Service function chain reliable deployment method based on deep reinforcement learning

Abstract

The invention relates to a service function chain reliable deployment method based on deep reinforcement learning, and belongs to the technical field of communication. The method comprises the following steps: S1, obtaining a reliability value based on the reliability measurement mode of the equipment use degree and the peripheral safety coefficient; S2, preliminarily determining a reliability demand of each virtual network function through the functional characteristics and the topological characteristics; S3, obtaining the deployable length of the link reliability requirement meeting the virtual link reliability requirement; S4, searching an optimal mapping scheme suitable for the virtual network environment and the base environment by using deep reinforcement learning based on each reliability demand; and S5, if the VNF reliability cannot be satisfied in the mapping process, using an importance-based node backup method, and if the link deployment result does not satisfy the link reliability, using a link backup importance-based link backup method. The method can effectively deal with basic faults on the basis of ensuring reliability requirements, reduces the number of failure SFC, and ensures load balance to make the whole virtual network more stable and reliable.

Description

technical field [0001] The invention belongs to the technical field of communication, and relates to a reliable deployment method of a service function chain based on deep reinforcement learning. Background technique [0002] With the rise of cloud computing, big data and other technologies, people are increasingly using the Internet for work, shopping, video and other activities in their daily lives. While these technologies bring convenience to people, they also impose more requirements on network performance and architecture. The traditional network structure lacks flexibility and needs to deploy a series of dedicated hardware devices to provide services. As a result, each network function forms a strict and rigid network function chain. Its defect is reflected in the fixed position of the network topology and service provider components. To modify the network function chain means modifying the network topology or changing the connection of the middle box. When providin...

AI Technical Summary

Problems solved by technology

Therefore, the existing reliability description model is no longer accurate

Secondly, the existing schemes based on reliability mapping simply meet the overall reliability requirements, while ignoring the different losses caused by the failure of different virtual network functions. How to give each virtual network function the corresponding reliability to reduce the In the case of SFC failure rate, reduce the loss caused by failure

In addition, most of the current solutions to this type of literature are based on heuristic methods, which are easy to fall into local optimal solutions while reducing complexity.

Images