Background: In order to obtain reliable cloud resources, reduce the impact of resource
node faults in cloud computing environment and reduce the fault time perceived by the application
layer, a task scheduling model based on reliability perception is proposed.
Methods: The model combines the two-parameter weibull distribution and analyzes various interaction
relations between parallel tasks to describe the local characteristics of the failure rules of resource
nodes and communication links in different periods. The model is added into the particle
swarm optimization (PSO) algorithm, and an adaptive inertial weighted PSO resource scheduling
algorithm based on reliability perception is obtained.
Results: Simulation results show that when A increases to 0.3, the average scheduling length of the
task increases rapidly. When it is 0.4-0.6, the growth rate is relatively slow. When greater than 0.8,
the average scheduling length increases sharply, it can be seen that the r-PSO algorithm proposed in
this paper can accurately estimate the relevant parameters of cloud resource failure rule, and the
generated resource scheduling scheme has better fitness, and the optimization effect is more significant
with the increase in the number of tasks.
Conclusion: With only a small amount of time added, the reliability of cloud services is greatly