A4MtCS

A2MtCS: Accountability Analysis for Multi-tenant Cloud Services

We propose a deployment framework that regroups the competences of the evidence analysis approach, as shown in this Figure.

In the following, we briefly introduce the elements of the Evidence analysis deployment architecture:

• Violation base: This database stores the violations in the form of a recordset. These violations are detected by MobMtCS, where each of them specifies the details of each violation.

• Mt Evidence handler: It looks for the suitable AltMt Evds to analyze. The Mt Evidence handler follows the two-stage optimization mechanism (i.e., multi-tenancy aware classification and optimization actions) so as to obtain Alt Evidences. Therefore, violation details and AltMt Evds serve as inputs for the analysis process.

• Accountability-obligation models repository: It contains accountability-obligation models stored with reference to the evidence model. Obligation models represent the main input of the Obligation handler component as they specify the main attributes of each violated obligation according to which the optimal evidence collection is ensured.

• Obligation handler: It is a component looking for the accountability-obligation model corresponding to the accountability obligation violated (received as input). Then, it ensures the transfer of the accountability-obligation model to the component responsible for evidence optimization (i.e., Optimization solver).

• Optimization Solver: It follows the two-stage mechanism for optimal evidence collection. The optimization solver receives Mt Evds from the MobMtCS that are related to the service where the accountability violation has been detected. Then, it applies the linear program and executes an action, in order to generate AltMt Evds . In fact, we have implemented stage1 using IBM-ILOG Cplex Optimization Studio V12.6.3. stage2 has been implemented using Java language.

• Alt evidences base: It contains Alt Evidences obtained by the Mt Evidence handler. They are permanently stored until the end of the analysis process.

• Analyzer: It is the administrator of the Evidence analysis components and ensures the analysis process. In fact, it follows the instructions given by the analysis algorithm adopted by the service provider.