Layered computing power network arrangement method and device and storage medium

Abstract

The invention provides a hierarchical computing power network arrangement method and device and a storage medium. According to the embodiment of the invention, the computing power distribution, the network connection and the application deployment of the users / tenants are realized by adopting the idea of hierarchical arrangement. One layer is a regional computing power network arrangement module arranged close to the region where the computing node is located, and the other layer is an end-to-end computing power network arrangement module which is arranged on the regional computing power network arrangement module and is responsible for global end-to-end computing power network arrangement. A regional computing power network arrangement module is arranged near a computing node region, andcompared with cross-region arrangement, the reliability of connection is improved. Meanwhile, each area is provided with an area computing power network arrangement module, scheduling can be conductedaccording to the strategy of the area, and good flexibility is achieved. Through the method, the computing resource utilization rate can be optimized, and the network efficiency is improved.

Description

technical field [0001] The present disclosure relates to the technical field of cloud computing, and in particular to a layered computing power network orchestration method, device and storage medium. Background technique [0002] Artificial Intelligence (AI), 5G technology promotes the widespread application of computing-intensive services, and promotes the vigorous development of emerging services such as the Internet of Things, big video, and Internet of Vehicles. The existing network architecture and cloud centralized deployment methods are difficult to meet business needs. Edge computing enables networks and clouds to migrate to the edge along with business flows, which has become a new trend in the industry. [0003] Computing power has become a key factor supporting the development of an intelligent society. Computing power is distributed at different network levels (terminal, edge, cloud), and computing power resources are different at each level. At the same time, t...

AI Technical Summary

Problems solved by technology

However, existing business applications lack network collaboration, and the best service processing node cannot be determined without user access information in the network, making it difficult to meet low-latency requirements

[0005] A large number of businesses need more guaranteed computing power resources. Existing computing resource nodes lack flexible interactive scheduling capabilities. When computing resources are overloaded or even unavailable, they currently rely on client application layer reconnection, and the waiting time reaches seconds / minutes level, leading to a decline in service experience

Today's business is statically deployed, resources cannot be reused, and there is also a lot of waste of idle resources

Images