Secure Networked Collaboration

From TCS Group Internal Wiki

Participants: Sandberg (Project leader), Dán, Stadler, Buchegger, Dam, Wikström, Johansson.

The operation of complex, networked systems relies on the information exchange between the constituent nodes. A networked control system will use interaction between neighbours to update their estimates of the systems state, and use this information as a basis for decision making. In a P2P-based social network the nodes use neighbour („finger”) interactions to maintain a distributed hash table (DHT) across which content is distributed. In emerging scalable architectures for network management, aggregation trees are often constructed and maintained through neighbour interactions, and used to aggregate global state information. All these examples raise important and challenging security concerns which are currently unresolved.

The core of the problem is that nodes need to collaborate in order to achieve their joint objective (of maintaining a DHT intact, for instance, or of accurately estimating or controlling a distributed state). This interaction opens up for attacks. An attack could be mounted by an intruder penetrating a node or communication channel, or by a node deliberately wishing to affect the outcome of the computation, e.g. by denying service to some competitor, or distorting the outcome in some specific direction. Also privacy concerns are relevant, as the neighbour interactions will in general cause information to be exchanged that may violate confidentiality, e.g. in the context of a multi-provider network management application.

Project objectives:

The project objective is to develop new algorithmic techniques to make complex networked systems robust to attacks. This includes the prevention and detection of attacks, localization and identification of non-cooperative nodes, and mitigation of attacks, including the use of cryptographic techniques. We propose to examine these problems initially in four concrete settings:

• Secure state estimation in electric power networks.
• Security and privacy in distributed network management.
• Secure autonomous flocking in autonomous robot networks.
• Detection of price and information manipulation in deregulated power grids.