Evaluating the scalability of distributed systems

A system design is scalable if it can be economically deployed at a range of scales, in both small and large configurations. Little attention has been paid to measuring and comparing the scalability of different designs of software for distributed operation. Recently a new measure of scalability (called here P-scalability since it bused on the “power” metric) has been defined specifically for distributed systems. The paper generalizes the metric, defines a scaling path embodying a strategy modifying the system as it is scaled up, and employs scalability enabling parameters which adapt the system to give the maximum value of the scalability metric, at any point along the scaling path