Ok the basic idea here is this:

hosts.table   -- Holds all _general_ information for a host, keyed on the
                 hostname, fully qualified.
h_services    -- Will be a list of services a given host performs (also
                 for dertimineing impact if host fails), Again this is
                 keyed on the hostname. Each service will be assigned a 
                 3 letter 'type' that will be looked up in another table
                 to see 'how' to check for this service. This will reily on
                 some sort of standard plug-in method or something for
                 adding services and what function to invoke.
services      -- A list of services, the type, the tuple of the service 
                 what function will check this service, a description of
                 the service, whos responsible for said service etc..
circuts       -- A list of circuts that we know about, the vendor for 
                 the circut. Id of the circut, location of circut,
                 trouble # for the circut etc..
vendors       -- A list of vendors we know about, contact info etc.. This
                 way if trouble appears on a node and its connected to 
                 a circut we can generate a reasonable report so that
                 we won't have to go digging around looking for who to call
                 what circut #'s and such. It should be all info we need
                 to report and resovle a trouble.

