Trustworthy Cloud Certification: A Model-Based Approach

Abstract

Cloud computing is introducing an architectural paradigm shift that involves a large part of the IT industry. The flexibility in allocating and releasing resources at runtime creates new business opportunities for service providers and their customers. However, despite its advantages, cloud computing is still not showing its full potential. Lack of mechanisms to formally assess the behavior of the cloud and its services/processes, in fact, negatively affects the trust relation between providers and potential customers, limiting customer movement to the cloud. Recently, cloud certification has been proposed as a means to support trustworthy services by providing formal evidence of service behavior to customers. One of the main limitations of existing approaches is the uncertainty introduced by the cloud on the validity and correctness of existing certificates. In this paper, we present a trustworthy cloud certification approach based on model verification. Our approach checks certificate validity at runtime, by continuously verifying the correctness of the service model at the basis of certification activities against real and synthetic service execution traces.

Modeling time, probability, and configuration constraints for continuous cloud service certification

“…To ll in this gap, we extend our model-based certication process in Anisetti et al (2014aAnisetti et al ( , 2015 with a methodology for the verication of the correctness of the system model. We assume the system model to be either manually gener-ated by the system provider with the help of the accredited lab or automatically generated according to existing solutions (e.g., Anisetti et al (2014b); Merten et al (2012); Ernst et al (1999)). Our certication process implements the pre-deployment and production evaluation steps, and extends the system model adding time, probability, conguration constraints, and attack paths tested on the ToC.…”

“…More recently, the works in Anisetti et al (2014aAnisetti et al ( , 2015Anisetti et al ( , 2014b; Ardagna et al (2016) have provided the basic building blocks for the denition of a certication process addressing cloud peculiarities. In this paper, we consider advanced aspects aimed to guarantee the trustworthiness of the certication process in a dynamic cloud, including constraints on time relations between dierent states of the model, transition probabilities, and congurations, and attack paths for certicate validity.…”

“…We note that, although our methodology is applied to a specic certication scheme, it is general and can be used for any model-based assurance technique. This paper develops on our previous work (Anisetti et al (2014b(Anisetti et al ( , 2015) by providing i) a trustworthy certication scheme extended with checks on model correctness (Section 2); ii) a methodology composed of four processes and corresponding algorithms for verifying the model structure, as well as time, probability, and conguration constraints (Section 3); iii) an architecture including components supporting online verication of model correctness (Section 4); iv) an extensive evaluation of our trustworthy certication scheme (Section 5). The evaluation considers the eHealth application of ATOS SA, a major IT enterprise, deployed on top of open source IaaS OpenStack.…”

Certification-Based Cloud Adaptation

“…A failure returned by the verification process is triggered by i) an inconsistency between the certification model and the execution traces, or ii) a violation of configuration constraints. Scenario i) requires a system adaptation and has been already tackled in our previous work in [13]. Scenario ii) is the target of the adaptation approach in this paper.…”