This paper presents experimental results which support our proposition that the increase in bandwidth efficiency achievable by dynamically and transparently matching transport protocols and messaging patterns to application requirements outweighs the necessary increase in complexity and overhead of the proposed architecture. In order to validate this claim, DIRECTOR, a middleware solution for managing Smart Grid applications communication requirements has been developed. The results show that distributed systems, such as the Smart Grid data network, can benefit from the scalability and transport layer flexibility provided by the proposed middleware solution by being able to rapidly negotiate transport level communication options at run time. This will allow applications to operate with increased network efficiency and better meet the Quality of Service requirements specified by the applications. The middleware solution will abstract these additional complexities from the applications and provide a simple communications interface which requires no network infrastructure modifications to be supported. The fundamental components of the middleware solution are demonstrated through the use of a real time implementation of the Open Automated Demand Response (OpenADR) Real Time Pricing (RTP) data model. The experimental test bed includes both real hardware nodes (Raspberry Pi Model B) and virtualised nodes residing in a custom network emulator.