United Technologies Corporation (UTC), the largest single supplier of equipment and systems to the global building industry, has a number of initiatives in sustainability and integration to impact the commercial buildings market. The UTRC Integrated Building Energy and Control Systems (IBECS) initiative is aimed at creating high performance buildings that meet or exceed the occupant comfort and safety needs at lower life cycle costs and resource utilization than typically realized using conventional methods. Real-time knowledge of dynamic indoor environment parameters, such as of occupant distribution, thermal and airflow state, are critical to the management and optimization of building occupant comfort and safety. Practical examples illustrating the role of multi-scale dynamics in building operations will be provided. Reduced-order models developed at UTRC will be described to investigate the spatially inhomogeneous and temporally rich dynamics of air and contaminant transport and people traffic. Progress in real-time estimation of airflow and traffic distribution in buildings, utilizing information from a large network of building sensors, to enable monitoring and control of the indoor environment will also be presented. Challenges in dealing with model complexity, multi-scale dynamics and sensor configuration will be discussed, for which several analytical methods are being developed at UTRC. In particular, we will present an approach that could be used to radically accelerate the numerical simulation, model reduction, and propagation of uncertainty in large multi-scale nonlinear networks of interconnected dynamic components. In this approach a large network is decomposed into subcomponents using spectral graph theory. Operator theory and geometric dynamics methods are used for model reduction and to analyze propagation of uncertainty in subcomponents. We will show how this approach would enable model-based design of robust building systems. We will indicate challenges in the design of future networked embedded systems for control of building systems with spatially distributed dynamics, and how solving them will require the tight integration of IT systems (including algorithms, computational architecture, and communications) with control, and dynamics.