At 09:16 AM 8/19/2002 -0600, Vernon Schryver wrote: >The simplest, most direct, and most accurate tactic seems to me to be >snooping on the routing traffic that the box sends. Since the box was >described as a "Router," why not set up a BGP session to it or snoop on >its RIP, IGRP, or OSPF traffic. I think you misunderstood him, or at least I understood him differently. He said >I need to design a heartbeat mechanism between our Router and a 3rd party >device, that supports IP stack. > >The intent is to develop software that monitor the general health of the >box by receiving/ and or sending regular keepalives. Cisco equipment has a similar function. It was originally developed to monitor printers and such in IBM network, and has been extended to be used for a variety of purposes. Basically, the question before the house is "is a given device alive, and if not can someone please be notified?" By extension, one might note the loss rate to the device (number of pings sent for which no corresponding reply was received), the mean round trip delay, or other data. Cisco's approach is to ping the device - every device that implements IP is supposed to implement ICMP, and if ICMP is implemented, ICMP Echo Request/Reply should be implemented. One could go on to wonder about SNMP or a list of other protocols, if there is specific information to be gained (do you want to know, for example, what kind of device it is?). However, it seems that ping covers the most cases if only general heartbeat is required.