An Availability Model for Distributed Computing Environments

One significant quality of service measure for a computing environment is availability. Traditional measures of availability of the central processing unit and operating system. A common method used to calculate this value is to have the operating system write time-stamped records onto an output device. These are then analyzed to determine the fraction of time that the records were written and this number is given as the availability of the computing environment. However, there are several significant problems with this method and similar approaches to availability calculations when used for modern time sharing environments: