Author:
George Kousiouris
Affiliation:
National Technical University of Athens, Greece
Keyword(s):
Cloud Computing, elasticity, DoS,eDoS, Markov Chains, DoS Identification as a Service
Related
Ontology
Subjects/Areas/Topics:
Cloud Computing
;
Cloud Computing Enabling Technology
;
Cloud Economics
;
Cloud Risk, Challenges, and Governance
;
Fundamentals
;
Monitoring of Services, Quality of Service, Service Level Agreements
;
Platforms and Applications
;
Security, Privacy, and Compliance Management
;
Service Monitoring and Control
;
Services Science
;
Services Security and Reliability
Abstract:
The problem of DoS attacks has significant effects for any computing system available through the public domain. In the case of Clouds, it becomes even more critical since elasticity policies tied with metrics like Key Performance Indicators (KPIs) can lead a Cloud adopter to significant monetary losses. DoS attacks increase the KPIs, which in turn trigger the elastic increase of resources but without financial benefit for the owner of the cloud-enabled application (Economic Denial of Sustainability). Given the numerous scenarios of DoS attacks and the nature of services computing (in many cases involving legitimate automated traffic requests and bursts), networking mitigation approaches may not be sufficient. In this paper, the concept of Key Completion Indicators (KCIs) is provided and an analysis framework based on a probabilistic approach is proposed that can be applied on the application layer in cloud-deployed applications and elasticity policies, in order to avoid the aforemen
tioned situation. KCIs indicate the level of completeness and provided revenue of the requests made towards a publicly available service and together with the KPIs can lead to a safer result with regard to elasticity. An initial architecture of this DoS Identification as a Service is portrayed.
(More)