In this paper, we contend that ubiquitous computing environments will be highly heterogeneous, service rich domains. Moreover, future applications will consequently be required to interact with multiple, specialised service location and interaction protocols simultaneously. We argue that existing service discovery techniques do not provide sufficient support to address the challenges of building applications targeted to these emerging environments. This paper makes a number of contributions. Firstly, using a set of short ubiquitous computing scenarios we identify several key limitations of existing service discovery approaches that reduce their ability to support ubiquitous computing applications. Secondly, we present a detailed analysis of requirements for providing effective support in this domain. Thirdly, we provide the design of a simple extensible meta-service discovery architecture that uses database techniques to unify service discovery protocols and addresses several of our key requirements. Lastly, we examine the lessons learnt through the development of a prototype implementation of our architecture.