RSMOA: A revenue and social welfare maximizing online auction for dynamic cloud resource provisioning

We study online cloud resource auctions where users can arrive anytime and bid for heterogeneous types of virtual machines (VMs) assembled and provisioned on the fly. The proposed auction mechanism RSMOA, to the authors' knowledge, represents the first truthful online mechanism that timely responds to incoming users' demands and makes dynamic resource provisioning and allocation decisions, while guaranteeing efficiency in both the provider's revenue and system social welfare. RSMOA consists of two components: (1) an online mechanism that computes resource allocation and users' payments based on a global, non-decreasing pricing curve, and guarantees truthfulness; (2) a judiciously designed pricing curve, which is derived from a threat-based strategy and guarantees a competitive ratio O(ln(p)) in both system social welfare and the provider's revenue, as compared to the celebrated offline Vickrey-Clarke-Groves (VCG) auction. Here p is the ratio between the upper and lower bounds of users' marginal valuation of a type of resource. The efficacy of RSMOA is validated through extensive theoretical analysis and trace-driven simulation studies.

[1]  Andrew V. Goldberg,et al.  Competitive auctions and digital goods , 2001, SODA '01.

[2]  Tim Roughgarden,et al.  Algorithmic Game Theory , 2007 .

[3]  Noam Nisan,et al.  Truthful approximation mechanisms for restricted combinatorial auctions , 2008, Games Econ. Behav..

[4]  Ronald M. Harstad,et al.  Computationally Manageable Combinational Auctions , 1998 .

[5]  Yoav Shoham,et al.  Truth revelation in approximately efficient combinatorial auctions , 2002, EC '99.

[6]  Noam Nisan,et al.  Competitive analysis of incentive compatible on-line auctions , 2000, EC '00.

[7]  Baochun Li,et al.  Revenue maximization with dynamic auctions in IaaS cloud markets , 2013, 2013 IEEE/ACM 21st International Symposium on Quality of Service (IWQoS).

[8]  Zongpeng Li,et al.  Dynamic resource provisioning in cloud computing: A randomized auction approach , 2014, IEEE INFOCOM 2014 - IEEE Conference on Computer Communications.

[9]  Kui Ren,et al.  When cloud meets eBay: Towards effective pricing for cloud computing , 2012, 2012 Proceedings IEEE INFOCOM.

[10]  William Vickrey,et al.  Counterspeculation, Auctions, And Competitive Sealed Tenders , 1961 .

[11]  Joseph Naor,et al.  Online Primal-Dual Algorithms for Maximizing Ad-Auctions Revenue , 2007, ESA.

[12]  Chaitanya Swamy,et al.  Truthful and near-optimal mechanism design via linear programming , 2005, 46th Annual IEEE Symposium on Foundations of Computer Science (FOCS'05).

[13]  Athanasios V. Vasilakos,et al.  A Framework for Truthful Online Auctions in Cloud Computing with Heterogeneous User Demands , 2016, IEEE Transactions on Computers.