5.22 Introduction to Performance Management

• Need for Performance Management

• Performance Management Architecture

• Configuring and Customizing Performance module

• Getting started with Performance Management

Need for Performance Management

The main task of Network management systems is to monitor and control the network infrastructure. As computer networks increase in size, heterogeneity and complexity, effective management of such networks becomes more important and much difficult. In the present business world, each and every hour counts. An hour of network downtime means an hour of lost business and hence lost opportunities. Pro-actively managing the Network's health and performance is indispensable for any mission critical business and is something which the administrators of large networks have to put in place.

The main challenges in this area include

• Identifying the data that needs to be collected

• Interpreting the collected data

• Disseminating the data and

• Presenting the data .... which helps in network Performance management.

Performance Management Architecture

Web NMS Performance module is made up of the following components:

Data Collection service (Protocol neutral Data collection)

• This provides multi protocol Data collection support. Data can be collected from TL1 devices, CORBA devices etc. Default data collection takes place with respect to SNMP devices.

• User can plug in his own Protocol provider to facilitate data collection for that protocol.

• Data collection process can be Customized to suit user requirements.

• Observers can be set to monitor data collection process and get notified .

Data Collection Objects (Modeling Polling units)

Data collection process has been well-studied and modeled using objects that define

• what data to collect (PollingObjects and PolledData) and

• where to store the collected data

Scheduler

This component takes care of scheduling

• Periodic data collection
  Data Collection takes place at specified time intervals. Based on the time interval given, the Scheduler schedules the Data Collection process.

• Periodic Report generation
  Based on the settings you specify i.e. which report should be generated when ( day and time) and how often, the time of report generation will be scheduled and at the appropriate day and hour, the report will be produced.

• Tables clean up
  You can specify the periodicity as to how often you want to delete the tables which hold collected data. If table clean up is not done then the number of tables will increase and soon database will be full.

Filter

The filter otherwise called as Poll Filter allows manipulation of PolledData objects before they are added to the database. The manipulations will be some kind of addition, modification or deletion of PolledData objects .

Decoder

The data collected for the device can be converted into any other format and stored in database. This process of conversion is called Decoding and is taken care of by Data decoder. Normally the collected data is of type Long or String. One of the decoding practice can be that the collected string data which is in the format 3,88,7,9 (denoting memory usage ) might be parsed and separated as individual long values like 3 88 7 and 9.

Threshold Rules

The collected data needs some kind of monitoring which is done by applying Threshold rules on the collected data. These rules are nothing but Threshold objects each having a value, type, severity etc associated with it. These threshold objects will be associated with PolledData objects that define what data to collect and thus monitor constantly whenever data gets collected. Any violation in Threshold rule will result in notifications sent to the administrator.
 

Notification

Notification refers to messages sent to the network users for intimating some aspect of network happening. By default Web NMS Performance module supports notifications via three means:

• Threshold Notification
  Whenever collected data exceeds Threshold value, a Threshold event is generated and sent to Fault module which handles it.

• Collected data Observers (Poll observers)
  You can set Observers to get intimated when data is collected. You might like to do so when you want to do something with data before it gets stored in database.

• Poll Unit Observers
  You may wish to receive notifications whenever any change is made in existing Data collection configuration. Poll units are objects which hold definition of what data to collect and from where.

Reporting

The collected data can be grouped into meaningful sets and represented in formatted manner called Reports. You can create your own reports and add it to existing set of reports. Reports can be scheduled to be generated periodically. Many types of reports are provided including " On demand " reports.

Security and Audit

Authorization privileges are available using which the administrator can create user accounts and associate Performance related operations permitted for him. User Based Views can also be created and coupled with permissions on what the user can do on Performance objects.

Programming Interface

This refers to a rich set of API methods which help you in customizing, extending and configuring Performance module to suit your requirements. Javadocs for API methods are also available along with extensive Help documents.

Configuring and Customizing Performance module

You may have your own set of requirements for which you would like to customize and configure Web NMS Performance module. Following gives you the list which you can configure and customize:

1. Data collection

2. Data storage

3. Reports

4. Threshold generation

5. Graphs

6. Distributed Polling

7. Performance Client

8. Managed Object inputs to Performance module

You can do the above listed using API methods and using Configuration files.

Configuration files

• Configuration files are available in XML format

• You can modify configuration files before Server startup and see the changes

• These are stored under <Web NMS Home>/conf directory

• These are updated when settings are changed via Client User Interface

API methods

• These are used when you want to configure Performance objects at runtime

• You are required to get the handle of the API to use it's methods

• PollAPI is the most importantly used interface to configure Data collection parameters

• PollAPI can be accessed through RMI. When RMI is enabled by running the RMI registry, it will be published with the RMI handle / PollAPI on the server

The following code snippet is used for obtaining the handle for PerformanceAPI.

Getting started with Performance Management

To help you cater to your requirements this help manual is organized into meaningful sections as explained below:

1. Understanding Performance Objects
   This page illustrates the basics of how Performance module works and what are the objects that deal with
   Performance management. Also it gives you information on the core APIs used.

2. Performance Startup options
   This page lists and explains the different parameters that can be set before Performance server is startup. Also this document explains dependency among those parameters and their default values.

3. Database Schema
   This page gives you a brief note on the list of tables created and maintained by Performance module. This information will be useful to you when you customize Performance module.

4. Configuring Data collection
   This section illustrates the process of setting up the environment for data collection. You can change various attributes related to data collection and effect the changes. The main objects that deal with data collection are PollingObjects and PolledData.

5. Receiving Notification
   This section explains the various notifications sent by Poll engine and how to handle them.

6. Defining Threshold
   This section details you with information on threshold types, how they are created and configured.

7. Configuring Data storage
   This section illustrates the concept of collected data being stored in database and various ways to customize it.

8. Customizing Graphs
   The default mechanism of graph display and how you can customize it has been explained in this section.

9. Customizing Reports
   You can develop your own reports and display the required data in your own format. This page illustrates how to customize the reports.

10. Managing Extended PolledData
    This section explains the purpose and procedure to create your own class extending PolledData class. This facilitates you to add your own properties for data collection.

11. Examples
    This section gives you unique samples of Performance features like Data decoder, Poll Filter, Poll observer etc. The examples can be used to learn more about Web NMS Performance module.

12. Developer Tips
    This page can be used for Quick reference for Frequently Asked Questions and Trouble shooting Tips pertaining to Performance features.