Research On Routing Optimization And Task Scheduling Algorithm Based On RapidIO Network

As a system internal interconnect technology,RapidIO is designed for low-latency,high-throughput,high-bandwidth,high-performance,and is designed for inter-board and inter-chip interconnect communication in high-performance embedded systems.RapidIO bus technology is applicable to various topologies and can be composed of switching devices to form communication networks of different sizes.However,RapidIO as a system interconnection technology will ultimately serve the embedded infrastructure.Therefore,it is necessary to study how to improve the performance of RapidIO network,and also need to study the computing system including RapidIO network module.With the deepening of the research on RapidIO technology,it is found that the following problems still exist:(1)RapidIO network is prone to repeated enumeration;(2)In the RapidIO network QoS process,the algorithm execution time is long;(3)Communication conflicts often occur during RapidIO network computing system task scheduling.Based on the above issues,firstly,this thesis starts from the characteristics of the RapidIO network structure.Secondly,mainly from three aspects of minimum isolation block mechanism to avoid repeated enumeration,routing strategy to improve routing efficiency and computing system task scheduling algorithm to avoid communication conflicts,this thesis constructs model of the problem and designs algorithm to optimize.Specifically,its main innovations are as follows:1.Aiming at the repeated enumeration problem of RapidIO routing network,a flow driven mechanism based on minimum isolation block was proposed.Firstly,the concept of minimum separation block was proposed,and the isolation block was divided into the routing network.Then,a depth first traversal search was carried out for the entire RapidIO routing network.Finally,the flow driven mechanism was introduced to compare the energy consumption value,power consumption value and the average operation time of the RapidIO switches before and after isolation,to verify the reliability of the algorithm through the flow driven energy algorithm and flow path allocation algorithm.The experimental results show that the number of enumerations of the RapidIO switch can be reduced by the isolation block,avoiding repeated enumeration,and the effect of this performance improvement is better as the number of routing structure layers is expanded.2.Aiming at the problem of QoS in RapidIO network,a RapidIO routing strategy based on improved glowworm swarm optimization algorithm was proposed.Firstly,gaussian mutation and storage mechanism were used to optimize the traditional firefly algorithm.gaussian mutation can effectively control the scattering degree of the solution in the search space of the algorithm,so that the algorithm avoids falling into a local optimum.The storage mechanism is conducive to evaluating and storing the historical state of each glowworm,preventing information loss.Then combine the improved glowworm swarm optimization algorithm with the actual RapidIO network QoS problem and select the final best routing strategy.The experimental results show that in the simulated RapidIO test network,the improved glowworm swarm optimization algorithm is more stable than other algorithm curves.It can find the optimal solution more quickly and show the best performance,playing a role in solving the QoS problem of RapidIO network.3.Aiming at the communication conflict in the task scheduling process of RapidIO network computing system,a cat swarm optimization task scheduling algorithm based on double arbitration mechanism and Taguchi orthogonal method is proposed.Firstly,the double arbitration mechanism is used to manage the task resources,and the task assignment is dynamically decided to effectively avoid communication conflicts.Then,the Taguchi Orthogonal Method is applied to the tracking mode of the cat swarm optimization process to reduce the algorithm running time.Experimental results show that the algorithm runs faster than other algorithms.The algorithm performs best in parallelism when dealing with a large number of tasks and has some advantages in heterogeneous environments.Finally,realize the full use of RapidIO network resources and computing resources,rational routing optimization and task scheduling,improve overall resource utilization,and better meet the needs of RapidIO in the development of embedded systems.