Why you should run TPC-DS: a workload analysis

  • Authors:
  • Meikel Poess;Raghunath Othayoth Nambiar;David Walrath

  • Affiliations:
  • Oracle Corporation, Redwood Shores, CA;Hewlett-Packard Company, Houston, TX;Sybase Inc, Concord, MA

  • Venue:
  • VLDB '07 Proceedings of the 33rd international conference on Very large data bases
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Transaction Processing Performance Council (TPC) is completing development of TPC-DS, a new generation industry standard decision support benchmark. The TPC-DS benchmark, first introduced in the "The Making of TPC-DS" [9] paper at the 32nd International Conference on Very Large Data Bases (VLDB), has now entered the TPC's "Formal Review" phase for new benchmarks; companies and researchers alike can now download the draft benchmark specification and tools for evaluation. The first paper [9] gave an overview of the TPC-DS data model, workload model, and execution rules. This paper details the characteristics of different phases of the workload, namely: database load, query workload and data maintenance; and also their impact to the benchmark's performance metric. As with prior TPC benchmarks, this workload will be widely used by vendors to demonstrate their capabilities to support complex decision support systems, by customers as a key factor in purchasing servers and software, and by the database community for research and development of optimization techniques.