Risk Assessment in Sensor Clouds

In sensor cloud, the security of a sensor network cannot be assessed by simply considering the feasible set of attacks in isolation. To accurately determine the security of a network, one must be able to determine the interdependencies between these attacks and how they can be exploited to execute a multistage attack on the sensor networks. These situations demand the need of risk assessment for a sensor network. Risk assessment using attack graphs will help the security administrators visualize the interdependencies between the various attacks and how they can be used in combination to cause more damage to the network.
With respect to sensor clouds which has two layers, a physical sensor network to perform the actual tasks and a virtualization of this sensor network which is hosted at the cloud platform, we will design risk modeling on both of these layers with the help of attack graphs and assess the interdependencies between the possible set of attacks on both of these layers and integrate them to assess the overall security of the sensor cloud platform. Attack graphs in a combination with principles like Bayesian networks will also help in estimating which of the given attack sets are most feasible to execute on the physical sensor layer and how that will affect the virtual layer on the cloud platform. We will also be able to determine the cost of implementing the security measures using the aforementioned principles, which will help in implementing network hardening strategies for the sensor cloud platform.

Off-line Risk Assessment of Cloud Service Providers

The acceptance of the cloud as an infrastructure to host applications is a growing trend. Facilitating and hosting applications on the cloud reduces support and maintenance costs. However, concerns about the security of these applications is one of the primary reasons organizations avoid complete adoption of cloud services. Although cloud service providers (CSPs) offer standard security, they don't address security with respect to application's security requirements. This project aims to propose an offline risk-assessment framework to evaluate the security offered by a CSP from the perspective (security needs) of an application to be migrated to it. Once the most secure CSP is identified for a given application, the proposed framework performs a cost-benefit tradeoff analysis in terms of the security dispensed and service costs to support the formation of an ideal cloud migration plan.

Researcher

Amartya Sen

Advisor

Dr. Sanjay Madria