The veri¯cation of latency time for distributed hard real-time systems is considered to be the most important performance metrics. It is important for safety critical applications and for performance evaluation of the transmission systems. The integration of a middleware in such application, can induce some communication delays. Therefore, we focused our interest to the computation of end to end latency. Some existing works, adopt optimization or scheduling methods in network computation.35,36 In our work, we propose the WCRT to take into account all possible delays. Latency is considered as the period of time since a message is written by a DataWriter until it is received by a DataReader. It consists of delay caused by signal processing, propagation, queuing and transmission. We detail the computation of latency period of a packet on its entire network path, and we verify its validity with DDS QoS:

对于分布式硬实时系统的延迟时间的验证被认为是最重要的性能指标。对于安全关键应用程序和传输系统的性能评估而言，它非常重要。在这种应用程序中集成中间件可能会引入一些通信延迟。因此，我们将重点放在计算端到端延迟上。一些现有的工作在网络计算中采用了优化或调度方法。在我们的工作中，我们提出了WCRT来考虑所有可能的延迟。延迟被认为是从消息被DataWriter写入到被DataReader接收的一段时间。它包括信号处理、传播、排队和传输引起的延迟。我们详细说明了数据包在整个网络路径上的延迟周期的计算，并通过DDS QoS验证其有效性。