"Uptime" and "Availability", two common IT terms

Uptime and availability, the two most often referred to terms in conveying the reliable of an IT service.  Service providers often boost their service uptime in 99.99% (four-9s), 99.999% (five-9s) or even higher percentages.  For some, these two terms seem to refer to the same things and thus are used interchangeably.  As a result, a service advertised with the five-9s (or higher) uptime gives the perception of a very reliable service for a perspective IT service customer.  However, not only these two terms are not synonymous; they are two different metrics in measuring the reliability of a service.

"Uptime" vs "Availability"

In general, “uptime” indicates an operational state of the service as defined by the service provider.  For instance, a hosting provider may define the server’s “uptime” as the server is responding to the provider’s monitoring system via ICMP. As such, the server would still considered as in “normal operational state” even when the web server application running on the server is out of system resource and not accepting any more web requests. This “uptime” does not account for whether the business applications hosted within the server are running nor does it indicate the services are accessible by the end-user.  As a result, it’s crucial to also look at the “availability” of the service.  In another word, “availability” indicates the service is in operation, serving its intended function, and accessible by its intended audience.  Although, the general definition of uptime and availability are presented here; these two terms may vary from one service provider to another.  Thus, the perspective IT service customer is encouraged to discuss and get clarification from the service provider; and put them into the SLA.

Compensation Matters

Aside from clear understanding of uptime and availability, it is equally important to look at how breaching of them and the resulting compensation model are specified in the SLA.  A service provider may advertise their service has a guaranteed five-9s uptime, a 4-hour RTO (recovery time objective), and compensate in service fee credit.  Looking at five-9s alone, one would assume the service provider is in breached of the SLA if the service is down for more than 5.26 minutes per year.  However, with the 4-hour RTO in the SLA, the service provider is not in breached of the SLA if the service is restored within four hours or less (for every occurrence of service outage).  The compensation, for breach of SLA, in the form of service fee credit is not a “compensation”; rather, it is a mean to extend the service contract – which the server provider failed to deliver. 

The Bottom Line

A discerning IT service customer should look for concise definition of uptime and availability, clear definition of service breach condition, and meaningful compensation model.  A meaningful compensation model should have some form of penalty charge-back in addition to the reimbursement of some service fee.