Cloud flat rates enabled via fair multi-resource consumption

Abstract

Many companies rent Virtual Machines (VM) from cloud providers to meet their computational needs. While this option is also available to end-users, they do not always take advantage of this option. One reason may be that it is common to pay on a per-VM-basis, whereas the telecommunications sector has shown that customers prefer flat rates. A flat rate for cloud services needs to define utilization thresholds, to cap the usage of heavy customers and thereby limit their impact on the flat rate price and the cloud performance. Unfortunately, customers consume multiple heterogenous resources in clouds, e.g., CPU, RAM, disk I/O and space, or network access. This makes the definition of a customer’s fair “cloud share” and according utilization thresholds complex. Backed by a questionnaire among more than 600 individuals, this paper designs the new Greediness Metric (GM) that formalizes an intuitive understanding of multi-resource fairness without access to consumers’ utility functions. This GM enables the introduction of attractive cloud flat rates and fair sharing policies for private/commodity clouds and provides incentive to customers to wisely determine VM configurations.