Probing the scheduling algorithms in the cloud based on OpenStack

Abstract

Among modern cloud infrastructures, live migration of virtual machines offers many advantages like scalability and elasticity but also leads to risks in the meantime. Security issues of live migration have been studied and classified into three threats: control plane, data plane and migration module. Lots of work has focused on the latter two aspects. However, the security of control plane has yet to be analyzed. This paper starts by introducing three classes of control plane threats: load balancing, scheduling and transmission. We then elaborate how scheduling attack can subvert the VM scheduling algorithm via the proposed scheduling algorithm reverse approach (SARA). We evaluate the effects of SARA using datasets gathered from OpenStack. This work is a beneficial attempt to compromise the control plane of VM migration, which can be used as a self-test tool for cloud service providers to test the defences against network intruders.