From my office up to 200 miles from the field sites, I was able to monitor, troubleshoot, and diagnose that system. Service technicians were able to review a location from their office and then ensure they had the appropriate tools and repair parts in their vehicle before driving over 100 miles to the reporting site. These sites were located in northeast Wyoming and reaching some of these sites could be a challenge at times.
Monitoring in levels is important. The ability to drill down through a summary alarm to the point where you are at the lowest replaceable unit ensures you will have that component available when dispatched to that site.
The result though isn't truly quantifiable as failure rates of various components cannot be predicted unless the system is developed to the level that trends can be established and predictive algorithms refined I had the opportunity a few years ago to work with three other engineers and mathematicians to develop a system to continuously test the state of a railway switch layout and predict failure conditions. As a condition approached a known failure point, the system would send an alarm allowing maintenance personnel to be dispatched prior to the failure. Over time we were able to optimize the test points; adding some and deleting others until we had only those that provided effective information.
As you might guess, I am still a believer in monitoring and feel that five nines of availability is achievable; but is it worth it? Bring in the accountants and calculate the cost per minute of that system being unavailable. In my experience, that cost will be higher than you might imagine.
Sr Systems Consultant