128 pages ; Context-aware applications require context information for their operation. However, context information is inherently associated with uncertainties, which need to be taken into account when processing such information. This Quality of Context Information (QoC) comprises numerous aspects, such as the uncertainty of sensed data, transmission and update protocols, consistency between data from several providers, or the trust placed into the information from individual providers. Because of the technical and economical challenges, it is our vision that a federation integrates many context models from different providers into a global context model. This model is then shared by a number of applications. Such a federated context management poses additional challenges with regard to QoC. Therefore, a universal QoC model is needed, which covers and integrates the different aspects of QoC on all abstraction levels. To enable a federated context management system such as Nexus to incorporate QoC, we present a QoC reference model in this document. Specifically, it consists of five main parts: 1. Abstract framework for quality aspects: This framework distinguishes three abstraction levels of context information -- sensor information, observable context, and high-level context --, as well as three fundamental quality aspects -- degradation, consistency, and trust. 2. Degradation model: Sensed data is typically inaccurate due to physical limitations of sensors or the employed update protocols. 3. Consistency model: Context information -- even with degradation information -- may be inconsistent to the physical world or between different providers. 4. Trust model: With arbitrary context providers, a comprehensive approach for assessing the reliability and trustworthiness of each provider is necessary. 5. Integrated QoC-aware processing model: We propose a universal QoC-aware processing model for queries on context information. It incorporates the specific models for degradation, consistency, and trust and their dependencies.