TPC photo
The TPC defines transaction processing and database benchmarks and delivers trusted results to the industry
    Document Search         Member Login    
     Home
About the TPC
Benchmarks
      Newsletter
      Join the TPC
      Downloads
      Technical Articles
      TPCTC



TPC Benchmark Status
September 2001

Overview
The TPC held a General Council meeting August 5, 2001 in Bellevue, Washington. The main focus of the work was on setting the framework for new standards, such as the work going on with TPC-OLTP, and refining existing benchmarks. Several of the current benchmarks, namely TPC-C and TPC-H/R, are complete and mature, needing only periodic summary reviews to maintain them.

TPC-C
On October 20, 2001 all TPC-C V3 results will be archived and removed from the web site. Only results from the current version V5 will be displayed after this date.

TPC-OLTP Working Group

The TPC-OLTP Working Group focused its efforts on developing the requirements for a Request For Proposal (RFP) that could form the basis for the TPC to solicit proposals for a new OLTP workload. Areas discussed were market relevance, technology drivers, metrics, scalability, availability, security, transaction complexity, backup/restore, database complexity, database size, extensibility, load variances, and user context. The group also worked on selection criteria and process.

The proposed schedule is as follows:
Aug/01 final requirements, draft RFP, selection criteria
Oct/01 final draft RFP, presentation to the General Council (GC), legal review of RFP
Dec/01 GC approval, RFP released
Apr/02 review of RFP responses, selection process

Companies interested in joining the TPC to participate in the OLTP Working Group should contact the TPC Administrator at info@tpc.org.


TPC-H/R
The TPC-H/R Maintenance subcommittee met in August to discuss clarifications to the existing benchmark. The general direction was discussed at the General Council meeting through an intent motion.

TPC-DS

The subcommittee focused during this meeting on refining the Data Maintenance Process (ETL) part of the TPC-DS specification. Significant process has been made in the following areas:

  • source schema (operational data)
  • target schema (data warehouse data)
  • SQL mapping between source and target schemas
  • dimensional data load
  • definition of a transformation framework

The source schema is complete. It includes web and catalog channels. Adjustments were made to the target schema to accommodate the implementation of slowly changing dimensions. As a basis for the ETL process SQL mappings between the source and target schema are complete.

The loading of dimensional data is a central task of TPC-DS. During the past meeting the subcommittee made many decision on how to load slowly changing dimensions and fact data.

The TPC-DS specification defines transformations in pseudo code or sample SQL and mandates benchmark sponsors to implement transformations in SQL.


TPC-W

The TPC-W subcommittee made a few clarifications to the specification, primarily related to data dependent routing and functional routing. In addition, overload runs and response time graphs have been removed from the benchmark.

The subcommittee also drafted a technical overview white paper to highlight the features and benefits of the TPC-W benchmark. An administrative white paper is also in the works.

Finally, the subcommittee looked at proposals for enhancements and extensions to the benchmark, including the integration of XML, click stream analysis, numerous database population changes, and multiple store fronts. The TPC-W subcommittee also met with the TPC-OLTP Working Group to review the goals of the respective benchmarks and ensure that there were no overlaps.

Public Relations Committee

The recruiting campaign to bring new members to the TPC is in full swing. Informatica joined the council while Brocade attended the meeting as a guest.

The General Council approved changes proposed regarding the display of results to make all the benchmark result lists consistent. As with TPC-C, clustered system results will now be separated from non-clustered systems for the TPC-H, R, and W benchmarks. Additionally, the Top Ten listswill also be enhanced so that results from substantially identical systems will occupy a single position. This change allows the ten positions on each list to be held by ten distinct systems. For more information, please click here.


All Benchmark Status Reports
 

Valid XHTML 1.0 Transitional Valid CSS!