In the vision of pervasive computing, technology is integrated throughout our environment. Wireless sensor networks (WSNs) provide support for pervasive computing applications through integration of environmental data. We believe that previous approaches to resource management in WSNs are designed to provide support for a specific application and do not adequately support the vision of pervasive computing. A separation of concerns between application logic and resource management is needed to provide adequate support for multiple concurrent applications using the WSN infrastructure. Additionally, sensor resource management middleware should provide appropriate primitives to support the creation of self-management frameworks. These frameworks should support the sharing of sensor network resources between multiple applications and should be extensible in order to comply with changing business needs, while leveraging on the existing infrastructure. This paper presents a component-based resource management middleware architecture that is designed to support the construction of large scale, self-managing distributed systems. We will show how the proposed architecture provides resource management support for concurrent applications while promoting reusability and extensibility for both sensor network infrastructure and software services. Furthermore our approach actively promotes point of action processing, which provides significant benefits in enterprise application deployments.