GoAlfred Wiki

Wiki of Alfred Crombez

Edit

Introduction

A network management system provides the means to configure, control, and diagnose several entities in the network. Several management functions can be considered:
  • Configuration and registration of the network and its entities.
  • Control giving access to the controlling functions of the network entities such as start-up, and shutdown.
  • Fault diagnostics providing assistance in diagnosing problems.
  • Notification providing alerting capabilities and allowing the definition of alarm rule expressions.
  • Performance statistics are required for performance analysis and parameter tuning. An historian can collect and store this data over selected periods of time, allowing a more long-term analysis.

Edit

Management Architecture

The management architecture is based on the entity model. In this model, each to manage object in the network is seen as an entity. Each entity has an agent for handling the management functions. The management director is a software system that manages the entities.
Image

Management Architecture



Edit

Management Information

Following management information is exchanged between the management system and the entities:
  • Directives are commands issued by management to an entity: a show command to read a value, a set, add or remove command to change a value. Directives can also specify controlling actions: enable or disable entity, start entity.
  • Events flow from the agent to the management module. An event is generated when some specific normal or abnormal condition occurs that is of interest to the management system.

Edit

Management Director Architecture

The management director software is used to control and monitor a collection of modules. The architecture has a modular design to simplify the addition of extensions:
  • Kernel. Provides a set of services that support and integrate other functions of the director.
  • Management Information Repository. Database of management information of the entities being managed.
  • Application Programming Interfaces. Set of APIs that define how the other three types of director components can be plugged into the kernel. They allow these components to be implemented independent of the particular operating system on which the kernel runs.
  • Presentation Modules. A presentation module handles a particular style of user interface between a manager and the director. Presentation modules are independent of the entities being managed and the functions that can be applied to them.
  • Function Modules. A function module handles a set of specific management actions that can be applied to a collection of entities. Function modules are independent of the entities being managed and of the user interface style a manager employs.
  • Access Modules. An access module handles the communication with one or more of the entities being managed. Access modules are independent of the functions that can be applied to the entities and of the user interface style a manager employs. An access module operates as a sink that receives information about events the managed entities generate.
Image

Director Architecture


Edit

Entity Attributes

The variables of an entity that can be inspected or set by a management action are called entity attributes. There are four types of attributes:
  • Identification. Uniquely identifies an entity instance to management.
  • Characteristics. Characteristics allow to control the operating parameters of an entity (for example, polling rate, routing cost). In general, they take default values when the entity is created, and can only be changed through a network management action.
  • Status. Status attributes allow to inspect the current state of an entity. Unlike characteristics, status attributes can be changed without management intervention.
  • Counters. A counter indicates the number of times an operation has been performed by an entity or the number of times a particular condition has been detected. As with status attributes, counters change in value during normal operation. Two types of counters are distinguished:
    • Statistics are normal conditions that are counted for determining the performance and the activities of an entity. Examples of statistics are number of messages sent, number of message sent, time since last reset of statistics.
    • Errors are counters for abnormal conditions that are detected. A difference is made between operational errors (such as communication errors) and software errors (such as an unexpected software condition).

Edit

Basic Management Functions

The basic management functions are the minimum management functions required for configuring, controlling, and diagnosing the entities in the network.

Edit

Configuration

With the configuration function, entities are made known to the management system. Each to manage entity is registered by creating an entry for it in the management information repository.

Edit

Control

The control functions allow to change entity parameters and to initiate controlling actions of the different entities. Some control functions are common to most of the entities:
  • Setting parameters or resetting them to default values.
  • The start-up, shutdown or restart of an entity.
  • Enabling or disabling entities or certain of their functions.

Edit

Fault Diagnostics

The fault diagnostics need to provide the necessary means for diagnosing problems:
  • View tests allow viewing the internal operation and data of the different entities.
  • Communication tests allow an interactive exchange of messages with the different entities and to request diagnostic data locally stored at entities.
  • Loop-back tests allow testing a certain communication path by looping a test message at various points in the path.

Edit

Event Logging

The event logging allows logging of different events:
  • Normal events are generated for specific normal conditions that are of interest for monitoring the system's operation.
  • Abnormal events such as errors: for such events it must be possible to log additional data that reflects the current status of the entity that generates the event.

Besides logging the information, the event logger module allows to set filters on what events at what moments of which entities that need to be logged.

Edit

Performance Analysis

The performance analysis is based on the event counters of entities:
  • Statistics
  • Errors (operational and software errors)

The performance analysis allows enabling and disabling the event counting and counts the events into the Management Information Repository.

Edit

Extended Management Functions

Edit

Auto-Configuration

Auto-configuration consists of the automatic discovery and registration of the network entities.

Edit

Maintenance Operations

Maintenance operations are related to system maintenance:
  • Download of system software.
  • Upload of dump allows to send memory dumps to the management system.

Edit

Notification

Notification allows alarming exception conditions by providing following functions:
  • The definition of exception and alarm rules.
  • The detection and notification or alarming of "out of limit" conditions.

Edit

Historian

The historian collects and stores management data over several selected periods of time, allowing a more long-term analysis.

Edit

Exporter

The export modules allow to export real-time and historian data to databases.

Edit

Management Solutions

OpenNMS Network Management Platform

AdventNet Network Management Solutions

Edit

References

OSI Model (Wikipedia)

Network Management-An Overview (SEI)

Common Management Information Protocol (CMIP) (SEI)

Simple Network Management Protocol (SNMP) (SEI)

Common Information Model (CIM) (Wikipedia)

Common Information Model (CIM) (DMTF)

Web-Based Enterprise Management (WBEM) (DMTF)

Design of the DECmcc Management Director

ScrewTurn Wiki version 2.0.19. Some of the icons created by FamFamFam.