CT-ER: Security and Privacy Solutions for Data-Centric Sensor Networks

题目：以数据为中心的传感器网络的安全和隐私解决方案PI: Sencun Zhu (szhu@cse.psu.edu), Co-PI: Guohong Cao （gcao@cse.psu.edu）随着传感器网络规模的扩大，产生的传感数据量也将随之扩大。大量的数据加上数据遍布整个网络的事实，产生了对有效的数据传播/访问技术的需求，以便从网络中找到相关的数据。这种需求导致了以数据为中心的传感器（DCS）网络的发展，其中传感器数据而不是传感器节点是根据事件类型或地理位置等属性命名的。然而，在网络中保存数据也会产生严重的安全问题，这是以前的工作没有解决的。本研究的目的是设计安全的DCS系统，提供两个基本的安全服务：数据保密性和位置隐私。考虑了攻击者能力不断增强的三种类型的攻击模型，包括本地被动攻击、传感器辅助被动攻击和基于妥协的主动攻击。设计了基于最小特权原则、虚拟流量和过滤、位置隐私和查询优化的防御技术。这项研究的成功将对传感器网络产生更广泛的影响，使其更经济实惠，适用于商业、民用和军事应用。它有可能促进传感器网络隐私方面的新研究。这项研究的结果将通过高质量的出版物和讲座广泛传播。这项拟议的研究也将与宾夕法尼亚州立大学的教育课程相结合。

Proposal Number: NSF-0627382TITLE: Security and Privacy Solutions for Data-Centric Sensor Networks PI: Sencun Zhu (szhu@cse.psu.edu), Co-PI: Guohong Cao (gcao@cse.psu.edu)As sensor networks scale in size, so will the amount of sensingdata generated. The large volume of data coupled with the factthat the data are spread across the entire network creates ademand for efficient data dissemination/access techniques to findthe relevant data from within the network. This demand has led tothe development of data centric sensor (DCS) networks, wheresensor data rather than sensor nodes are named based on attributessuch as event type or geographic location. However, saving data inthe network also creates critical security problems, which havenot been addressed by previous work.The objective of this research is to design secure DCS systemswhich provide two fundamental security services: dataconfidentiality and location privacy. Three types of attack modelswith increasing attacker capabilities are considered, includinglocal passive attacks, sensor-assisted passive attacks, andcompromise-based active attacks. Techniques based on the principleof least privilege, dummy traffic and filtering, location privacyand query optimization are designed to defend against the attacks.The success of this research will have a much broader impact onmaking sensor networks more affordable and amenable to commercial,civilian, and military applications. It has the potential tofoster new research in the privacy perspective of sensor networks.The results from this research will be disseminated widely throughhigh quality publications and talks. The proposed research willalso be integrated with the education curricula at Penn State.