System availability refers to the percentage of time that a system is available to provide service. What can cause a system to become unavailable? Equipment failure may be one of the biggest causes, then the system remains unavailable until actions are taken to restore service.
There can be confusion between the terms “Availability” and “Grade of Service (GoS)”. Availability indicates that the system is up and running – whether the system has enough capacity to serve everyone that want to use the system at the same time, is another question. That capacity question is the system’s Grade of Service.
The prime factors that affect the System Availability include the the natural failure rate of components and the system design. If the system design includes redundancy the availability will be improved – but that will cost? A cost-benefit analysis should be considered to determine which components it makes sense to make redundant. If there are redundant components consideration must be given to how the system will switch to that back-up. If the system self monitors and automatically switches to the back-up then the unavailability will be minimal. If however activating the back-up requires human intervention (for example, installing a spare card) the downtime involved in dispatching a worker to perform the can be significant and must be considered in the unavailability calculation.
The system unavailability should also consider things such as the unavailability of microwave radio links (if used) due to environmental conditions and allow for system maintenance outages (if necessary).
Availability should be considered from a user’s perspective – the end user doesn’t care why a system is unavailable, only that it isn’t. Vendors may try to define availability that excludes scheduled downtime for maintenance purposes, but this should be considered unacceptable in any service level agreement.
What is a reasonable availability for a public safety system? This is something an agency will have to evaluate based on their operational requirements and budget. Availability is often expressed as a percentage, approaching 100%. Typical availability values and their corresponding annual outage times are:
- 99.900% = “three 9’s = 8.8 hours out of service per year
- 99.990% = “four 9’s = 53 minutes out of service per year
- 99.999% = “five 9’s = 5.3 minutes out of service per year
Availability will be predicted in the system design stage and then the availability of the active system should be monitored on a regular basis. If a trend away from the target availability is noticed, strategies for restoring will have to be addressed.