ABSTRACT
Cloud spot markets enable users to bid for compute resources, such that the cloud platform may revoke them if the market price rises too high. Due to their increased risk, revocable resources in the spot market are often significantly cheaper (by as much as 10×) than the equivalent non-revocable on-demand resources. One way to mitigate spot market risk is to use various fault-tolerance mechanisms, such as checkpointing or replication, to limit the work lost on revocation. However, the additional performance overhead and cost for a particular fault-tolerance mechanism is a complex function of both an application's resource usage and the magnitude and volatility of spot market prices.
We present the design of a batch computing service for the spot market, called SpotOn, that automatically selects a spot market and fault-tolerance mechanism to mitigate the impact of spot revocations without requiring application modification. SpotOn's goal is to execute jobs with the performance of on-demand resources, but at a cost near that of the spot market. We implement and evaluate SpotOn in simulation and using a prototype on Amazon's EC2 that packages jobs in Linux Containers. Our simulation results using a job trace from a Google cluster indicate that SpotOn lowers costs by 91.9% compared to using on-demand resources with little impact on performance.
- Scientific Computing Using Spot Instances. http://aws.amazon.com/ec2/spot-and-science/, June 2013.Google Scholar
- ClusterK. https://clusterk.com/, July 10th 2015.Google Scholar
- Google preemptible instances. https://cloud.google.com/compute/docs/instances/preemptible, July 10th 2015.Google Scholar
- J. Barr. New - EC2 Spot Instance Termination Notices. https://aws.amazon.com/blogs/aws/new-ec2-spot-instance-termination-notices/, January 6th 2015.Google Scholar
- C. Clark, K. Fraser, S. Hand, J. Hansen, E. Jul, C. Limpach, I. Pratt, and A. Waeldr. Live Migration of Virtual Machines. In NSDI, May 2005. Google ScholarDigital Library
- B. Cully, G. Lefebvre, D. Meyer, M. Feeley, N. Hutchinson, and A. Warfield. Remus: High Availability via Asynchronous Virtual Machine Replication. In NSDI, April 2008. Google ScholarDigital Library
- S. Di, D. Kondo, and F. Cappello. Characterizing and Modeling Cloud Applications/Jobs on a Google Data Center. Supercomputing, 69(1), July 2014. Google ScholarDigital Library
- M. R. Hines, U. Deshpande, and K. Gopalan. Post-copy Live Migration of Virtual Machines. SIGOPS Operating Systems Review, 43(3), July 2009. Google ScholarDigital Library
- S. Khatua and N. Mukherjee. Application-centric Resource Provisioning for Amazon EC2 Spot Instances. In EuroPar, August 2013. Google ScholarDigital Library
- M. Mao and M. Humphrey. A Performance Study on VM Startup Time in the Cloud. In CLOUD, June 2012. Google ScholarDigital Library
- M. Mattess, C. Vecchiola, and R. Buyya. Managing Peak Loads by Leasing Cloud Infrastructure Services from a Spot Market. In HPCC, September 2010. Google ScholarDigital Library
- M. Mazzucco and M. Dumas. Achieving Performance and Availability Guarantees with Spot Instances. In HPCC, September 2011. Google ScholarDigital Library
- I. Menache, O. Shamir, and N. Jain. On-demand, Spot, or Both: Dynamic Resource Allocation for Executing Batch Jobs in the Cloud. In ICAC, 2014.Google Scholar
- A. Mu'alem and D. Feitelson. Utilization, Predictability, Workloads, and User Runtime Estimates in Scheduling the IBM SP2 with Backfilling. TPDS, 12(6), June 2001. Google ScholarDigital Library
- C. Reiss, J. Wilkes, and J. L. Hellerstein. Google Cluster-usage Traces: Format + Schema. Technical report, Google Inc., November 2011.Google Scholar
- P. Sharma, S. Lee, T. Guo, D. Irwin, and P. Shenoy. SpotCheck: Designing a Derivative IaaS Cloud on the Spot Market. In EuroSys, April 2015. Google ScholarDigital Library
- R. Singh, D. Irwin, P. Shenoy, and K. Ramakrishnan. Yank: Enabling Green Data Centers to Pull the Plug. In NSDI, April 2013. Google ScholarDigital Library
- Y. Song, M. Zafer, and K. Lee. Optimal Bidding in Spot Instance Market. In Infocom, March 2012.Google ScholarCross Ref
- S. Tang, J. Yuan, and X. Li. Towards Optimal Bidding Strategy for Amazon EC2 Cloud Spot Instance. In CLOUD, June 2012. Google ScholarDigital Library
- D. Tsafrir, Y. Etsion, and D. Feitelson. Modeling User Runtime Estimates. In JSSP, 2005. Google ScholarDigital Library
- W. Voorsluys and R. Buyya. Reliable Provisioning of Spot Instances for Compute-Intensive Applications. In AINA, 2012. Google ScholarDigital Library
- S. Yi, D. Kondo, and A. Andrzejak. Reducing Costs of Spot Instances via Checkpointing in the Amazon Elastic Compute Cloud. In CLOUD, July 2010. Google ScholarDigital Library
- M. Zafer, Y. Song, and K. Lee. Optimal Bids for Spot VMs in a Cloud for Deadline Constrained Jobs. In CLOUD, 2012. Google ScholarDigital Library
- S. Zaman and D. Grosu. Efficient Bidding for Virtual Machine Instances in Clouds. In CLOUD, July 2011. Google ScholarDigital Library
- Q. Zhang, E. Gürses, R. Boutaba, and J. Xiao. Dynamic Resource Allocation for Spot Markets in Clouds. In HotICE, March 2011. Google ScholarDigital Library
Index Terms
- SpotOn: a batch computing service for the spot market
Recommendations
The Aneka platform and QoS-driven resource provisioning for elastic applications on hybrid Clouds
Cloud computing alters the way traditional software systems are built and run by introducing a utility-based model for delivering IT infrastructure, platforms, applications, and services. The consolidation of this new paradigm in both enterprises and ...
A Survey on Spot Pricing in Cloud Computing
Amazon offers spot instances to cloud customers using an auction-like mechanism. These instances are dynamically priced and offered at a lower price with less guarantee of availability. Observing the popularity of Amazon spot instances among the cloud ...
Price efficiency in futures and spot trading: The role of information technology
During the last years information technology has had a profound impact on financial markets. The speed of trading and the amount of available information has increased substantially. Nearly all exchanges have upgraded their trading systems to meet the ...
Comments