Framework of evidence collection with temporal logic and first-order logic for providing accountability in Cloud Service

Abstract

With the introduction of cloud computing, many types of services have been introduced within the umbrella of this technology. With these services, some of the computations are brought into network, cloud of service machines. Although the technology gives lots of benefits and flexibility to its users, there are some areas that need to be taken to concern. The lack of mechanism to verify the policies are in place in the real system is one of the main reasons for difficulty of cloud computing adoption. There is a need of way to control the processes within the service chain and make sure that they are complied with service level agreement.

A framework for collecting evidence based on source of information about system’s running is proposed. The process is based on the obligations or policies defined for services in the system. This framework of evidence collection can be used as basis for providing accountability in cloud.

Accountability policies are, in the context of this framework, expressed in A-PPL which is an accountability policy representation framework. A use case is selected to demonstrate how the approach for evidence collection works. It is about health care service in the cloud. Test environment to represent healthcare service in cloud is set up. The purpose of this is to have data as source of evidence to be processed using proposed method on selected use case and related defined policies or obligations. The environment is set up using VMs (Virtual Machine) in Linux.

Two approaches on processing source of evidence and policy are shown and compared. The first approach processes them as MFOTL using MonPoly. The second approach processes them as Prolog (FOL) using Pyke.

Testing on those two approaches using the implementations done on this thesis shows that representing accountability policies in MFOTL gives more expressiveness than representing them in pure Prolog (FOL). However processing of MFOTL used in MonPoly gives no more flexibility in terms of practical usage and improvement than using Prolog with Pyke.