A systems thinking view on cloud computing and energy consumption

Abstract

Cloud Computing is often regarded as form of "green computing". However, the growth of cloud computing also goes hand in hand with an increasing demand for energy, caused by a continuously growing demand for content, data-intensive applications and bandwidth. The debates on the energy implications of the substantial shift to cloud computing among industry experts, researchers and civil society organizations mainly focus on stand-alone aspects of this shift, whereas a comprehensive view on the whole system, its inter-relations and driving forces are still rare. In the course of an EU-funded project (www.SCP-RESPONDER.eu) the syntax of causal loop diagrams and analytical tools of system dynamics has been used to develop a system map of the diverse effect pathways that a shift to cloud computing may have on energy consumption. The map was developed in an iterative participatory expert consultation process involving 25 experts from industry, research and policy-making. The map allows for integrating different strands of knowledge, shedding light on the dynamics of the whole system and identifying leverage points for business strategies as well as policy interventions. It shows that the effects on overall energy consumption depend on (1) relative strengths of the individual effect pathways which are, inter alia, (2) dependent on energy efficiency gains achieved through a broad diversity of measures in very different areas, while (3) the key drivers of the whole system are economy of scales (through reductions of service cost per unit) and (4) new user needs for data-intensive services. While some experts perceive these dynamics as typical rebound effects (increased efficiency is compensated by growth), others regard them as success strategies for individual businesses and the sector as a whole. Both groups share the view that service reliability, privacy and security of cloud services are the most important leverage points of the whole system. Furthermore, even if, through the growth of data-intensive services, the shift to cloud computing would cause growth of the overall energy consumption, this increase might potentially be off-set by third-order effects in the form of savings in the areas of mobility or housing (e.g. teleworking).