Design Of Deterministic Radio Resource Scheduling Algorithm For 5G EMBB

Deterministic network refers to an end-to-end ultra-reliable data network with extremely low latency and jitter.It originally originated from a wired network,but until now,the implementation of deterministic network in 5G has also become a new popular application scenario.When the concept of deterministic network is extended to 5G,it includes not only ultra-high reliability and low-latency communication(URLLC)services,but also real-time streaming services that have explicit requirements for latency in enhanced mobile broadband(e MBB).More broadly,it includes all services that have reliability requirements for delay.How to reasonably allocate resources among multiple users under the condition of limited radio resources is an important research direction for achieving determinism in 5G systems.After extensive research and comparison of existing radio resource scheduling algorithms,it is found that these algorithms have their own shortcomings from the perspective of achieving determinism,including that cannot guarantee the delay of flows during their lifetime,the implementation of the algorithm requires some specific assumptions,and does not match the actual scheduling framework so it cannot be directly used in practice,etc.In response to the above problems,this thesis proposes a deterministic radio resource scheduling algorithm with a three-layer architecture for e MBB,which can guarantee the queuing delay of as many flows as possible when the system capacity is limited.The key idea of the algorithm is to first convert the delay requirement of flows into the requirement of the amount of data to be transmitted,and then establish the optimization problem in order to allow as many flows as possible to meet the data transmission requirements,solve the problem using greedy strategy to get specific scheduling algorithm and scheduling priority metric,and finally adapt to the actual scheduling framework,so that the algorithm can be directly used in practice.However,the above deterministic radio resource scheduling algorithm does not take into account the slicing characteristics of 5G.To solve this problem,this thesis expands it to support the slice.It is expected that while achieving determinism,it can also meet the service level agreement(SLA)proposed by slicing as much as possible.The key idea of the expansion is to introduce a warning counter in the design of the scheduling priority metric to measure the deviation of the index of each slice in the actual scheduling process from the requirements proposed,thereby affecting the scheduling decision,and pull the corresponding index back into the range of requirements set by the slice.Finally,the two algorithms proposed in this thesis are simulated on the 5G-air-simulation simulation platform.The simulation results verify that the deterministic radio scheduling algorithm can indeed guarantee the delay requirements of some user flows and that the algorithm's slicing expansion can indeed meet the SLA requirements proposed by the slice.