On Characterized Once-in-a-Lifetime Workload patterns for Periodic Static Scaling

Abstract

Measurability is a concept in periodic static scaling based on the following two conditions: (a) a cloud service provider should be cautious, that is, should not exclude any cloud consumer's resource pooling pattern strategy from consideration; and (b) a cloud service provider should consider the cloud consumers' resource pooling pattern preferences, that is, should deem a cloud consumer's resource pooling pattern strategy k i infinitely more likely than k' i if it premises the cloud consumer to prefer k i to k' i . A resource pooling pattern strategy is measurable if it can optimally be chosen under common resource pooling pattern conjecture in the events (a) and (b). In this paper we present an algorithm that for every finite periodic static scaling operation computes the set of all measurable resource pooling pattern strategies. The algorithm is based on the new idea of an Once-in-aLifetime Workload preference limitation, which is a pair (k i , V i ) consisting of a resource pooling pattern strategy k i , and a subset of resource pooling pattern strategies V i , for cloud service provider i. The interpretation is that cloud service provider i prefers some resource pooling pattern strategy in V i to k i . The algorithm proceeds by successively adding Once-ina-Lifetime Workload preference limitations to the periodic static scaling.