Traditionally, Remote Laboratories have been focused on specific solutions for specific problems. We canfind a wide range of Remote Laboratories in the literature [1], assisting very different types of subjects (electronics, robotics, optics, fluids mechanics...), but commonly bound to a restricted set of requirements. Because of this, little attention has been paid on working on a scalable, maintainable, secure, open architecture that addresses the requirements of a wide set of experiments, and that could be open enough to support or adapt itself to new experiments. In this paper, we describe several aspects that might be taken into account when designing a Remote Laboratory architecture, resulted from the iterative development of our Remote Laboratory.