High-concurrency Car Network Distributed Database Load Balancing System Oriented To Environment Awareness

The vehicle network platform mainly acquires terminal environment-aware sensor data,including positioning,speed,and other road condition information,processes sensor data through big data technology and deep learning algorithm,analyzes terminal vehicle attitude and surrounding driving environment,and finally controls vehicles to avoid road obstacles.,reliable and safe to reach the destination.However,with the maturity of vehicle network technology,relevant vehicle companies will accelerate the deployment of vehicle terminals in a rhythm manner.At the same time,more and more vehicle terminals are successively connected to the vehicle networking platform,and various environmental sensors placed in the vehicle body collect vehicle road condition information.Vehicle attitude parameters,vehicle position,etc.,these huge sensor data are delivered to the vehicle networking platform in real time,causing the vehicle network platform system to load the blowout.If the load control capability of the platform is limited,the vehicle network server will be washed,resulting in the downtime of the vehicle networking platform.Interrupting customer service,and even causing terminal vehicles to lose control,posing a risk of car accident.This system takes the vehicle environment awareness and the high-speed concurrent interaction of the vehicle networking platform as the entrance,and studies the environmental parameters,position status,vehicle interior information,abstract environment-aware data type,data size and data priority involved in the driving process.Analyze the high concurrency technology of the network,explore the network design model,analyze the synchronous blocking model,the synchronous non-blocking model and the asynchronous non-blocking model,select the asynchronous non-blocking model as the optimal network design mode,and promote the efficient loading of vehicle sensor data into the vehicle network.platform.Analyze the stand-alone storage engine,do an in-depth study of the InnoDB data engine,MyISAM data engine and Redis storage engine,sort out the principle of the storage engine,design experiments to demonstrate the advantages of each engine,and use the Redis storage engine as the system database.The distributed database data distribution strategy is designed.The virtual node-based consistency hash model and the fragment-based consistency hash model are compared,and the fragmentation strategy is used to process the data distribution.The load balancing strategy is the research focus of the system,starting from three aspects: load evaluation,load balancing triggering and load movement.The load evaluation strategy studies the hardware resource usage of the cluster nodes,such as processor,memory,disk,and network throughput.The weight-based load evaluation algorithm is designed to calculate the current load status of the cluster nodes.In terms of load balancing triggering,the statistical data node receives the probability of the user request,studies the average load of the data unit in the time period,calculates the node load change rate,establishes the node load switching model,and estimates the future node load status.If the system threshold is exceeded,the load balancing is triggered.Device.In the load mobility strategy,the cluster heat data and cold data nodes are classified by the load evaluation algorithm,and the resource average maximization model is used to migrate the hot data to the idle nodes,and the cold data is distributed to the overload nodes to promote the even distribution of the hot data and the cold data in the cluster nodes..The load balancing strategy is efficient,stable and reliable,and solves the load pressure of distributed storage systems.It is practical and convenient for the vehicle network platform.It supports high-performance,high-concurrency database system to ensure the safety of vehicle network terminals,smooth driving,and bring high quality to customers.Driving experience.