Cloud Computing Spot Pricing Dynamics: Latency and Limits to Arbitrage

This study examines cloud computing spot pricing dynamics and the influence of latency on those pricing dynamics. Using the Amazon Elastic Compute Cloud U.S. East and West market spot instance pricing and latency intraday data from April 9, 2010, to May 22, 2011, we find considerable time variation in spot instance prices, and prices are often persistently higher in the West. Bivariate vector autoregressive model results show that within-market autoregressive pricing effects are larger than across-market effects. We also document that over 70% of the relative price discovery occurs in the East market. Our regression results further show that East–West latency differentials have a significantly positive effect on East–West pricing differentials. Latency creates a dynamic pricing wedge that widens or narrows conditional on the latency differentials. Using an error correction model, the speed of adjustment from long-run pricing convergence errors causes the short-run price differential to narrow, but the adjustment does not completely offset the price differential.

[1]  R. Baillie,et al.  Price discovery and common factor models , 2002 .

[2]  C. Granger,et al.  Co-integration and error correction: representation, estimation and testing , 1987 .

[3]  Mohammad S. Rahman,et al.  Battle of the Retail Channels: How Product Selection and Geography Drive Cross-Channel Competition , 2009, Manag. Sci..

[4]  James Davidson,et al.  Cointegration and error correction , 2013 .

[5]  Ram Ramesh,et al.  Efficient Risk Hedging by Dynamic Forward Pricing: A Study in Cloud Computing , 2013, INFORMS J. Comput..

[6]  Michele Mazzucco.  Towards Autonomic Service Provisioning Systems , 2010, 2010 10th IEEE/ACM International Conference on Cluster, Cloud and Grid Computing.

[7]  Bruce N. Lehmann,et al.  Some desiderata for the measurement of price discovery across markets , 2002 .

[8]  Zhongju Zhang,et al.  Research Note - Online Price Dispersion: A Game-Theoretic Perspective and Empirical Evidence , 2012, Inf. Syst. Res..

[9]  Jan Stallaert,et al.  A Clock-and-Offer Auction Market for Grid Resources When Bidders Face Stochastic Computational Needs , 2011, INFORMS J. Comput..

[10]  Anindya Datta,et al.  World Wide Wait: A Study of Internet Scalability and Cache-Based Approaches to Alleviate It , 2003, Manag. Sci..

[11]  Quanyan Zhu,et al.  Dynamic Resource Allocation for Spot Markets in Cloud Computing Environments , 2011, 2011 Fourth IEEE International Conference on Utility and Cloud Computing.

[12]  Robert B. Miller,et al.  Response time in man-computer conversational transactions , 1899, AFIPS Fall Joint Computing Conference.

[13]  Michele Mazzucco,et al.  Achieving Performance and Availability Guarantees with Spot Instances , 2011, 2011 IEEE International Conference on High Performance Computing and Communications.

[14]  Ramayya Krishnan,et al.  Pricing and Resource Allocation in Caching Services with Multiple Levels of Quality of Service , 2005, Manag. Sci..

[15]  Subhajyoti Bandyopadhyay,et al.  Cloud computing - The business perspective , 2011, Decis. Support Syst..

[16]  James D. Hamilton.  Time Series Analysis , 1994 .

[17]  Ram Ramesh,et al.  Risk Management and Optimal Pricing in Online Storage Grids , 2008, Inf. Syst. Res..

[18]  Erik Brynjolfsson,et al.  Frictionless Commerce? A Comparison of Internet and Conventional Retailers , 2000 .

[19]  Joel Hasbrouck,et al.  One Security, Many Markets: Determining the Contributions to Price Discovery , 1995 .

[20]  Jan Stallaert,et al.  A Market Design for Grid Computing , 2008, INFORMS J. Comput..

[21]  Jesus Gonzalo,et al.  Estimation of Common Long-Memory Components in Cointegrated Systems , 1995 .

[22]  Frank de Jong,et al.  Measures of Contributions to Price Discovery: A Comparison , 2002 .