The former can be connected to the latter via the Internet of Things, consisting of appliances embedded with sensors, electronic chips, and connectivity to communications networks. Devices serving as a bridge between physical electricity networks and the digital world (e.g., digital meters or sensors) are considered part of the digital world rather than part of the physical electricity network, although they also possess the characteristics of a physical entity (ITU-T, 2015a, p. 55).

The physical microgrid consists of a low-voltage electricity network, connecting the participating prosumers and providing access to physical electricity networks, either of other microgrids or of distribution networks, via the microgrid node as illustrated in Fig. 13.3.

On the one hand, the aggregator is assumed to balance production and consumption of low-voltage electricity within the microgrid. On the other hand, the aggregator ensures the import or export of electricity from the wholesale market of the medium-voltage distribution grid, based on injection or extraction fees charged at the microgrid node (Knieps, 2016a). Participation within a microgrid is voluntary; other prosumers in the local neighborhood may either belong to another microgrid or be served by a local/regional utility. The term (physical) node is used throughout this chapter to describe points of electricity injection or extraction within the distribution and transmission networks where location matters. As within (low-voltage) microgrids the location of consumption, storage or production of electricity does not matter, the term prosumage gains relevance. Within a prosumage unit, several loads of different home appliances are relevant and, together with the small-scale storage and generation facilities as an aggregate, constitute the load of a prosumage unit.

It is important to differentiate between the ICT requirements inside a virtual microgrid and the outside requirements resulting from the connection either to other microgrids or to distribution networks. Standardization efforts within virtual microgrids are focused on different subparts, such as building-to-grid, home-to-grid, industry-to-grid, vehicle-to-grid, distributed renewables, generators, and storage (Fang et  al.,  2012; NIST,  2014, p. 43).

In the meantime, the challenge for ICT networks to take into account the requirements of the Internet of Things has gained increasing attention. “[T]he advancement of mass storage units, high speed computing devices, and ultra broadband transport technologies … enables many emerging devices such as sensors, tiny devices, vehicles, etc. The resultant new shape of ICT architecture and huge number of new services cannot be well supported with current network technologies.” (ISO/IEC, 2012, p. vi). The term Future Networks has been coined, drawing attention to the overall challenges to ICT developments: “network of the future which is made on clean-slate design approach as well as incremental design approach; it should provide futuristic capabilities and services beyond the limitations of the current network, including the Internet” (ISO/IEC, 2012, p. l).

The evolution of ICT physical network architectures is characterized by a change from specialized network infrastructures toward shared multipurpose IP-based communications infrastructure. In this context the concept of network virtualization has evolved: “Network virtualization is a method that allows multiple virtual networks, called logically isolated network partitions (LINPs) to coexist in a single physical network” (ITU-T, 2012, p. 2).

ICT logistics of microgrids possess the character of virtual networks reflecting the QoS requirements of data packets within microgrids and their interconnection to the multipurpose All-IP Internet. From the perspective of virtual networks, Recommendation ITU-T (2016a) on home networks can be considered as an important innovation complementary to All-IP communication infrastructures of access and core networks with seamless data packet transmission, irrespective of the physical communication network infrastructure (Cisco, 2014).

Focusing on virtual microgrids, the requirements of complementary application services and virtual networks comprise the identification-based connectivity between a “thing” and the Internet of Things, application support, security and privacy protection, interoperability, reliability, high availability, and adaptability. ICT-based networking of virtual microgrids also entails a large potential for enabling virtual power plants capable of dispatching significant amounts of energy to the wholesale markets as described in chapter by Steiniger.

Depending on the real-time price signals, prosumage behavior is incentivized. Given the sun and wind conditions for generation and the preferences for timing flexibility for consumption, rational prosumage decisions can take place to sell electricity to the virtual power plant. However, price signals should reflect the opportunity costs of marginal prosumage behavior. Separate metering of all embedded generation with different charges for generation and consumption leads to ample disincentives, which can only be avoided by net metering treating consumption and generation within a microgrid equally as proposed in chapter by Biggar & Dimasi. More generally it can be shown that the opportunity costs of generation and consumption (loads) within a microgrid are completely triggered by its outside options either to import or export from the wholesale market of the distribution network (via the microgrid node) or to trade directly with other microgrids in the neighborhood (Knieps, 2016a, pp. 276 ff.).

Adoption of IP provides end-to-end bidirectional communication capabilities between any devices in the network. New application fields entail demand/response, distributed energy resource integration, and electric vehicle charging. Significant challenges for data packet transmission networks arise to deal with the strongly increasing amount of IP addresses, rapidly growing data traffic, as well as mobility and QoS requirements due to complex bidirectional communications with increasing traffic volumes and low latency requirements.

The physical microgrid network can be connected to the virtual microgrid via the Internet of Things based on All-IP communication infrastructures, enabling the seamless application of different communication infrastructures, such as fixed telecom access networks, cable access networks, and mobile access networks (Knieps and Zenhäusern, 2015, pp. 339 ff.). Although sensor networks and other ICT components within microgrids could be based on non-IP standards, the advantages of IP protocol evolution toward IPv6 from the perspective of the Internet of Things is pointed out. The microgrids of the future are embedded within overall Internet-of-Things architectures of smart sustainable city infrastructures, and thus definitively based on the IP protocol (ITU-T, 2015a).

Hardware and software requirements of the application services are beyond the scope of the All-IP communication infrastructures and traffic management of NGNs and related QoS differentiations. Nevertheless, application protocols have also been developed by the IETF, for example, for billing and identification purposes (Knieps, 2015, p. 743). Additional efforts have been provided by ITU-T in general and also in relation to several cases focusing on the role of virtual networks in general and of home networks in particular.

Prosumer units act independently (noncooperatively), basing their consumption and generation decisions on scarcity signals (price signals) provided by the aggregator. The basic idea is that price signals from wholesale markets provide incentives for the prosumers to adjust generation, storage, or consumption using highly automated machine-to-machine communications, as illustrated in chapter by Steiniger. The aggregator collects the resulting consumption and generation quantities and orders the remaining quantities from the distribution grid operator. All decisions are on real-time quantities; location within a microgrid where consumption or generation takes place is irrelevant, only the total amount of net import or net export at each given moment in time is relevant. However, the location of the microgrid node within a distribution network is relevant because injection or extraction charges are to be based on the location-dependent opportunity costs of distribution network usage (Knieps, 2016a, pp. 275 ff.).

The European Commission (2014) considers data protection and security as an important challenge for smart metering, and more general smart grid systems. The aim is to guarantee end-to-end security and to refer to the interplay of general rules for data protection and sector-specific rules. The question arises, as to how guarantees of end-to-end security can be achieved via the principle of layer-based security for ICT networks beyond home networks. For example, data security can be promoted by prohibitions to exchange data between prosumer units and prohibitions for aggregators to send data of individual units to other units or to distribution operators, with only aggregated data allowed to be transmitted. Security standards within home networks are developed by ITU-T (2016b, pp. 348–376) based on advanced encryption, authentication, and confidentiality standards developed by NIST. The focus is on information metering and measurement, as well as information transmission, and the goal to avoid vulnerabilities enabling attackers to obtain user privacy, gain access to control software, or alter load condition to destabilize the grid in unpredictable ways (Fang et al. 2012, p. 26).

Network layer security (IP security/IPsec) is developed by the IETF in Kent and Seo (2005) and Baker and Meyer (2011). The focus is on the IP network layer, offering a set of security services for traffic at the IP layer in IPv4, as well as IPv6, including access control confidentiality (via encryption), data origin authentication, etc. These security services are provided at the IP layer, guaranteeing protection in a standardized fashion for all protocols that may be carried over IP. This architecture differs between a more time-consuming authentication and key exchange protocol step on the one hand, and the actual data traffic protection on the other. In addition to network layer security, further security measures can be installed on upper layers of the IP stack, including application layer–dependent security mechanisms, such as digital signatures (Baker and Meyer, 2011, pp. 15 f.).

Security within data link layer specifications for wireline-based home networking transceivers (ITU-T 2016b, pp. 348–376) can be implemented by means of encryption and authentication and key management procedures. Advanced encryption standards and encryption algorithms referring to standard specifications recommendations of the NIST are developed to guarantee the required security inside a network domain and the security of a network containing more than one domain.

In a topical report published in November 2016 the Broadband Internet Technical Advisory Group (BITAG, 2016) issues the following recommendations (BITAG, 2016, pp. iv–vii):

Powerline Access services are communication services connecting households or businesses to the aggregator or the utility. “With the arrival of G.9960 technologies, power line communications (PLC), which has traditionally been a “best effort, moderate speed” communications option, is now a very high speed, high quality communications path” (ITU-T, 2010, p. 6). However, the communication protocol between the home networks and the utility or the aggregators (e.g., between the meter and the billing entity) is considered outside the scope of G.9960 and could more adequately be located at network layer protocols, including IPv6 (ITU-T, 2010, p. 7).

The Automated Metering Infrastructure (AMI) is based on two-way communications systems requiring real-time data delivery and a high degree of cybersecurity for a large number of consumers (NIST, 2009). The communication protocol between meter and utility or billing entity is outside the scope of home networks G.9960 initiatives, but is compatible with IPv6 protocol of data packet transmission (ITU-T, 2010, p. 7). As G.hn supports the IP, it can be readily interconnected with IP-based access and core networks (Cisco,  2010,  2014).

The evolution from best effort TCP/IP Internet toward NGNs with active traffic management is required for future virtual networks, not only for microgrid applications, but also for the increasing variety of other smart networks, such as networked cars, smart road infrastructures, etc. (Knieps, 2015, pp. 743 f.). The basic definition of NGNs is as follows: “A packet-based network able to provide telecommunication services and able to make use of multiple broadband, QoS-enabled transport technologies and in which service-related functions are independent from underlying transport-related technologies. It enables unfettered access for users to networks and to competing service providers and/or services of their choice. It supports generalized mobility which will allow consistent and ubiquitous provision of services to users” (ITU-T, 2004, p. 2). The question arises as to how the required transition from best effort TCP/IP protocol toward active traffic management can be achieved to fulfill the QoS requirements of virtual networks and in particular of virtual microgrids.

Considering virtual microgrids and their relation to NGNs, complex capacity allocation problems arise regarding QoS guarantees of bandwidth capacities that are not yet identified within the following definition: “network virtualization should provide the capability of regulating the upper limit bandwidth usage by each LINP in order to maintain the overall throughput and performance” (ITU-T, 2012, p. 2).

Although both levels are vertically complementary they constitute two independent decision units: traffic network providers (QoS for All-IP network) and virtual service network providers. One basic principle is that property rights and decision competency for the traffic network providers of NGNs are different from those for the virtual service network providers. The advantage of network virtualization has been described as follows: “The virtual resource management is recommended to allow LINP operators to configure LINPs quickly without imposing complexities of physical network operation including management of network addresses and domain names” (ITU-T, 2014, p. 3).

Virtual networks are based on QoS differentiated traffic capacities. They constitute an essential input for the provision of downstream applications in the Internet of Things and cannot be bypassed by separated traffic management within virtual networks. NGNs may entail a large variety of network services implemented via the concept of virtualization of different application service networks that are all competing for the same physical infrastructure capacities. A large variety of different QoS requirements emerges due to Internet of Things, as well as the various Internet and telecommunications services. The question arises as to how to measure the opportunity costs of different QoS guarantees. Traffic architecture should enable the variety of application services within NGNs in such a way that the question of how to optimally solve the network capacity allocation problem can be addressed. This permits as a corollary to approach the question of how to allocate capacities for cloud computing, cooperation, and coordination between different virtual networks, and the division of labor between intravirtual network capacity allocation and QoS traffic capacity imported from the All-IP network. Under the assumption that the QoS class implementation of NGNs is chosen in such a way that traffic logistics are sufficient to enable the variety of application services, the capacity allocation problem can be analyzed.

Irrespective of how the microgrid ICT traffic management is organized, from the perspective of the traffic network providers which manage All-IP infrastructure capacities, only the bandwidth usage of the virtual service network is relevant. As it turns out, heterogeneous virtual service networks rely on traffic capacities that give differentiated QoS levels. Thus, not only are incentives created for the organizers of virtual networks to economize bandwidth consumption by optimizing intraservice network usage, but also the viability of virtual networks can only be guaranteed if the traffic network providers that operate the NGNs provide required QoS levels.

During the last decades there have been intensive efforts within standardization committees, in particular the ITU-T and IETF, to develop guidelines for IP QoS traffic classes (Babiarz et  al.,  2006; Ash et  al.,  2010, pp. 4 f.; ITU-T, 2011a, Table 2, pp. 12 f.; ITU-T, 2013). In particular, a monotone decreasing ordering of traffic classes regarding delay, jitter, and packet loss requirements has been developed. To encompass the variety of heterogeneous traffic requirements based on the All-IP broadband infrastructures, the concept of Generalized DiffServ architecture is applied, which enables QoS differentiation of different traffic classes with deterministic and stochastic traffic quality guarantees (Knieps, 2015, pp. 739 f.).

Describing the special case of stochastic QoS guarantees based on prioritization between data packets, a pricing scheme based on interclass externality pricing was developed by Knieps (2011). Knieps and Stocker (2016) present an extended model describing the case of deterministic traffic classes (without explicitly including a hierarchy) and lower traffic classes with stochastic QoS guarantees, and introduce an incentive-compatible pricing scheme based on interclass externalities between deterministic and stochastic traffic classes. The more general case of a hierarchy of deterministic traffic classes is modeled in Knieps (2016b). Traffic classes requiring deterministic guarantees for higher QoS levels consume relatively more bandwidth. If QoS parameters are more strictly defined (e.g., the highest traffic class guaranteeing very low latency and packet drop probability) bandwidth capacity requirements are higher than those required for providing less stringent levels of QoS.

This permits the implementation of economic incentives for the underlying capacity allocation problem in NGNs, resulting in economic priority ordering between different QoS traffic classes. Differentiated pricing for a hierarchy of traffic classes with deterministic, as well as stochastic QoS guarantees, is based on the opportunity costs of network usage for providing the relevant QoS requirements in different virtual networks. Thus, active entrepreneurial traffic management of NGNs providing QoS guaranteed ICT traffic services is essential for meeting the QoS requirements of smart grids and virtual microgrids.