC applications on Amazon clouds: checkpoints and replicated execution , 2015, SC15: International Conference for High Performance Computing, Networking, Storage and Analysis.

[13]  References , 1971 .

[14]  Prateek Sharma,et al.  SpotCheck: designing a derivative IaaS cloud on the spot market , 2015, EuroSys.

[15]  Nandini Mukherjee,et al.  Application-Centric Resource Provisioning for Amazon EC2 Spot Instances , 2012, Euro-Par.

[16]  Yang Song,et al.  Optimal Bids for Spot VMs in a Cloud for Deadline Constrained Jobs , 2012, 2012 IEEE Fifth International Conference on Cloud Computing.

[17]  Rajkumar Buyya,et al.  Managing Peak Loads by Leasing Cloud Infrastructure Services from a Spot Market , 2010, 2010 IEEE 12th International Conference on High Performance Computing and Communications (HPCC).

[18]  George Kesidis,et al.  Neutrality in Future Public Clouds: Implications and Challenges , 2016, HotCloud.

[19]  J. A. Hartigan,et al.  A k-means clustering algorithm , 1979 .

[20]  Qianlin Liang,et al.  An Empirical Analysis of Amazon EC2 Spot Instance Features Affecting Cost-Effective Resource Procurement , 2018, ACM Trans. Model. Perform. Evaluation Comput. Syst..

[21]  Prateek Sharma,et al.  How Not to Bid the Cloud , 2016, HotCloud.

[22]  George Kesidis,et al.  Effective Capacity Modulation as an Explicit Control Knob for Public Cloud Profitability , 2016, 2016 IEEE International Conference on Autonomic Computing (ICAC).

[23]  Christopher Stewart,et al.  Blending on-demand and spot instances to lower costs for in-memory storage , 2016, IEEE INFOCOM 2016 - The 35th Annual IEEE International Conference on Computer Communications.

[24]  Martin Schulz,et al.  Exploiting redundancy for cost-effective, time-constrained execution of HPC applications on amazon EC2 , 2014, HPDC '14.

[25]  Prateek Sharma,et al.  SpotOn: a batch computing service for the spot market , 2015, SoCC.

[26]  Shaojie Tang,et al.  Towards Optimal Bidding Strategy for Amazon EC2 Cloud Spot Instance , 2012, 2012 IEEE Fifth International Conference on Cloud Computing.