Server Management software is all-too often on the light side when it comes to integrating with fault management solutions, often sending just generic SNMP Traps and not including an enterprise MIB. Because of this lack of SNMP MIB support, many servers are less capable of sending critical information to a network or fault management platform that is standing by ready and willing to act on their desperate pleas for help. In this way, would-be attackers and unforeseen issues like disk utilization problems often go unseen. Interestingly enough, the largest maker of computer software, Microsoft, has just generic MIB support in their default SNMP agent, thereby rendering these agents fairly useless when it comes time to send some serious information to a management system. So this means that the majority of the computer systems in the world do not have adequate abilities to send unsolicited critical information.

What are SNMP Traps? These little nuggets of computer encoded instructions tell a management system about issues happening somewhere on the network. It could be a message about something happening on the local system, or something a neighbor heard about another system. Really, the sky is the limit and the software as well as the enterprise MIB gives the agent the ability to encode interesting information into traps so they then can be sent to interested managers.

Thankfully, there are several vendors out there that do in fact write interesting MIBs and provide good solid agents that can provide this useful information, either solicited by using an SNMP poller, or unsolicited, by using traps. Both IBM and DELL have agents which do a really good job at this. And there is also a freeware agent, called NetSNMP, which can be used on a variety of platforms, and has good, solid coverage. So all it takes is a little research, and one can find an affordable (or free) agent, that can be installed on their servers, and then these servers will be able to provide the critical information that is needed in order to adequately manage a heterogeneous server farm.