[ Contents ] [ Index ]

How SNMP Works

Simple Network Management Protocol (SNMP) is one of the most widely used protocols for network management. It does what its name says: it manages network devices in a relatively simple manner. SNMP is a simple, unacknowledged, connectionless protocol.

SNMP was originally developed to work under the umbrella of the Internet suite of protocols, primarily in TCP/IP (Transmission Control Protocol/Internet Protocol) transport protocols. Since then, SNMP has been implemented to work over other common transport protocols, such as Novell’s IPX/SPX (Internet Packet Exchange/Sequenced Packet Exchange).

If you have an enterprise management console on your network (for example, HP OpenView or IBM NetView), you can configure SANscape servers to send SNMP traps about events to the console. See Setting Up Servers to Send Traps for details.

SNMP Trap Messages

Agents use SNMP to send information to enterprise management systems.

Once a server is configured to send traps, SANscape event information flows to that system.

SANscape’s trap messages contain the following information:

  • OID (Object Identifier) 1.3.6.1.4.1.2294.1.2

  • Event Date (MM, DD,YY, such as 01,22,98)

  • Event Time (HH:MM:SS, such as 15:07:23)

  • Server Address and Name (IP Address Name, such as 192.187.249.187 Administration)

  • Card Name

  • Event Severity (Informational, Warning, Critical)

  • The text message

The information that is displayed and its format are dependent on the SNMP management console you are using.

The file RST_OID.MIB must be loaded into the enterprise management console for traps to be received. On a Solaris, Linux, and HP-UX OS, this file is located in /opt/dothill/ssagent. On an IBM AIX OS, this file is located in /usr/dothill/ssagent. On Microsoft Windows, this file is located in \Program Files\Dot HIll\ssagent.

Agents and Managers

The SNMP software model consists of agents and managers. An agent is the software that enables variables in the managed device to be monitored. The agent software is either embedded in the device that is managed or running on a computer to which the device is connected.

The manager or management console is the network monitoring software that uses SNMP to send requests to the agents about the variables. The manager polls the agents periodically because the values of variables change. SNMP managers typically analyze the data received from the agents and make it available to the user in a graphical display. The data can be stored in a database for historical and trend analysis.

In addition to responding to the SNMP manager’s requests, agents can send unsolicited notifications or traps to the management station. This is similar to interrupts in conventional software.

Management consoles can take various actions in response to a trap. Usually, traps are logged, and the information in them is forwarded through a notification scheme to the interested user. Traps can also cause the SNMP manager to initiate further action, such as querying the agent with more requests to find out the current status of the network entity or launching an application.

Management Information Base (MIB)

On a network, each device instrumented by SNMP agents has one or more variables or objects associated with it. An example of a typical variable might be one that keeps track of the name of the device. These variables or objects are described in a text file called Management Information Base (MIB). A MIB is a precise textual definition of variables being managed. The procedures for defining a MIB have been established by the Internet Engineering Task Force (IETF) of the International Standards Organization (ISO).

SNMP Objects

Some variables for network management are common to all LAN network devices. Many of these variables have been standardized and are referred to as MIB-II variables and MIB-II extensions. The IETF has also defined the procedures for MIBs that are not in a standard MIB-II category. These variables are listed in private enterprise MIBs.

Variables are listed under the entry of the private enterprise MIB. For example, the object ID 1.3.6.1.4.1.2294.1.2 refers to an object ID (or OID) of a variable for keeping track of the agent version under one of SANscape’s MIBs.

Every private enterprise wanting to develop agents to recognize its own devices can apply to the IETF for its own tree node.

SNMP Request Types

The SNMP operations that support accessing MIB data are Get, GetNext, and Set. Get retrieves a scalar value from a table entry field value, given an index value. GetNext is used to traverse the instances of table entries, with each command returning the next sequential value in the column (or field) of interest. The Set operation is used to create or change the value of a MIB variable.

SNMP Security

Setting values raises the issue of security. SNMP V1 (version 1) has a simple security scheme. Each protocol data unit (PDU) contains a community string, which is like a combination of user name and password. Each agent can be configured for a particular community string. An agent does not respond to a request sent to it except when the community string in the request PDU matches the community string in its own configuration.

[ Contents ] [ Index ]