Click Here!
home account info subscribe login search My ITKnowledge FAQ/help site map contact us


 
Brief Full
 Advanced
      Search
 Search Tips
To access the contents, click the chapter and section titles.

Oracle Performance Tuning and Optimization
(Publisher: Macmillan Computer Publishing)
Author(s): Edward Whalen
ISBN: 067230886x
Publication Date: 04/01/96

Bookmark It

Search this book:
 
Previous Table of Contents Next


The TPC-E benchmark consists of a combination of both OLTP and batch components, made up of the five TPC-C transactions, two additional OLTP transactions, and a concurrent batch activity. The new OLTP transactions (in addition to the five TPC-C transactions) are listed here:

  Customer-Inquiry. This transaction type represents a real-time customer service query that returns account information. It is frequently accessed and has stringent response time criteria.
  Customer-Status. This transaction type represents a real-time corporate-wide Management Information Systems (MIS) query that extracts customer payment information. This transaction is a heavy read-only transaction executed at a low frequency and has relaxed response time criteria.

The third new transaction is the Customer-Demographics transaction, which represents a real-time corporate-wide MIS query that evaluates customer buying activity based on demographic factors. This query is characterized by a heavy read-only transaction that is infrequently executed but consumes significant resources. The response time requirement is relaxed for this query.

The first metric used in the TPC-E benchmark is the Maximum Qualified Throughput (MQTh)—the number of New-Order transactions per minute, which is reported as tpmE. The and price-per-tps (price/performance) metric indicates the total system cost divided by the MQTh. This second metric is designed to demonstrate the value of the system. Along with these primary metrics, there are several secondary metrics such as CPU busy, system I/O rates, and so on.

As with the other OLTP benchmarks, the database size scales with throughput. This requirement provides for a fair comparison between systems by requiring larger systems to access larger databases.

The TPC-E benchmark will probably see less activity than the other TPC benchmarks because of the tremendous cost and effort required to execute it. In addition, I think that some companies will use the TPC-E benchmark while others migrate to the TPC-C/S or TPC-Server benchmark, depending on their market focus.

TPC-C/S

The TPC-C/S (Client/Server) benchmark is an OLTP benchmark designed to reflect the performance and price characteristics of the client/server solution to OLTP problems.


NOTE:  At the time this book goes to press, the TPC-C/S benchmark is still under development; it has not yet been accepted as an official benchmark (although I believe it will be accepted soon).

The TPC-C/S benchmark was developed to test system configurations designed with a client/server solution in mind. A client/server solution refers to the division of work between the client (front-end) and server (back-end) machines. The TPC-C/S benchmark is characterized by the following elements:

  Use of modern Graphical User Interfaces (GUIs) to improve user friendliness and end-user productivity
  Use of application development tools to speed development time and promote ease of maintenance
  Use of decision support and office automation tools
  Significant network activity
  Access to data when the server/host is not available
  A move closer to the new, decentralized business model
  Provisions for the user to access the large number of applications available for work-stations and personal computers
  Exploitation of the perceived price/performance advantage of the client/server solution
  Use of online and deferred execution modes
  Use of multiple online sessions
  Significant disk input/output
  Significant client processing of multiple data types: numeric, text, graphics, and images
  Transaction integrity (ACID properties)
  Parallel processing by the client and server

As with the TPC-A, TPC-C, and TPC-E benchmarks, the TPC-C/S benchmark actually simulates user access. Unlike the other benchmarks, the TPC-C/S benchmark simulates that access through a workstation GUI. This arrangement provides a multiuser, full-system emulation. The TPC-C/S benchmark requires input/output screen formatting through this interface.

The performance metric reported by the TPC-C/S benchmark is a “business throughput” measured by the number of New-Order transactions per minute. The metric is the tpmCS (transaction per minute C/S). As with all the other TPC benchmarks, the associated price/performance must also be disclosed.

With the rapid emergence of client/server computing, the TPC-C/S benchmark may soon overtake the other TPC benchmarks in popularity. The TPC-C/S benchmark—as with most other TPC benchmarks—is very time consuming and expensive to run.


Previous Table of Contents Next


Products |  Contact Us |  About Us |  Privacy  |  Ad Info  |  Home

Use of this site is subject to certain Terms & Conditions, Copyright © 1996-2000 EarthWeb Inc.
All rights reserved. Reproduction whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Read EarthWeb's privacy statement.