Monday, October 31, 2011

Next Generation Utility Networks -PoIP, CoIP, IoIP and others

Economies of scale along with the benefits of statistical multiplexing are moving the life cycle of Time Domain Multiplexed services to the winter of their years as Ethernet and its derivatives take the throne. 

Utilities for decades with staff trained and experienced in synchronous and asynchronous communications are now faced with change as ques, buffers and ports filled with frames, statistical multiplexing with prioritized packets, flags, shims, headers and all the glorious complexities of IPv4 and IPv6 overhead in the IP world. It means a great deal of learning to operational staff and owners who not only must find comprehension of the new technology but deploy it where failures can affect commerce and safety of industrialized nations.

Emotionally this creates uncertainty and fear of how this will impact daily operations of an electrical utility and resentment in a few and excitement in others about the changes. Change even advances are not always welcomed by all. New protocols like 61850 that open new possibilities for electrical operation have been emerging for nearly a decade as many leading edge thinkers and innovators saw the need and opportunity with IP to go beyond traditional SCADA networks. Protection and Control engineers have been working hard to learn how to implement these new protocols and methods with existing systems while also being confronted with the heady task of learning networking.

Secondly, the IP protocol is most often associated with an open network originally envisioned by DARPA to be reachable no matter the catastrophe through statistically a high number of paths where the probability of absolute failure (inability to connect) was near nil.  Utility engineers in protection and control view in general conversation this same "internet" approach as being dangerous to their
critical infrastructure since a router anywhere in the world can send a packet to any other discoverable router on the internet. The routers job is to find a route. However, the next generation utility operational network should have fundamentally different approaches than the internet first and most obvious is that it is a private network. (We will ignore the case where leased 3rd party services are used since this does expose a vulnerability do to a lack of privacy in channels and facilities.)

Private next generation intelligent utility networks should emerge than as closed private network domains where instrumentation, control and protection devices need only communicate within prescribed domains and across domains where the internal and external communication enhances situational awareness and control. Intelligent grids should have their IP networks aligned in architecture and domains with the electrical system with traffic very planned and controlled amongst domains where router penetration is limited to high tier facilities and core backbones. An intelligent grid then is not "open" like the internet and device communities are well established by design.  However, the survivability aspect of the internet aided by the overhead should be high even with a finite amount of redundant paths. The advantage being with Carrier Ethernet and MPLS that protective tunnels and FRR are soft connections only and while providing higher levels of network protection do not consume capacity under normal conditions.
(Security is fundamental to the architecture and design of a closed but functional system.)

Creating the intelligent grid network then as a private network with high survivability using the concepts of domain communities and the "soft-state" main and protection paths for instrumentation, control and protection requires defining those communities and their subsequent requirements for survivability, security and membership. New terms will emerge as Virtual LANS and VPLS are used to define device communities in alignment with "Networked Electrical Systems Topology" concepts and distributed switching is used to minimize routing at the electrical grids distribution edge. Terms like
Protection over IP (PoIP©), Control over IP, (CoIP©) and Instrumentation over IP, (IoIP©) will become common parlance over time as IETF and other standards forums focus networking schemes and protection toward utility application much as enterprise has focused IP efforts around specific voice and video service requirements. Historically, we have evidence of utility standards bodies influencing communication standards in the emergence of C37.94 and the DS0 requirements and coding for noisy utility environments which indicates as utilities learn the technology they are quick to "tweak it" to their particular needs. An intelligent grid network will emerge that is private, service focused and highly reliable suiting the specific needs of electrical utilities from generation to consumption because the utilities staff will understand IP and Protection and Control engineers will find a new friendship and close relationship with the network engineer.



Cooperatively, they can enable advanced substation automation services (SAS), grid automation services (GAS) and ensure a more stable consistent power delivery in the face of growing energy demand and high levels of intermittent green energy integrated in the network of power sources.