Giovanni Vigna, Fredrik Valeur, Jingyu Zhou and Richard A. Kemmerer
Reliable Software Group
Department of Computer Science
University of California Santa Barbara
Security analysis should take advantage of a reliable knowledge base that contains semantically-rich information about a protected network. This knowledge is provided by network mapping tools. These tools rely on models to represent the entities of interest, and they leverage off network discovery techniques to populate the model structure with the data that is pertinent to a specific target network. Unfortunately, existing tools rely on incomplete data models. Networks are complex systems and most approaches oversimplify their target models in an effort to limit the problem space. In addition, the techniques used to populate the models are limited in scope and are difficult to extend. This paper presents NetMap, a security tool for network modeling, discovery, and analysis. NetMap relies on a comprehensive network model that is not limited to a specific network level; it integrates network information throughout the layers. The model contains information about topology, infrastructure, and deployed services. In addition, the relationships among different entities in different layers of the model are made explicit. The modeled information is managed by using a suite of composable network tools that can determine various aspects of network configurations through scanning techniques and heuristics. Tools in the suite are responsible for a single, well-defined task. Each tool has an abstract specification of the input, the output, the type of processing, and the requirements for carrying out a task. Tool descriptions are expressed in a Network Tool Language. The tool descriptions are then stored in a database. By using the network model and the tool descriptions, NetMap is able to automatically determine which tools are needed to perform a particular complex task and how the tools should be scheduled to obtain the requested results.
Keywords: Network Security, Network Modeling and Analysis, Network Discovery and Validation
Read Paper (in PDF)