Each network manager must decide which open source and proprietary solutions offer methods to use to gather and make sense of the performance data. This means a delicate balance of monitoring networks and recognizing that monitoring can add to network traffic. Here we will look into a few of the remote monitoring options.
Remote Monitoring (RMON). RMON was originally developed to manage multiple LAN segments and remote sites from a central location. RFC 1757 defines 10 RMON groups for the gathering of information on Ethernet and Token Ring. The objective was to give the managers data to monitor the entire network, not just the devices. RMON also makes sure that the traffic that SNMP adds to the links is kept to a minimum.
As an extension of the SNMP MIB, the RMON requests and responses are part of the added SNMP traffic. The advantage of using RMON is that it collects statistical information and reports groups of data instead of small poll responses.
RMON 2 goes beyond the restricted monitoring of Ethernet and Token Ring networks performed by its predecessor. By monitoring activity in layers 3 through 7 of the OSI model, RMON 2 offers significant benefits.
By decoding and monitoring application-level traffic, the RMON 2 probe records application traffic and statistics both to and from servers. This traffic can include file transfer, telnet, e-mail, Web access, etc. The second benefit of RMON 2 is the ability to decode IP addresses. An RMON 2 probe can record the true source of traffic originating outside of its local network segment. It can also record traffic originating in the local network and destined for remote sites.
RMON 2 considers any protocol above the network layer of the OSI model to be at the application level. This includes TCP and UDP.
Switch Monitoring (SMON). The IESG approved SMON as a proposed standard MIB in 1999 (RFC 2613). SMON lets SNMP-based network management stations monitor switch traffic remotely. It includes the ability to provide virtual LAN objects. The most obvious challenges to switch monitoring are:
- Switches operate differently than other remote monitoring probes.
- There are many diverse proprietary switch-monitoring solutions.
- Inbound and outbound activity must be monitored.
- Any monitoring requires a large amount of system resources.
The RFC was intended to overcome these challenges by providing a framework to create a standard for monitoring switches. This approach seems to be working as vendors join the standards bodies.
Reproduced from Global Knowledge White Paper: Network Performance Monitoring