Previous | Table of Contents | Next |
The manager/agent model includes a number of interactive components. The OSI network management framework defines the roles of those components. ISO/IEC 7498-4 [1-2] defines the OSI network management framework. ISO/IEC 9595, or CCITT X.710, defines the Common Management Information Service, or CMIS [1-3]; ISO/IEC 9596-1, or CCITT X.711, defines the Common Management Information Protocol, or CMIP [1-4].
Mark Klerers paper The OSI Management Architecture: An Overview divides the OSI management environment into several models: organizational, informational, and functional [1-5]. Another resource, Yeminis The OSI Network Management Model [1-6], discusses the relationships between OSI management systems and agents.
The organizational model uses a management domain (see Figure 1-8a). The domain may contain one or more management systems, managed systems, and subdomains. The managed system may, in turn, contain one or more managed objects. Each object is a network resource that one of the management systems may monitor and/or control.
Figure 1-8a. An abstract organizational model of the management environment (©1988, IEEE)
From this example, we can make several observations. First, the managers network management application must be compatible with that of the agent. Since this example used an OSI-based network management scheme, the Common Management Information Protocol (CMIP) could provide compatible communication. Second, the other layers of the two computing architectures must also be compatible. If the Network layer of the console uses the ISO Connectionless Network Protocol (CLNP), defined by ISO 8473, the router must also understand that protocol. (For example, if the console used the CLNP network addressing scheme to insert the 20-octet CLNP address in its Network layer process, the routers Network layer process would have to recognize that address in order to respond.) Finally, the same physical path must connect the manager and agent, although that path doesnt have to be on the same LAN or even in the same country.
An information model associated with the organizational model defines the structure of the management information and the management information base (MIB). As you can see in Figure 1-8b, a tree structure groups objects sharing similar characteristics into classes. These objects are represented as an entry in the management information tree; each entry has defined attributes and values.
Figure 1-8b. Management information tree (©1988, IEEE)
The functional model defines five areas of network management used for specific purposes. The next section (Section 1.5.2) explores these five areas in detail.
Figure 1-8c demonstrates how the various elements work together. This model relates the system management application process (SMAP) to the management information base (MIB) and the seven layers of the network management system. It defines interfaces for system management (the system management interface, or SMI) and layer management (the layer management interface, or LMI). The layer management functions are specific to a particular OSI layer entity. Examples of these functions would include layer specific parameters, tests, or services which would reside in a layer management entity (LME). The model also specifies a protocol for manager/agent communication, known as the Common Management Information Protocol, or CMIP.
Figure 1-8c. Architectural model of OSI management (©1988, IEEE)
The OSI management environment includes five areas of network management, which are called the OSI specific management functional areas (SMFAs) (see Figure 1-9). These are fault management, accounting management, configuration management, performance management, and security management. ISO 7498-4 discusses these functional areas using academicand often incomprehensiblelanguage, so Ill explain how these functional areas relate to the management of real networks.
Figure 1-9. OSI network management functional areas
The standard says that fault management encompasses fault detection, isolation, and the correction of the abnormal operation of the OSI environment. The standard goes on to consider error logs, fault identification, and diagnostic testing.
In plain English, fault management means that you need to first identify, then repair, network faults. There are two ways to manage faults: reactively or proactively. A reactive manager waits for a problem and then troubleshoots it. A proactive manager examines the manager and agents to determine whether they are exceeding critical operational thresholds, such as network utilization. If excesses occur, the proactive administrator determines their source and reduces them accordingly.
The standard says that accounting management enables charges to be established for the use of resources in the OSI environment, and for cost to be identified for the use of those resources. Other considerations include informing the users of the costs and resources consumed, setting accounting limits, and incorporating tariff information into the overall accounting process.
In the real world, accounting means dealing with real people using real network resources with real operating expenses. Examples of these costs include disk usage and data archiving, telecommunication expenses for access to remote data, and charges for sending electronic mail messages. You can also use accounting management to determine whether network resource utilization is increasing because of growth, which might indicate the need for additions or rearrangements in the near future.
Previous | Table of Contents | Next |