Improving Security and Expanding the Range of Socially aware Single System Images
MetadataShow full item record
Cloud computing has enabled users to utilize computing resources as per their need and has freed them of the responsibility of maintaining their own dedicated computing resources. This enables the user to save on purchase and maintenance costs, while he/she pays the cloud provider a fee based on resource usage. But in addition to this, the user incurs an additional cost in the form of giving up his/her privacy. Recent revelations have revealed severe deliberate violations of privacy by major cloud computing companies. Given the remote and proprietary nature of computing resources belonging to cloud providers, there is no way to ensure privacy while using their resources. Thus there is a need to provision ways to be able to use computing resources belonging to people you can trust. SaSSI, at its core uses Single System Image (SSI) clustering technology which hides the discrete nature of the participating resources, and makes them appear as a single homogeneous computing resource to the user. Such a cluster can utilize all the available processing power and memory from its participating resources. SaSSI integrates social networking features with SSI technology, thus enabling friends / people who trust each other to share their resources with each other to create clusters. This environment provides the benefits of cloud computing while limiting the possession of sensitive user data to only a few trusted users. The major limitation in the current publicly available SSI software is that SSIs can only be formed using nodes which are within one-hop distance of each other. This implies that machines have to be within the same broadcast domain / Local Area Network (LAN) in order to form an SSI cluster. SaSSI aims to allow friends in different geographic locations to share resources to form SSIs over a WAN environment. In this research, we look at enhancing the underlying communication protocol used by SSIs to overcome the one-hop limitation and examine its feasibility in enabling SSI clusters over WAN nodes. Modern lightweight cloud enabled operating systems like ChromiumOS and Android are very user-friendly, have a host of security features present and could be used as frontend clients for accessing SSI clusters. In this research, we take a detailed look at their security features, describe some approaches to extract sensitive data by bypassing all security features and also discuss mitigation schemes to protect against the same.