Research On A Dynamic Heartbeat Mechanism That Maintains A Long-term Connection Of The Internet Of Things

With the development of the Internet of Things,how to implement a reliable long connection like PC in a low-performance,low-bandwidth embedded environment has become one of the research hotspots.The heartbeat mechanism is a widely used method for maintaining long connection between two parties.Dynamically settingsssss the keep-alive interval is particularly important in realizing reliable long connections with as little resources consumption as possible.In this paper,in order to overcome the poor flexibility and serious waste of resources in the traditional heartbeat mechanism.It attempts to use BP neural network prediction model based on teaching-learning-based optimization algorithms(TLBO)to predict keep-alive interval,which is suitable for the current network conditions and adapted to the dynamic network.The improvement of the TLBO algorithm is firstly to introduce the idea of teaching students their aptitude in the teacher phase.At this phase,to increase the learning level of the students themselves,and to adjust the position replacement formula.Secondly,before starting the learner phase,grouping the groups.It can enhance the local search ability of the algorithm,and avoid the loss of key features due to too close to the optimal individual.Finally,different learning strategies are adopted according to the level of learning objects encountered in the learner phase.So that the current solution always evolves toward the optimal direction,avoiding comparison.The BP neural network model based on the improved TLBO algorithm,is able to predict appropriate keep-alive interval,according to the changing network parameters in the Internet of Things environment.Making the heartbeat mechanism adjust to the dynamic network can improve the flexibility and adaptability of the heartbeat mechanism.Simultaneously,it reduces the flow and time loss by traditional algorithms in finding the optimal heartbeat.The experimental results show that the dynamic heartbeat mechanism implemented by the BP neural network prediction model based on the improved TLBO algorithm has obvious advantages in the number of iterations,flow and time loss compared with the adaptive heartbeat mechanism using the traditional algorithm.