Enabling the automated deployment and maintenance of current and future services over the Internet is a difficult task and requires self-aware functions that can adapt the services offered by the network to changing customer demand, business goals, and/or environmental conditions using policies. This paper describes a process to realise static and dynamic service deployment, given an understanding of available resources that exist in a communications network to build services . This process contributes to the realisation of the AutoI autonomic management architecture for the Internet, which aims to develop a self-managing virtual resource overlay that can span across heterogeneous networks and supports service mobility, security, quality of service and reliability. The core of the process is to take advantage of the substantial DEN-ng information model that decouples the definition and design of services from the resources available in the network. In this way, the creation and modification of services within a network can be planned, deployed, and even dynamically composed to meet context-aware demands. Providing such a service-aware process is central to the architecture being defined in the AutoI project and information modelling is seen as a major facilitator to this task. In this respect, we introduce a method of analysing a description of a service (including its demands on resources) within the context of currently available and deployed resources and services, in order to make informed decisions as to whether the service can be effectively deployed or not. The specific scenario we investigate is that of a secure VPN service that requires a set of security related functionality from the network in order to be effectively deployed. In accordance with the virtual resource overlay aspect of the AutoI architecture, virtual resources are modelled and used to realise the deployment of the new service. The scenario is implemented and validated in our test bed, where the service and resource characteristics are altered to determine the impact on the deployed service.