论文部分内容阅读
为提高因出现故障而导致网络设备资源不足时机载网络的可靠性,定义了两种机载网络传输性能降级约束,降级约束使用3个参数精确描述机载网络中消息实例传输成功或失败的数量及其分布。提出动态失效、关键函数和最小确定性将来序列等概念来对网络传输性能降级约束进行理论分析:通过计算关键函数值预先确定下一条消息实例传输失败是否会产生动态失效;使用最小确定性将来序列确定在不发生动态失效时将来消息实例传输成功的最少数量。给出两种实施约束的策略:消息源节点静态过滤和网络动态仲裁。提出一种适用于网络动态仲裁策略的调度算法:基于网络传输性能降级约束的双层优先级调度,该算法利用关键函数的预判性来避免动态失效;分析了该算法的可调度性条件。理论分析和仿真表明,当机载网络设备资源不足时,通过实施网络传输性能降级约束可以把网络性能降级的程度限制在可控范围内,从而增强机载网络的可靠性。
In order to improve the reliability of on-board network caused by the failure of network equipment resources, two kinds of degradation constraints on the performance of on-board network transmission are defined. The degradation constraint uses three parameters to describe the successful or unsuccessful transmission of message instances in airborne network Number and distribution. The concepts of dynamic failure, key functions and minimum deterministic future sequences are proposed to theoretically analyze the degradation of network transmission performance. By calculating the value of the key function, it is determined in advance whether the transmission failure of the next message instance will result in a dynamic failure or not. The minimum deterministic future sequence Determine the minimum number of successful message instances to transmit in the event of a dynamic failure. Two strategies for enforcing constraints are given: static filtering of source nodes and dynamic arbitration of networks. This paper presents a scheduling algorithm suitable for dynamic network arbitration strategy: two-level priority scheduling based on degradation of network transmission performance. The algorithm uses the predictive performance of key functions to avoid dynamic failure. The condition of schedulability of the algorithm is analyzed. Theoretical analysis and simulation show that when the resources of airborne network equipment are not enough, the degradation of network performance can be limited within the controllable range by implementing the network transmission performance degradation constraint, so as to enhance the reliability of airborne network.