Version 9.6	SEARCH   CONTENTS   |   PDF BOOKS   |   HOME   UP   PREV   NEXT

This document describes the accounting records for Broker that can be used for several purposes, including:

• application chargeback
  for apportioning EntireX resource consumption on the conversation and/or the application level;

• performance measurement
  for analyzing application throughput (bytes, messages, etc.) to determine overall performance;

• trend analysis
  for using data to determine periods of heavy and/or light resource and/or application usage.

This document covers the following topics:

• EntireX Accounting Data Fields

• Using Accounting under UNIX and Windows

• Example Uses of Accounting Data

EntireX Accounting Data Fields

In the EntireX Accounting record, there are various types of data available for consumption by applications that process the accounting data:

Note:
Accounting fields of any version greater than 1 are created only if the attribute ACCOUNTING-VERSION value is greater than or equal to the corresponding version. For example: accounting fields of version 2 are visible only if ACCOUNTING-VERSION=2 or higher is specified.

Using Accounting under UNIX and Windows

• Broker Attribute File Settings

• Retrieving Accounting Data

Broker Attribute File Settings

ACCOUNTING = NO | YES | (YES, SEPARATOR=Separator Characters) (Default is NO)

Set this parameter to "NO" (i.e., do not create accounting data) or "YES" to create accounting data. Up to seven separator characters can specified using the SEPARATOR suboption, for example ACCOUNTING = (YES, SEPARATOR=;). If no separator character is specified, the comma character will be used.

Retrieving Accounting Data

The accounting file will be located in the Broker's installed directory. The file's name is based on the ETB_LOG environment variable and the current date and time (for uniqueness). Example: If ETB_LOG is set to BROKER1.LOG, the accounting data file will be named BROKER1_YYYYMMDDHHMMSS.csv. If ETB_LOG is not set, the Broker's ID will be used, with an extension of CSV (e.g. ETB048_YYYYMMDDHHMMSS.csv). See Environment Variables in EntireX.

Example Uses of Accounting Data

• Chargeback

• Trend Analysis

• Tuning for Application Performance

Chargeback

Customers can use the EntireX accounting data to perform chargeback calculations for resource utilization in a data center. Suppose EntireX Broker is being used to dispatch messages for three business departments: Accounts Receivable, Accounts Payable, and Inventory. At the end of each month, the customer needs to determine how much of the operation and maintenance cost of EntireX Broker should be assigned to these departments. For a typical month, assume the following is true:

The use of Broker resources here is based upon both the amount of traffic sent to the Broker (bytes) as well as how often the Broker is called (messages). The average of the two percentages is used to internally bill the departments, so 52.5% of the cost of running EntireX Broker would be paid by the Inventory Department, 25% by the Accounts Receivable Department, and 22.5% by the Accounts Payable Department.

Trend Analysis

The Accounting Data can also be used for trend analysis. Suppose a customer has several point-of-sale systems in several stores throughout the United States that are tied into the corporate inventory database with EntireX. The stubs would be running at the stores, and the sales data would be transmitted to the Broker, which would hand it off to the appropriate departments in inventory. If these departments wish to ascertain when the stores are busiest, they can use the accounting data to monitor store transactions. Assume all of the stores are open every day from 9 AM to 10 PM.

The owner of the stores can examine the data and make decisions based upon the data here. For example, on weekdays, he or she can see that there is little business until lunchtime, when the number of transactions increase. It then decreases during lunch hour; then there is another increase from 5 PM to 8 PM, after people leave work. Based on this data, the owner might investigate changing the store hours on weekdays to 10 AM to 9 PM. On the weekend the trends are different, and the store hours could be adjusted as well, although there is a more regular customer flow each hour on the weekends.

Tuning for Application Performance

Assume that a customer has two applications that perform basic request/response messaging for similar sized messages. The applications consist of many Windows PC clients and Natural RPC Servers on UNIX. An analysis of the accounting data shows the following:

A further analysis of the accounting data reveals that there are a lot of non-zero response codes in the records pertaining to Application 2, and that a lot of these non-zero responses indicate timeouts. With that information, the customer can address the problem by modifying the server code, or by adjusting the timeout parameters for SERVER2 so that it can have more time to get a response from the Service.

www.softwareag.com

Copyright © 1997-2014 Software AG, Darmstadt, Germany.

SEARCH   CONTENTS   |   PDF BOOKS   |   HOME   UP   PREV   NEXT