基于云计算的航空飞行试验数据中心任务调度优化架构设计

Optimization architecture design of mission scheduling for aviation flight test data center based on cloud computing

摘要:

为解决航空飞行试验数据中心任务调度行为明显滞后的问题，实现对航空飞行试验数据的实时调度，设计基于云计算的航空飞行试验数据中心任务调度优化架构。设置WiRo中心网络，联合试验数据预测器与飞行任务分配器，完善中心任务调度优化架构体系的基础应用结构设计。根据PSO优化度量值的取值范围，求解惯性权重指标与粒子编码条件，并按照云计算法则，推导函数表达式条件，实现基于云计算的航空飞行试验数据调度模型的构建。在动态数据权限的约束下，计算中心调度任务的资源占用率与长尾延迟参数，实现对任务调度架构的优化配置，联合WiRo中心网络及EMU调度结构，完成基于云计算的航空飞行试验数据中心任务调度优化架构的设计。实验结果表明，云计算技术作用下，单位时间内的数据吞吐量达到了9.85B/s，由数据吞吐量有限造成的中心任务调度行为滞后的问题得到较好解决，符合实时调度航空飞行试验数据的实际应用需求。

Abstract:

In order to solve the problem that the task scheduling behavior of the aviation flight test data center is obviously lagging behind, and realize the real-time scheduling of the aviation flight test data, the optimization architecture of the aviation flight test data center task scheduling based on cloud computing is designed. Set up WiRo center network, joint test data predictor and flight task distributor, and improve the basic application architecture design of the center's mission scheduling optimization architecture. According to the value range of PSO optimization measures, the inertia weight index and particle coding conditions are solved, and the function expression conditions are derived according to the cloud computing rules to realize the construction of the aviation flight test data scheduling model based on cloud computing. Under the constraint of dynamic data permissions, the resource occupancy rate and long tail delay parameters of the scheduling tasks of the computing center can realize the optimal configuration of the task scheduling architecture, and complete the design of the task scheduling optimization architecture of the aviation flight test data center based on cloud computing in combination with the WiRo central network and EMU scheduling architecture. The experimental results show that the data throughput per unit time has reached 9.85B/s under the effect of cloud computing technology, and the problem of central task scheduling behavior lag caused by the limited data throughput has been well solved, which meets the practical application requirements of real-time scheduling aviation flight test data.