Visible to the public Double-Blind Reputation vs. Intelligent Fake VIP Attacks in Cloud-Assisted Interactions

TitleDouble-Blind Reputation vs. Intelligent Fake VIP Attacks in Cloud-Assisted Interactions
Publication TypeConference Paper
Year of Publication2018
AuthorsKonorski, J.
Conference Name2018 17th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/ 12th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDataSE)
ISBN Number978-1-5386-4388-4
Keywordsbest-response fake VIP attack strategy, Chained Attacks, client-server interactions, client-server systems, cloud computing, cloud-assisted interactions, demand undue QoS level, demanded QoS level, digital signatures, double-blind reputation scheme, Fake VIP attack, game theory, Games, handwriting recognition, intelligent fake VIP attacks, Markov processes, probabilistic fake VIP attack strategy, pubcrawl, QoS, QoS as a service paradigm, quality of service, Relays, reputation, resilience, Resiliency, Scalability, Servers, signature detection cost, Stackelberg game, Stackelberg game setting, Trust, trust strategy, two-dimensional Markov chains

We consider a generic model of Client-Server interactions in the presence of Sender and Relay, conceptual agents acting on behalf of Client and Server, respectively, and modeling cloud service providers in the envisaged "QoS as a Service paradigm". Client generates objects which Sender tags with demanded QoS level, whereas Relay assigns the QoS level to be provided at Server. To verify an object's right to a QoS level, Relay detects its signature that neither Client nor Sender can modify. Since signature detection is costly, Relay tends to occasionally skip it and trust an object; this prompts Sender to occasionally launch a Fake VIP attack, i.e., demand undue QoS level. In a Stackelberg game setting, Relay employs a trust strategy in the form of a double-blind reputation scheme so as to minimize the signature detection cost and undue QoS provision, anticipating a best-response Fake VIP attack strategy on the part of Sender. We ask whether the double-blind reputation scheme, previously proved resilient to a probabilistic Fake VIP attack strategy, is equally resilient to more intelligent Sender behavior. Two intelligent attack strategies are proposed and analyzed using two-dimensional Markov chains.

Citation Keykonorski_double-blind_2018