Due to the diversity, magnitude and nature of the contextual information, a context-based service delivery mechanism is required that not only identifies the most appropriately matching services for the interested clients in a given context, but can also ensure the confidentiality of contextual information by restricting access to unauthorized clients. Various attributes of a user s context dictate the relevance of services that may belong to diverse domains. Ignoring the semantics and context of the user's query in the service-matching process can lead to low precision results with less or no relevance to the query at all. A context aware system can easily become invasive from pervasive if access to information is not closely monitored. The security requirements of the system are dynamic and change with the changing context. A two-pronged approach is proposed in this thesis to facilitate discovery of most relevant services without violating access restrictions (i) semantics-based search with context-based filtering for service look-up and (ii) dynamic policies and implicit context credentials for access control.