Reliability, Redundancy, and Maintainability

Building highly available systems requires that system architects fully understand and pay attention to three key factors: reliability, redundancy, and maintainability.

Reliability. The greater the reliability of each component of a system, the better overall availability will be. In this context, reliability refers to the absence of hardware or software failures over long, continuous operational periods. However, designing and building more reliable systems drives up system costs, and too much hardware and software engineering and testing can delay product introductions to the point of diminishing returns, which means that other measures—redundancy and maintainability—must augment reliability.

Redundancy. Given that perfect reliability is impossible and changing conditions may require system maintenance during an application’s service lifetime, redundancy plays an essential role in providing high availability. This redundancy can be built in at many levels, from hardware components, such as dual power supplies, to entire server farms located in different geographic locations.