The integration of web technologies and web services has been, in the recent years, one of the major trends in upgrading and developing control systems for accelerators and large experimental apparatuses. Usually, web technologies have been introduced to complement the control systems with smart add-ons and user friendly services or, for instance, to safely allow access to the control system to users from remote sites. In spite of this still narrow spectrum of employment, some software technologies developed for high performance web services, although originally intended and optimized for these particular applications, deserve some features that would allow their deeper integration in a control system and, eventually, use them to develop some of the control system's core components. In this paper we present the conclusion of the preliminary investigations of a new paradigm for an accelerator control system and associated machine data acquisition system (DAQ), based on a synergic combination of network distributed cache memory and a non-relational key/value database. We investigated these technologies with particular interest on performances, namely speed of data storage and retrieve for the network memory, data throughput and queries execution time for the database and, especially, how much this performances can benefit from their inherent scalability. The work has been developed in a collaboration between INFN-LNF and INFN-Roma Tor Vergata.