Adaptive security in smart spaces [Älytilojen mukautuva tietoturva]

Smart spaces – like smart homes, smart offices and smart cities – exploit variousresources in order to offer enriched services and information for the end users.Achieving security in such a dynamic and heterogeneous environment with pre-defined and static security mechanisms is a challenging task. Hence, solutions forself-adaptive security are needed. Self-adaptive security is able to automaticallyselect security mechanisms and their parameters at runtime in order to preservethe required security level in a changing environment.The research problem of the dissertation is how to achieve security adaptationin a smart-space application. For this dissertation, architecture and knowledge baseobjectives were set. The objectives were satisfied with security-adaptation architec-ture that contains an adaptation loop and an ontology-based knowledge base forsecurity. The adaptation loop conforms to the Monitor, Analyse, Plan, Execute andKnowledge (MAPE-K) model, which is a widely applied reference model in auto-nomic computing. The ontology-based knowledge base offers input knowledge forsecurity adaptation. The research was carried using five cases, which iterativelydeveloped the architecture and the knowledge base for security adaptation.The contributions of the dissertation are: Firstly, reusable adaptation architec-ture for security is presented. The architecture strictly conforms to the MAPE-Kreference model and defines all phases in it. Moreover, the architecture is the firstthat specifically separates security knowledge from the adaptation loop. Secondly,the architecture supports the utilisation of security measures to recognise an ad-aptation need. Security measures are presented by means of a three-level struc-ture in order to achieve systematic monitoring. Due to the suggested architecture,it is possible to reuse and extend the defined security measures. Thirdly, this is thefirst time that an ontology has been applied for security adaptation. Hence, theInformation Security Measuring Ontology (ISMO) acts as the knowledge base forthe security adaptation. The ISMO is applicable at design-time and runtime alike.At design-time, the ISMO offers knowledge for the software architect, in order todesign an application with security-adaptation features. In contrast, the applicationsearches knowledge from the ISMO at runtime, in order to automatically performthe security adaptation. Utilising the ontology as a knowledge base ensures thatthe knowledge is presented in a reusable and extensible form. Moreover, theapplication does not need hard-coded adaptation knowledge.