Q-Logic MKII-BASE16 Interrupt-Based Data Collection, Trap-Directed Polling Data Collection

Managing the Switch Using SNMP

Switch

SNMP

Agent

How Much Traffic?

12.5% at 4:15 PM

Mgmt

Station

Figure 3-14 Polling-Only Data Collection

Interrupt-Based Data Collection

This method provides immediate notification to the network management station if an extraordinary event occurs (assuming the node has not crashed and the communication path between the node and the management station remains intact). A drawback of this method involves the resources required to generate the error (or trap). If the error message contains a lot of information, the managed node expends valuable time and resources generating the error rather than performing its primary function.

Also, if several identical events occur back-to-back, excessive network bandwidth is tied up with repeated information. It would be particularly unfortunate if the errors pertained to network congestion. To avoid this, thresholds can be established for the managed node regarding when to report problems. However, threshold monitoring also requires agent resources to determine the appropriateness of generating a trap.

Figure 3-15 illustrates how a managed node agent interrupts its operation to report errors to the SNMP management station.

Switch

SNMP Agent

ALERT! 55% at 4:20 PM

Mgmt Station

Figure 3-15 Interrupt-Based Data Collection

Trap-Directed Polling Data Collection

This method is a combination of the polling-only and interrupt-based data collection processes and is a highly effective and efficient means of performing network management. The management station polls the agents in the managed nodes to collect data and then displays the data on the console in either numeric

GigWorks MkII-16 Switch Model MKII-BASE16		Switch Management 3-53
Installer's/User's Manual	59003-01 Rev. A