Availability and reliability analysis of cloud computing under economic denial of sustainability (EDoS) attack: a semi-Markov approach

Abstract

Economic denial of sustainability (EDoS) attack is a new type of distributed denial of service (DDoS) attack which targets the economic resources of cloud adopters by exploiting the auto-scaling features of the cloud. EDoS attack has become a significant threat to cloud adopters as it can lead to bankruptcy or withdrawal from cloud services and this in turn entails unavailability of the cloud services. In fact, availability has been one of the most significant issues of cloud computing securities. To address this issue caused by an EDoS attack, we develop a novel state-based availability model of the cloud under the EDoS attack with the help of a semi-Markov process. Based on the proposed model, we determine the steady state availability of the cloud under the EDoS attack. In addition, we derive the mean time to failure (or reliability) of the cloud in the presence of the EDoS attack to characterize the period of time for which the cloud adopter experiences the state of being bankrupt or service unsustainable situation. Moreover, we present a cost management strategy to prevent the cloud adopter from undergoing bankruptcy or experiencing unexpected exorbitant bill by incorporating the cloud shutdown feature into the proposed model. Numerical illustrations are presented which give insights into how availability and reliability of the cloud services are affected by the EDoS attack. The proposed model is also validated via simulation evidences.