Improving the performance of online auction sites through closing time rescheduling

The workload of auction sites exhibits some interesting features as indicated in previous work by the authors. Of particular importance to this paper is the fact that over thirty percent of the bids of an auction arrive in the last five percent of the auction's lifetime. This creates a surge in the load seen by auction sites as an auction's closing time approaches. The site's performance is degraded if the site is not able to cope with these load spikes. With performance degradation comes frustrated customers and lost business. To mitigate this problem, we propose that auction sites reschedule auction closing times within a relatively small window centered around the closing time originally proposed by a customer. The scheduler's goal is to more evenly distribute the number of closings of auctions in order to produce a more uniform distribution of the number of bids on the auction site. This paper describes the auction closing time scheduling algorithm, applies it to a trace of auctions and bids obtained from an actual auction site, and uses a simulation model to study the effects of the rescheduled trace on the site's performance. In addition to that, the paper presents an algorithm to obtain a complete knowledge (CK) but nonrealizable schedule. This schedule could be used as a benchmark for comparison with the realizable rescheduled load.

[1]  David Leake,et al.  Managing, Mapping, and Manipulating Conceptual Knowledge* , 1999 .

[2]  Stefan Agamanolis,et al.  Synergistic storyscapes and constructionist cinematic sharing , 2000, IBM Syst. J..

[3]  I. Nonaka,et al.  The Concept of “Ba”: Building a Foundation for Knowledge Creation , 1998 .

[4]  Heinz Ulrich Hoppe,et al.  Integrating collaborative concept mapping tools with group memory and retrieval functions , 2002, CSCL.

[5]  Virgílio A. F. Almeida,et al.  In search of invariants for e-business workloads , 2000, EC '00.

[6]  Joseph D. Novak,et al.  Learning How to Learn , 1984 .

[7]  Virgílio A. F. Almeida,et al.  Business-oriented resource management policies for e-commerce servers , 2000, Perform. Evaluation.

[8]  Justine Cassell,et al.  Literacy learning by storytelling with a virtual peer , 2002, CSCL.

[9]  E. Thompson Colour Vision: A Study in Cognitive Science and Philosophy of Science , 1994 .

[10]  Brian R. Gaines,et al.  Collaboration through concept maps , 1995, CSCL.

[11]  Shoshana Blum-Kulka,et al.  “You gotta know how to tell a story”: Telling, tales, and tellers in American and Israeli narrative events at dinner , 1993, Language in Society.

[12]  Virgílio A. F. Almeida,et al.  A methodology for workload characterization of E-commerce sites , 1999, EC '99.

[13]  Heinz Ulrich Hoppe,et al.  Document Management in a Computer-Integrated Classroom , 2002, CRIWG.

[14]  D.A. Menasce,et al.  Towards workload characterization of auction sites , 2003, 2003 IEEE International Conference on Communications (Cat. No.03CH37441).

[15]  Joseph D. Novak,et al.  A theory of education , 1977 .

[16]  Justine Cassell,et al.  Virtual peers as partners in storytelling and literacy learning , 2003, J. Comput. Assist. Learn..

[17]  D. Boje The storytelling organization: A study of story performance in an office-supply firm. , 1991 .

[18]  E. Yalow Educational psychology: A cognitive view. 2nd ed. , 1978 .

[19]  Daniel A. Menascé,et al.  Scaling for E-Business: Technologies, Models, Performance, and Capacity Planning , 2000 .

[20]  Peter J. Denning,et al.  Operating Systems Theory , 1973 .

[21]  C. Amza,et al.  Specification and implementation of dynamic Web site benchmarks , 2002, 2002 IEEE International Workshop on Workload Characterization.