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


Here are some of the attributes of a well-performing host-based system:

  RDBMS and screen-handling processes are not bound by disk, memory, or network or terminal-handling components
  Good throughput and response times are maintained
  Effective backup of the system is possible within a specified time interval

Of course, all installations are unique; you may have different criteria for what is important.


NOTE:  If the screen-handling processes are too intense for one system to handle, you may be able to move the screen handling off the database system by separating the application into two parts: one for the RDBMS functions and one for the screen-handling functions. By separating the functions, you can then offload the screen-handling function to a client system. Remember that a client is defined in this book as a system that handles presentation services. A client can be a multiuser machine that serves terminals.

Terminal-Based Checklist

As with client/server computing, it is important with terminal or host-based computing to look at the efficiency of the entire system. Any single component can drag down the system performance. Look at each component as an individual contributor to the entire system. By minimizing resources used in the application code, the effect of the reduction is multiplied by the number of users simultaneously using this code.

Here is another checklist of items that may indicate whether you have a well-tuned system:

  Do you have good response times? If not, is the problem in the application code, in the database, or in the operating system?
  Can the system handle the required number of users effectively or do response times drastically increase with the numbers of users?
  Is the user community satisfied with system performance?
  Are time-critical tasks such as backup and loading accomplished within the required time interval?
  Is the system disk bound?

Later chapters examine how to determine whether some of these things are problems and how to fix them if they are.

Batch Computing

Batch computing refers to large jobs that take a significant amount of time and that typically aren’t accessed in an interactive manner. Response times can be minutes or even hours. Depending on the application, there may be large amounts or very small amounts of data returned. Usually, batch jobs consist of two phases: the batch loading and the batch processing. In both phases, processing time should be optimized.

What determines a well-tuned batch-processing system? The system should be optimized for load time as well as for batch processing. If the system is dedicated to batch processing, OLTP queries may see significant increases in response times.

The system should load and process data in the required time periods. The system should be CPU bound. The system should be able to transfer large amounts of data effectively over the network or through removable storage as specified by the application.

If large amounts of data are to be moved across the network, you should ensure that the network does not become a bottleneck. If necessary, install faster network hardware or segment the network to have a dedicated connection between the batch system and the loading system. If this is not possible, schedule the activity at a time when it least affects other users on the network. You may also be able to break the data into pieces for transfer or employ alternative methods such as tape.

Because the job in batch-processing systems is to move large amounts of data in, process that data, and move the results out, backup and recovery may not be important considerations. But remember that if all the input data is saved, you can always rerun the job to achieve the same results.

A well-tuned system can effectively handle these bursts of activity. A major difference between an OLTP system and a batch system is that—usually—the OLTP system has a constant workload; the batch system has bursts of activity.

Batch-Processing Checklist

Batch-processing systems are different from OLTP systems in that they must be able to handle the loading of large volumes of data and process large amounts of data in bursts. Typically, there are only certain time windows in which this processing can take place. Concern for response time is replaced by concern for processing times that can take minutes or hours to complete.

In a batch-processing system, the checklist for a well-tuned system is much shorter:

  Are you able to process the data in the required time? This includes loading, processing, and offloading the data.
  Is the system disk bound?

This book examines how to determine whether some of these things are problems and how to fix them if they are.

Exceptions

I always find it interesting to look at some of the cases that are exceptions to the typical way most systems operate. The following sections describe some of the systems that handle things a little differently.


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.