Cryptography for (partially) compromised sensor networks Petr Hanáček hanacek@fit.vutbr.cz Faculty of Information Technology Brno University of Technology Brno, Czech Republic Petr Švenda svenda@fi.muni.cz Faculty of Informatics Masaryk University Brno, Czech Republic Abstract This paper * will introduce the technology of wireless sensor networks with a special focus on its security issues. This relatively young technology started to evolve together with the advance in miniaturization of electronic devices, decreasing costs and general spread of wireless communication. Data sensed by the miniature devices in a target area (e.g., temperature, pressure, movement) are locally processed and then transmitted to end user who obtains the possibility to continuously monitor target environment. The usage of the technology starts from medical monitoring of the patients over agriculture and industrial monitoring or early warning emergency systems, ending with uses for military purposes as well – that is where the technology originally started. We will cover the issue of design of a key distribution and establishment protocols secure against the partial network compromise in more details. Special focus will be given to possibility for its automated generation of protocols for particular network scenario. Opposite direction will be covered as well – automated search for attacker's. We will also cover possibility to introduce low-cost tamper resistant hardware to sensor nodes without increasing the node cost and battery consumption significantly – the scenario usually not assumed in current research papers in the field. Keywords: key exchange protocols, evolutionary algorithms, wireless sensor networks, tamper resistance 1 Introduction Advance in miniaturization of electronics opens the opportunity to build devices that are small in scale, can run autonomously on battery and can communicate on short distances via wireless radio. These devices can be used to form a new class of applications, Wireless Sensor Networks (WSNs). WSNs consist of a mesh of a several powerful devices (denoted as base stations, sinks or cluster controllers) and a high number (10 2 - 10 6 ) of a low- cost devices (denoted as nodes or motes), which are constrained in processing power, memory and energy. These nodes are typically equipped with an environment sensor (e.g., heat, pressure, light, movement). Events recorded by the sensor nodes are locally collected and then forwarded using multi-hop paths to a base station (BS) for further processing. Wireless networks are widely used today and they will become even more widespread with the increasing number of personal digital devices that people are going to be using in the near future. Sensor networks form just a small fraction of future applications, but they abstract some of the new concepts in distributed computing. WSNs are considered for and deployed in a multitude of different scenarios such as emergency response information, energy management, medical and wildlife monitoring or battlefield management. Resource- constrained nodes render new challenges for suitable routing, key distribution, and communication protocols. Still, the notion of sensor networks is used in several different contexts. There are projects targeting the development of very small and cheap sensors (e.g., [1]) as well as research in middleware architectures and routing protocols for self-organizing networks – to name a few. * The paper is significantly edited version of the papers originally presented at SPI’09 and MKB’10.