Toward a standardized way for reporting on energy efficiency in the metro area network

Abstract

Energy is converted from one form to another through the activity of physical processes. The study of energy use, as it is converted from one form to another, therefore necessarily requires detailed understanding of the laws of physics that describe the behaviour of the entity responsible for the conversion (component: level 1 complexity). The complexity of the problem grows rapidly when these fundamental laws are not the ordinary means by which the behaviour of the entity is understood. This condition is common in systems: such aggregates encapsulate the behaviour of their components and obtain physical processes that are functions of the internal organization of these components (system of components: level 2 complexity). The complexity of the problem is compounded further when the ordinary means of interaction with the entity are no longer physical and material, but parametric representations of the entity’s function(s). These representations might be summarized as key performance indices; a more granular knowledge of the entity’s energy use may be obtained through study of the behaviour of its functions under a variety of operating conditions (multi-layered system of components: level 3 complexity). A fourth level in the hierarchy of complexity emerges with a localized system of systems; the fifth and final level of complexity is that of the geographically-dispersed system of systems. The complexity of the study of energy use by telecommunications networks falls into this fifth level. Several problems take root in this complexity. Diversity of components; diversity of systems; diversity of architectures; laxity in terminology; diversity of players, each interested in specific roles and layers, and abuse of abstractions are just some of the highly impactful ones. These problems lead to poorly defined studies of energy use, incorrect cross-comparison of studies, weak analytical technique and over-extrapolated prognoses. It must be conceded that, notwithstanding grave limitations, these works have sown interest in the field and spurred research into better methods. Perhaps this is a common trajectory in the development of our scientific knowledge of this wonderful world. I have primarily addressed the spatial aspect of the problem domain. Seeded by the observed laxity in architectural description and terminology, and driven by a documented failure arising out of misunderstanding of architectures, I have modelled the access portion of the metro area network in sufficient detail to support coherent analysis. Study was restricted to the metro area of the telecommunications network, as this was found to be the extent within a globally-spanning telecommunications network where fastest traffic growth was predicted. The market has been surveyed and the input gathered has been applied to validate my understanding, correct it, and to establish a firm foundation for future cycles of architecturally rigorous descriptions in support of the energy analyst. This work develops mutual understanding between industrial and academic practitioners in two disciplines: sustainability in ICT, and telecommunications operations. The two groups have been approaching one another over the past ten – fifteen years, and much effort has been put in by both sides to cooperate. Sustainability researchers want to reduce telecommunications’ Scope 1 (and beyond) greenhouse gas emissions; moreover, telecommunications network operators are keen to minimize the significant impact that energy use has on their operational expenditure. However, sustainability researchers have been hindered by the complexity of the object of their study, by the immaturity of methods, by the lack of methodology, and it is only recently that some consensus has emerged on good practice and the actual size of the problem (which, in the 1 – 2 % range of GHG emissions, is well short of more dire anticipations). On the other hand, while the operators are willing to share judiciously crafted questions, the detail of network architecture is not a matter of the public domain. The desire for rapprochement is there, but the modus operandi is still somewhat elusive. This work offers a contribution towards a solution of this problem. The standardized methodology of the implementational model has been applied to map the access network, and work is in progress to describe aggregation and metro-core. The models can be integrated with the software-defined networking paradigm. Since the implementational model describes functions and locates them relative to reference points, then it can be used within controllers to interact with service functions in the data plane. The prerequisite is standardized application programming interfaces, and standardized data models that incorporate energy and/or power usage. The former role can be fulfilled by NETCONF (RFC 6241); the latter role can be fulfilled by YANG (RFC 6020), but a valid contender for the latter role is the Green Abstraction Layer (ES 203 237, ES 203 682). The Green Abstraction Layer’s potential is investigated and its likelihood of adoption in the current data-plan driven exchange of link-state data is found to be poor. Regardless of whether GAL or YANG fulfil NETCONF’s content and operations layers, the energy-related notification data in the content layer cannot be generated without real-time power use models, as virtualization containers are not amenable to direct measurement of power use. The field of models is surveyed in a novel manner and contentious problems, productive approaches and significant developments are elicited.