SNMP Tutorial
SNMP Tutorial
Contents
1. Overview ................................................................................................................................... 2
3.1. Agents................................................................................................................................. 4
3.2. Manager ............................................................................................................................. 4
3.3. The MIB.............................................................................................................................. 4
3.4. MIB Objects........................................................................................................................ 4
8. Traps ......................................................................................................................................... 9
9. Community Strings................................................................................................................. 10
1. Overview
SNMP (Simple Network Management Protocol) was first defined by the Internet Engineering
Task Force (IETF) in 1989. Since then, SNMP has become an industry standard for
controlling networking devices from a single management application. For information on the
SNMP standard, refer to RFC 1098.
SNMP is a set of network management protocols and functions that communicate using the
Internet Protocol (IP) stack. SNMP allows network managers to isolate and troubleshoot faults
on multi-vendor networks, configure devices on a network, and monitor network performance
and status.
As an Application Layer protocol in the seven-layer OSI Model, SNMP normally uses UDP
(User Datagram Protocol) and defines a method of communication. SNMP consists of two
parts:
• Agents and Proxy Agents – These reside on network devices and generate
information such as Ethernet addresses, TCP/IP addresses and traffic statistics about
the device on which they reside. The information is then stored in Management
Information Bases (MIBs). Proxy agents act on behalf of a device that has not
implemented SNMP.
SNMP is used to communicate information between network management stations and the
agents in the network elements.
3. SNMP Manager and Agents
In accordance with the SNMP model, managed devices such as routers, hubs, bridges and
workstations contain software components called agents. The agent monitors the operation of
the managed device by maintaining a collection of variables, called objects, in the
Management Information Base (MIB). The MIB reflects the operation of the managed device.
Here follows a more detailed description of some SNMP concepts.
3.1. Agents
Agents are any devices on the network that need to be managed and that have the
SNMP protocol and the Management Information Base.
Agents monitor the desired objects in their environment, package this information in
the appropriate manner, and send it to the management station either immediately or
upon request. Information is generated by the Agent, stored in its MIB, and made
available to the Manager. Proxy Agents act on behalf of a device that has not
implemented SNMP.
3.2. Manager
A manager program, which normally executes on a network server, exchanges
messages with the agent to access the agent's MIB. The manager reads from, and
writes to, objects in the MIB according to predefined access privileges that have been
assigned to the MIB objects.
SNMP defines the protocols and message formats used to perform the read and write
operations; these are called gets and sets, respectively.
The MIB is the definition of the data, or objects, that are stored in the agent for the
manager to access.
• Interface Information
• IP datagrams
• UDP datagrams
Most networking devices that support SNMP support MIB II. As MIB II is a published
set of data definitions, any SNMP Manager can access MIB II data. V vendors have
also created their own sets of definitions, called custom MIBs, so that their own
Managers can gather more product-specific information than is available from MIB II.
4. Structure of the MIB
The standard MIB's structure is represented by a tree. There are three components of the
tree:
• Root - The unnamed root of the tree contains a set of characters common to all MIB
objects located beneath that root. Objects beneath "unnamed" are said to be in that
root's domain.
• Subtree - The subtree contains a subset of the information available in the root; a
subtree may also serve as a root and have subtrees of its own.
• Leaf - The leaf is a subtree with no additional subtrees in its domain; a leaf
represents a single MIB object whose characteristics are unique from any other MIB
object.
Each MIB object can be located by following a path from "unnamed," through the subtrees, to
the leaf. In order to simplify finding an individual MIB object, the paths are defined by a
sequence of numbers.
5. Remote Network Monitoring
RMON is a standard for monitoring and reporting network activity using remote monitors.
RMON is designed to supplement the management information obtained and used by SNMP.
It provides functions for getting information about the operation and performance of entire
networks or of subnetworks in an inter-network.
Remote monitors are expected to do their work in a way that is minimally disruptive to
network activity and that makes minimal demands on the available resources. Much of the
information that remote monitors provide is summary information, some of which can be
obtained passively (by counting packets, error signals, and so on).
As a supplement to the SNMP management function and to the data in MIB II, RMON is
included in the global tree under MIB II. In the notation used to describe elements in the tree,
RMON is MIB II 16. RMON provides MIB elements of its own. The table below lists these
elements.
SUBTREE DESCRIPTION
The SNMP PDUs are five commands, or operations, implemented by SNMP. They are used
between the Agent and the Manager to pass information and make requests. They are:
• LinkDown - attached interface has changed from the up to the down state.
The strings provide a measure of security for information contained in the objects, although
they are not passwords. The most commonly used community strings are public and private.
The receiving entity first determines if the SNMP command has a valid community string, then
the access to the requested objects is verified as either read-only or read-write.
When an SNMP command is received, its community string is compared to the community
string associated with the requested object to determine the appropriate access level.