A new agent based security framework for collaborative cloud environment

With the increasing adoption of cloud technologies, the concern regarding the security of the deployed environment is gaining importance. Vulnerabilities are often added to the virtual machines (VM) without the knowledge of the user due to the dynamics of the cloud technologies. The situation becomes worse when applications are deployed in a collaborative manner. To support collaboration user must allow inbound traffics to its own VM from VMs of other Cloud Service Users (CSU). The usage of Security Groups (SG) has been introduced by cloud service provider to control the inbound traffic. This often leads to certain vulnerabilities in the system. In this paper, an agent based security framework is proposed for providing multilevel security for a collaborative cloud deployment. A secure communication protocol is also proposed for communication among the nontrusted groups. This can be used by the agents to control access between two VMs belonging to the same domain or in two different security groups. The proposed framework ensures that the trust-level of an entire domain does not fall due to malicious activities of only a small minority of members. The trust value of each member changes dynamically depending on its behavior.