Reputation as a service: A system for ranking service providers in cloud systems

Abstract

Today cloud service providers guarantee the quality of their services by defining a set of Service Level Agreements (SLAs) with their customers. SLAs binds the provider to a set of service level metrics typically related to service reliability, availability, performance, security, and billing. Generally, the SLA formally specifies the minimum expected service metrics that the provider is committed to supply and that the customer agrees to accede. A detailed description on SLA terms, levels, and the various legislations and conditions that accompany their specification is comprehensively presented in [34]. Unfortunately, SLAs typically lack any technical means of enforcement which leaves the customer's data and software processes under the total control of the cloud service provider. Any failure to meet the SLA terms and obligations will have disastrous effects on the cloud customer and provider. The effects range from losing reputation and client trust to legal compliance, and financial penalties that may lead to putting an end to the entire business. This fact will put pressure and responsibility on the customers when selecting a particular cloud service provider for running their business processes and storing data. The severity of this selection is further aggravated when we estimate the serious losses incurred when dealing with misbehaving cloud providers or the technical difficulties, financial losses, and service downtimes accompanying the process of switching between service providers. Terabytes of data migration tasks over expensive communication links, software reconfiguration and adaptation, and data leakage and privacy implications are some factors that render the migration process highly expensive.