Network slice dynamic resource distributing method based on MDP

Abstract

The invention relates to a network slice dynamic resource distributing method based on a Markov decision process (MDP), and belongs to the field of the mobile communication. The method comprises the following steps: guaranteeing serious delay requirements of all delay sensitive operations in each network slice to realize dynamic resource distribution performed by the taking the compromise betweenthe maximum network throughput capacity and the minimum energy consumption as a target; and distributing appropriate computational resource amount for a virtual network function (VNF) on a service function chain (SFC) of each user in the slice and dynamically adjusting the amount of the started server for providing the computational resource according to cache queue state information of the user in each network slice and real-time price state information of the electric power consumed by the server on each discrete time slot. Through the network slice dynamic resource distribution method basedon the MDP provided by the invention, the delay constraint of the user operation can be satisfied while the compromise between the maximum network throughput capacity and the minimum energy consumption is realized.

Description

technical field [0001] The invention belongs to the field of mobile communication, and relates to an MDP-based network slice dynamic resource allocation method. Background technique [0002] There will be a variety of application scenarios in the future 5G network, among which three typical application scenarios are enhanced mobile bandwidth (eMBB), massive machine type communication (mMTC) and ultra-high reliability and low latency communication. Application scenarios The difference of each means that the requirements in terms of mobility, security, delay and reliability are different. Therefore, how to meet the diverse service requirements and the access requirements of massive terminal connections under the expectation of low cost and high capacity has brought huge challenges to the existing network. It is unrealistic to build separate physical networks for different application scenarios, but wireless virtualization technology can realize the sharing and isolation of ne...

AI Technical Summary

Problems solved by technology

[0003] In the existing network slicing resource allocation schemes, most of the research on network slicing virtual resource allocation does not consider both delay-sensitive services and non-delay-sensitive services in network slicing, that is, delay-sensitive services are not considered Strict latency requirements, and does not take into account that the cost of electricity in the infrastructure depends not only on the energy consumption of the server, but also on the real-time price of electricity

In addition, the resource allocation problems currently studied are either static, or do not consider the impact of the decision made in the current time slot on the future income, that is, the optimization goal is to optimize the single-step income, without considering the system in a long-term earnings

Images