Refine
Document Type
- Doctoral thesis (3)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3)
Language
- English (3)
Keywords
- Security (3) (remove)
Institute
With the introduction of the Internet at the end of the last century the modern society was fundamentally changed. Computer systems became an element of nearly all parts of our daily live. Due to the interconnection of these systems local borders are mostly vanished, so that information is accessible and exchangeable anywhere and at anytime. But this increased connectivity causes that physical fences are no longer an adequate protection for computer systems. Whereas the security of commodity computer systems was improved continuously and similarly with their increased connectivity, deeply embedded systems were then and now mostly protected by physical fences. But the ubiquitous availability of embedded systems in personal and commercial environments makes these systems likewise accessible and moves them strongly into the focus of security investigations.
Deeply embedded systems are usually equipped with tiny scale micro controllers, which are limited in their available resources and do not feature secure mechanisms to isolate system resources. Hence, a single error in a local software component is not limited to the component itself, instead the complete system may be influenced. The lack of resource isolation makes tiny scale systems prone for accidental errors but in particular vulnerable for a broad variety of malicious software. For a safe and secure operation of computer systems it is strongly recommended that software components are isolated in such a manner that they have access only to those resources, which are assigned to them. Even though a substantial number of approaches in the context of embedded system’s safety were investigated during the last fifteen years, security was mostly neglected.
This thesis is focused on security aspects where malicious software wittingly tries to bypass available protection mechanisms. The thesis introduces a security platform for tiny scale systems that enforces an isolation of software components considering security aspects. Due to the limited resources of tiny scale systems the proposed solution is based on a co-design process that takes the static and predefined nature of deeply embedded systems into account and includes hardware, compile-time, and run-time partitions to reduce the number of additional run-time components, to avoid performance drawbacks, and to minimize the memory as well as the components footprint overhead. To prove the applicability of the presented platform it was applied and evaluated with two real applications. In addition, an investigation of technologies of commodity computer systems that are suitable to build secure systems is presented. The thesis analyzes their enforcement based on the features provided by the introduced security platform. The contributions of this thesis include an enforcement of a security isolation of system resources on tiny scale systems and enable the development of a broad variety of secure tiny scale system applications.
The development of secure systems is already a very challenging task. In the domain of wireless sensor networks this challenge is even aggravated by severe constraints of the sensor node devices and the exposed character of the networks. To cope with this issue, this thesis proposes a tool-supported development flow named configKIT, that helps users to integrate secured applications in the domain of Wireless Sensor Networks. It is a component-based framework that selects and composes configurations of hardware and software components for WSN applications from high-level user requirements, automatically. Therefore, the composition process utilizes a flexible meta-model to describe properties of the components, the requirements, and the system semantics, which allows the assessment of the behavior of the composed system. Based on this modeling technology five practical security models are investigated, which base on different technical views on a general security ontology for WSNs. Each model is discussed theoretically and practically, based on a practical integration in the configKIT framework. The configuration toolkit and the security models are finally evaluated by applying the techniques developed to the non-trivial example of secure in-network aggregation. The evaluation shows that all five practical security models developed in this thesis work correctly and with reasonable model overhead. These results promote the notion of a practically applicable toolkit to configure secure applications in WSNs.
Key management is a fundamental security service to enable secure wireless ad hoc networks (WAHN). To date existing key management solutions based on either public key infrastructures (PKI) or key pre‑distribution scheme (KPS) exhibit limitations for WAHNs. We firstly develop the Hybrid Key Management Infrastructure (HKMI) for WAHNs composed of moderate‑resource devices. The HKMI complements PKI with trust and cooperation protocols to construct an performance efficient security solution. We secondly develop the Deterministic Pairwise Key Pre-Distribution Scheme (DPKPS) for large‑scale dynamic WAHNs composed of low‑resource devices. The DPKPS applies a combinatorial design for the pre-distribution of multiple bivariate polynomial shares to WAHN nodes. Future work comprises further improving the resiliency of the DPKPS, completing a key management infrastructure on the basis of the DPKPS, the design of DPKPS‑based access control mechanisms, and the integration of the HKMI with the DPKPS in a unified key management architecture.