Přehled o publikaci

Cyber ranges enable security experts to emulate computer networks where they can perform cyber security exercises and experiments. Although their architecture may differ, the following four types of services are usually provided: (a) Resource management, i.e. the allocation of network infrastructure with requested topology and running applications. (b) Interaction of users with hosts, i.e. allowing users to log into hosts and run applications in the same way they interact in real computer networks. (c) Monitoring services, when network activities are monitored on the fly and measured data are stored and mediated to end users on demand. (d) Learning and understanding of cyber security processes by providing users with a continuous overview of events and developments by means of analytic tools, interactive visualizations, and other enhanced user interfaces. This paper deals with (c) and (d) service types. The low-level infrastructure of modern cyber ranges is usually generic, enabling to instantiate topologies and hosts of many types and then to support end user with variable objectives and tasks. However, this diversity of objectives put big demands on the design of a monitoring subsystem and corresponding user interfaces providing insight into the measured data and cyber security processes. It is because the monitored data and their manipulation strategies may differ scenario to scenario. This paper discusses a generic approach to data storage using traditional entity-relationship databases. Our approach exploits data analysis patterns enabling us to define scenario-specific security phenomena without modifying rational scheme or other parts of the cyber range architecture. A flexible web-based system of user interfaces adaptable to the variable data is discussed as well. Our approach has been developed as part of a KYPO Cyber Range. Its usability has been evaluated within many diverse successfully performed cyber exercises. This paper presents several types of exercises that differ in required phenomena and interactions and then sufficiently demonstrate generality and usability of our approach.