Previous | Table of Contents | Next |
Because it is almost impossible for any one network management product to be aware of every vendor-specific set of managed objects located under the private/enterprise node, the use of these path identifiers gives network managers and administrators a way to walk through the global naming tree structure, so they can set or retrieve information from any vendor subtree. This is a capability network managers might not otherwise obtain by relying on the standard features of a network management system.
The SNMP standard defines a database of network management information as a MIB. The MIB consists of a combination of hardware and software settings that represent objects needed to manage different types of products.
To simplify network management, objects were organized into units known as groups. Groups have a common management function. Depending on the operating characteristics of a device, it may or may not support one or more of the groups shown under the MIB-2node in Exhibit 1. (The original MIB for managing a TCP/IP network is known as MIB-1; its new version is labeled MIB-2. MIB-2is recognized by the IANA under the management node in the global naming tree.)
Object Groups The system group permits configuration information to be defined to include what the device is, where it is located, and the person or persons to call when something goes wrong. Because of the importance of the system group, it is required to be supported by every device.
In comparison, other groups are optional and are only required to be implemented if applicable to a specific device. Examples of optional groups include:
One group that requires some explanation is the transmission group. This group more correctly represents a node position in the global naming tree under which groups applicable to different transmission technologies are placed.
Three examples of transmission technologies are shown in Exhibit 1 under the transmission node DOT3, DOT5, and FDDI. The DOT3 and DOT5 nodes reference local area networks standardized by the IEEE as 802.3and 802.5. Those standards are better known as Ethernet and Token Ring.
Each object in a device to be managed is represented by a unique address within the global naming tree. That address, which is referred to as an object identifier in standards documents, can be expressed in several ways.
The most commonly used method to express an object identifier is through the use of a string of integers separated by dots to form a path to the object. For example, the path to the system group shown in Exhibit 1 would be 1.3.6.1.2.1.1.The first object in that group would be located at 1.3.6.1.2.1.1.1 in the global naming tree.
Some object identifiers can have more than one value. For example, a bridge or router would have at least two interfaces, which would make it necessary to append a digit to the identifier path to denote the specific interface the administrator wishes to retrieve information from.
However, many objects represent a one-of-a-kind value, such as the location of a device. To provide consistency, an index is always added at the end of an identifier. Thus, if the object is a one-of-a-kind object, the administrator would add a zero (0) to its path. Because the first object in the system group is a one-of-a-kind object, its path identifier becomes 1.3.6.1.2.1.1.1.0.The omission of the trailing zero is a common error when users of a network management system use path addresses to retrieve object values.
Other methods used for object identifiers can include linking text labels with underscores or combinations of text labels and numerics. Because programming operations are easier and faster when working with numerics rather than text identifiers, most network management systems that allow users to enter tree identifiers do so by supporting integer strings with dots used as separators.
In actuality, there are two types of addresses network administrators need to assign to manage a device. The first address is the IP address of a device installed in a network, which defines its location.
The second address is the set of addresses within the global naming tree that defines the location of counters, registers, and memory locations that the administrator may be able to read from, write into, or read and write, depending on the access method defined for the object.
By having a standard method for defining objects in the global naming tree, different vendors can independently develop managed objects, avoiding the potential for the occurrence of addressing conflicts. In addition, vendors can design their products to support applicable SNMP and RMON groups in a standardized manner, and other vendors can develop network management systems without having to know full details of different vendor products.
In short, if a network management system supports the global naming tree structure, any user who understand tree addressing concepts is able to read, write, or read and write to and from different managed devices manufactured by different vendors.
For purposes of further illustrating the use of the global naming tree in network management operations, this section uses the example of an application developed using SimpleView from Triticom (Eden Prairie, MN).
Previous | Table of Contents | Next |