Research And Implementation Of Fault Management For 5G Network Slice

The design of the 5G network no longer follows the networking mode of the traditional network,but uses network slicing technology to enable operators to build multiple dedicated,virtualized,and isolated logical networks based on a shared underlying physical network.Network slicing technology ensures that different types of end-to-end network services are available to customers with different needs.This thesis mainly studies the fault management of network slicing.Aiming at the characteristics of many fault points,large amount of fault data and scattered fault management functions in network slicing,this paper proposes a fault management scheme based on association rules mining algorithm.The scheme mines the association rules by collecting a large number of fault alarm data,finds the correlation between them,and uses the generated rules to construct the association rule base.Then,in the actual fault management,the alarm data generated by the network slice is analyzed by the rules in the rule base to perform fault location and early warning.Finally,the processing flow of the network slicing fault is explained,so that the faulty resource can be deleted or replaced in time after the fault is located.When implementing the association rule mining algorithm,the algorithm is improved in two aspects.First,according to the characteristics of the alarm data generated by the network slice,the weighted idea is injected into the algorithm,and the corresponding weight is set for each alarm data.Second,the selection of the minimum support degree in the algorithm is adjusted,and the alarm data is more rationally screened.After simulation verification and result analysis,the improved weighted association rule algorithm is more suitable for network slice fault management,which improves the performance of network slicing.