The emergence of novel pervasive networks that consist of tiny embedded nodes have reduced the gap between real and virtual worlds. This paradigm has opened the Service Cloud to a variety of wireless devices especially those with sensorial and actuating capabilities. Those pervasive networks contribute to build new context-aware applications that interpret the state of the physical world at real-time. However, traditional Service-Oriented Architectures (SOA), which are widely used in the current Internet are unsuitable for such resource-constraint devices since they are too heavy. In this research paper, an internetworking approach is proposed in order to address that important issue. The main part of our proposal is the Knowledge-Aware and Service-Oriented (KASO) Middleware that has been designed for pervasive embedded networks. KASO Middleware implements a diversity of mechanisms, services and protocols which enable developers and business processing designers to deploy, expose, discover, compose, and orchestrate real-world services (i.e. services running on sensor/actuator devices). Moreover, KASO Middleware implements endpoints to offer those services to the Cloud in a REST manner. Our internetworking approach has been validated through a real healthcare telemonitoring system deployed in a sanatorium. The validation tests show that KASO Middleware successfully brings pervasive embedded networks to the Service Cloud.