City Research Online

Abstract

In this report, I introduce a hybrid approach for certifying security properties of cloud services that combines monitoring and testing data. This report argues about the need for hybrid certification and examines the basic characteristics of hybrid certification models.

The certification of cloud service security has become a necessity due to the on-going concerns about cloud security and the need to increase cloud trustworthiness through rigorous assessments of security by trusted third parties. Unlike the certification of security in traditional software systems, which is based on static forms of security assessment (e.g., the Common Criteria model), the certification of cloud service security requires continuous assessment. This is because cloud services are provisioned through dynamic infrastructures operating under security controls and other configurations that may change dynamically introducing unforeseen vulnerabilities. Cloud service security can also be compromised because of attacks on co-tenant services.

Recent work on cloud service certification applies dynamic forms of security assessment, notably dynamic testing or continuous monitoring. These overcome some of the limitations of traditional security certification and audits (e.g. they produce machine readable certificates incorporating dynamically collected evidence). However, there are cases where existing approaches cannot provide an adequate level of assurance. Testing, for instance, may be insufficient for transactional services, as it is normally performed through a special testing (as opposed to the operational) service interface. Monitoring-based certification may also be insufficient if there is conflicting or inconclusive evidence in monitoring data; such data may, for example, not cover all traces of system events that should be seen to assess a property.

To overcome such problems, I am working on a hybrid approach for certifying cloud service security that can combine both monitoring and testing evidence. For that reason, I designed a new cloud certification approach supporting the automated and continuous certification of security properties of cloud services based on the combination of dynamically acquired testing and monitoring evidence that can deliver the high level of assurance and can overcome the limitations of assessments based on each of these types of evidence in isolation. My approach is based on the cloud certification framework of the CUMULUS EU FP7 project.

Publication Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments:	School of Science & Technology > Computer Science

Export

Downloads