Energy Internet (also referred to as Smart Grid 2.0) is another promising application of the Industrial Internet of Things (IIoT), for example, in the way energy is being produced, traded, distributed, and consumed. This is partly due to the lowering of barriers (e.g., costs and Internet connectivity) and advances in the underlying technologies, such as smart meters, electric vehicles, and actuators. This has also resulted in significant growth in the volume, velocity, variety, veracity, and value of data (i.e., the 5 Vs of big data). However, efficiently and effectively handling such big data remains challenging. One solution currently being explored in the literature (including industry) to cope with the increasing network traffic is to use conventional cloud infrastructure, but the key limitations of such an approach include long response time and high bandwidth consumption. Therefore, there have been attempts to introduce next-generation Internet of Things networks to satisfy (real-time) network service demands and guarantee quality of service, for example, by pushing computing capabilities closer to the users (e.g., edge of the network). However, computation-intensive energy analytics can be challenging to perform at the network edge by edge or fog computing devices. Hence, there have also been attempts to utilize software defined networking (SDN) and network function virtualization (NFV) in order to improve network functionality while adding programmability and flexibility features to the network infrastructure. Specifically, NFV facilitates virtual network function (VNF) deployment and orchestration, while SDN controls them for specific use cases. However, with the rise of next-generation mobile networks (i.e., 5G), applications and services require fast and smooth operations with greater flexibility, efficiency, and scalability. In order to align with 5G and leverage potential benefits of edge computing, VNFs should possess critical processing requirements (e.g., high throughput, low latency, and minimal computation overheads). In other words, virtualization plays an important role. To date, several techniques have been introduced in order to achieve the desired objective using virtual machines (VMs) and containers in isolation. However, a hybrid approach using VMs and containers is likely to bring potential benefits for the large-scale deployment of VNFs across the heterogeneous edge and cloud platform. Thus, in this article, a novel architecture for SDN integrated with NFV, specifically for the Energy Internet ecosystem, is presented by leveraging the advantagesof edge computing. In the considered setup, the deployment of VNFs is achieved using a mix of both virtualization and containerization. Findings from our evaluation demonstrate the potential for VNF placement across a hybrid execution setup powered by VMs, as well as the benefits of using containers.

中文翻译：

---

SDN-NFV辅助的边缘云互动，可实现5G能源互联网生态系统

能源互联网（也称为智能电网2.0）是工业物联网（IIoT）的另一个有前途的应用，例如，在能源的生产，交易，分配和消费方面。部分原因是由于降低了壁垒（例如，成本和Internet连接）以及诸如智能电表，电动汽车和执行器之类的基础技术的进步。这也导致数据的数量，速度，多样性，准确性和价值（即大数据的5 V）显着增长。但是，有效地处理如此大的数据仍然具有挑战性。为了应对日益增长的网络流量，文献（包括行业）目前正在探索的一种解决方案是使用传统的云基础架构，但是这种方法的主要限制包括响应时间长和带宽消耗高。因此，已经尝试引入下一代物联网网络来满足（实时）网络服务需求并保证服务质量，例如，通过将计算能力推近用户（例如，网络边缘）来实现。 。但是，计算密集型能源分析可能难以在边缘或雾计算设备的网络边缘执行。因此，也已经尝试利用软件定义网络（SDN）和网络功能虚拟化（NFV）来改善网络功能，同时向网络基础结构中添加可编程性和灵活性功能。具体来说，NFV促进了虚拟网络功能（VNF）的部署和编排，而SDN则针对特定用例控制它们。但是，随着下一代移动网络（即5G）的兴起，应用程序和服务需要快速，流畅的操作以及更大的灵活性，效率和可扩展性。为了与5G保持一致并利用边缘计算的潜在优势，VNF应该具有关键的处理要求（例如，高吞吐量，低延迟和最小的计算开销）。换句话说，虚拟化起着重要作用。迄今为止，已经引入了几种技术，以便单独使用虚拟机（VM）和容器来实现所需的目标。但是，使用VM和容器的混合方法可能会为跨异构边缘和云平台的VNF大规模部署带来潜在的好处。因此，在本文中，通过利用边缘计算的优势，提出了一种与NFV集成的SDN的新型架构，特别是针对能源互联网生态系统的架构。在考虑的设置中，可以同时使用虚拟化和容器化来实现VNF的部署。我们评估的结果表明，在由VM支持的混合执行设置中，VNF放置的潜力以及使用容器的好处。