Future homecare networks will consist of a very wide range of embedded services and software that will often rely on numerous other components to achieve their tasks. They will rarely operate in a self sufficient manner. The ability to discover and use services is not however a trivial task. Services may provide raw data, such as temperature readings, or higher contextual data, such as user activity and availability. Networks may change over time and may not be subject to a single management regime, implying the need for a great deal of self-reliance for any software component seeking services from elsewhere within the network. This chapter describes work carried out at the University of Stirling to improve service discovery and allow it to operate effectively in networks with a significant turnover in services. Simple syntactical keyword lookups are insufficient, and so semantics are introduced into the discovery process by using ontologies. However ontologies are known to grow and change over time and so maintaining them can be difficult and error-prone. The described approach employs a hierarchical approach that fosters re-use and sharing of ontologies to alleviate some of the more acute problems of building and maintaining large ontologies.