Near-End Cloud Computing: Opportunities and Challenges in the Post-Cloud Computing Era

Cloud computing is a form of network computing paradigm that can provide on-demand access to shared pools of configurable computing resources, such as computation, storage, and services. Since its introduction in 2007, it has attracted a lot of attention and has developed rapidly over the last decade. In recent years, with the rapid development of pervasive smart devices and ubiquitous network technologies, various kinds of new network applications and services are continuously emerging, such as the Internet of things, smart city, smart grid, virtual/augmented reality, unmanned vehicles. However, it is very difficult for cloud computing to meet the diverse requirements of these new network applications and services due to the characteristic of centralized computation and storage. The capabilities of heterogeneous devices and edge equipment are not fully leveraged in cloud computing. Cloud computing also encounters the bottleneck of heterogeneous and long-distance networks. Moreover, the demands of new complex network application scenarios, such as the Internet of Things and 5G, cannot be satisfied by cloud computing. To overcome the limitations of cloud computing, a variety of new network computing paradigms, such as fog computing, multi-access edge computing, and edge computing are proposed by industry and academia. Although they employ different terms, the basic idea of them is to extend cloud computing and move the computing infrastructure from remote data centers to edge routers, base stations and servers that are close to end-devices, thus overcoming the technology and application bottlenecks of cloud computing. In this paper, we collectively refer to these new network computing paradigms as near-end cloud computing paradigm. To make it easier to understand the underlying causes of the near-end cloud computing paradigm and to grasp the application scenarios and challenges of it, we systematically summarize the emerged near-end cloud computing paradigms in recent years and sum up the basic characteristics of them. First, the main technology and application bottlenecks encountered in the development of cloud computing are summarized and analyzed in detail. Second, the concept and architecture of the near-end cloud computing paradigm are introduced, and several near-end cloud computing paradigms, including fog computing, multi-access edge computing, and edge computing are summarized. This shows that these near-end cloud computing paradigms are not completely different from cloud computing, but rather a natural extension of cloud computing from large-scale centralization to small-scale centralization and distribution, which can be regarded as a historical regression to the distributed computing paradigm. Third, the development opportunities associated with near-end cloud computing are discussed with application scenarios, including the Internet of Things, 5G, smart grid, virtual/augmented reality, and unmanned aerial vehicle. Fourth, the technological challenges including architecture and model, network access and management, resource management and scheduling, pricing and incentive mechanism, security and privacy protection are discussed in details. Finally, the future development of near-end cloud computing is prospected. It is believed that as problems faced by cloud computing become increasingly prominent, near-end cloud computing will undoubtedly become a hot research topic in both the industrial and academic communities in the near future. © 2019, Science Press. All right reserved.