ABSTRACT
This chapter presents lean application capacity management in which capacity decision processes continuously monitor application and resource usage along with other characteristics, and automatically transition online application capacity into service ahead of rising user demand and retire application capacity behind falling demand. The goal is just enough online capacity available just in time to continuously serve offered user workload with acceptable service quality while holding sufficient reserve capacity to cover inevitable failure events and unforecast demand surges without wasting resources on excessive capacity. The chapter also presents common application demand management strategies, as well as rapid elasticity and scalability, that are used to address cyclical and random variations in user demand. Highly available services are engineered to have no single point of failure (SPOF). This means that sufficient redundant capacity is held online (e.g., active) or nearline so that user service can be recovered following a failure without producing unacceptable service impact or outage for users.
