Overview

Let's face it, electronic equipment will fail video servers are no exception. No matter how much redundancy or fault tolerance is built into a system, at some time, an element in the system will fault and that system will come down. Ideally, systems engineers would like to achieve what is referred to as "five-9s" uptime (99.999 percent on-line uptime), a very high level of reliability and protection. To a broadcast facility operating 24x7, this kind of performance amounts to slightly more than 315 seconds per year of downtime, or in sales/traffic lingo "the loss of no more than ten 30-second spots in a year."

Information technology looks at such goals for data-intensive servers, and it is by no means an easy or inexpensive goal to achieve. In the world of compute-intensive and transaction servers, the costs in additional hardware and software systems to get from 99.9 percent to 99.999 percent (an increase of two 9s) may not be something a facility can afford for what amounts to 22/100 of a second difference in video-equivalent performance.

There are elements, systems, and configurations that can result in better overall performance, which in turn can substantially change the comfort level when considering a professional video server solution. Looking out for that single point of failure and minimizing its potential to bring some or all of the server system down is the topic of this month's installment.

Predicting and Preparing

Today's system administrator is tasked with predicting and preparing for the failures of power...