Research on Edge Intelligence-based Security Analysis Method for Power Operation System

Edge computing and cloud computing



259
Modern cloud computing platforms rely on virtualization service technology to effectively integrate and manage various system resources to provide users with efficient computing services and application requirements. Cloud computing is a kind of simple distributed computing. It can decompose huge tasks into countless small tasks, use the server farm for processing and analysis, and finally merge the calculation results back to the user [23], [24]. However, the large number of terminal devices access exposes the limitations of the cloud platform computing model. According to Cisco statistics, as many as 1.25 billion terminal devices connected to the Internet in 2010, it is expected to reach 50 billion units by 2020. Cloud computing is to provide elastic physical resources and virtual resources in a shared manner for service provision and management, while edge computing is to provide data-based services with distributed processing and storage at the edge nodes of the network. Edge computing reduces the pressure on the core nodes of the cloud network by performing data processing at the edge of the network. It is an ideal solution to achieve distributed intelligent management of large-scale smart terminals in the future. However, edge computing is not a substitute for cloud computing, but a supplement and extension to cloud computing. It provides rich, convenient, and flexible elastic resources for terminal devices on edge [25].



Fig. 1. Joint training framework for cloud-edge collaboration and edge-edge federation collaboration.
Generally speaking, edge computing has four characteris­tics: 1) Intelligence: Edge computing can be combined with artificial intelligence technology to enable terminal devices to handle more complex services. 2) Low latency: The edge computing platform sinks the computing task to the edge, and uses distributed computing to efficiently process at the source of the data, which can effectively shorten the response time. 3) Low energy consumption: The distributed architecture of the edge computing can reduce the data transmission between the cloud and the occupation of the network channel, thereby reducing the data processing cost and equipment operation energy consumption. 4) Reliability: Distributed edge equipment can provide on-site calculation and management functions for the system, and ensure the stable operation of local systems when the cloud center does not process in time or communication failures. Edge computing is suitable for real­time and short-cycle data analysis and local decision-making scenarios. In contrast, cloud computing is suitable for big data analysis of non-real-time and long-cycle data. Therefore, the collaboration between edge computing and cloud computing has many advantages. Edge computing is close to the data generation side. It is a data collection unit that provides data for cloud computing. It can support big data applications in the cloud and can ease the pressure on the cloud platform’s network bandwidth and computing storage. The calculation results and business rules formed by the cloud through big data analysis can also be transmitted to the edge side to improve terminal business processing capabilities.
III. Methods

1.