The Design And Implementation Of LED Display Centralized Management Platform Based On Django Framework

As an important medium for the release of resource information,LED displays are widely used in transportation,commerce,advertising,finance and other fields.The centralized management system of the display device has changed from C/S(client/server)architecture and LAN management mode to B/S(browser/server)architecture and Internet management mode.With the increase in the number of outlets and the scale of equipment,it is necessary to design and develop an LED display centralized management Web platform to effectively solve the throughput bottleneck of gigabit(gigabit)bit-level uplink data during the peak period of equipment use and the database read and write pressure.In order to comprehensively research and analyze the existing solution system architecture in view of the above problems,this paper proposes to use Django framework to build a back-end server side,in cooperation with Kafka message middleware and distributed database middleware Mycat,to improve system performance indicators and simplify existing systems.Aiming at the huge amount of data generated during the peak usage period of the device,the system designed a message filtering and distribution strategy,a message re-investment strategy,and a synchronous flashing strategy based on Kafka message middleware to handle peak data,improve data throughput,and achieve high reliability and high Performance messaging.Aiming at the storage and query of massive historical data,the database based on the distributed database middleware Mycat,is split from the vertical and horizontal dimensions to improve the database storage capacity.The hash modulus fragmentation rule is instead by the consistent hash fragmentation rule to improve Cluster scalability.The database use master-slave node replication technology to ensure data consistency after read and write operations been separated;designing customized query sharding strategies for different data tables to reduce database response time.Use the kafa-perf tool to test the performance of the cluster,set the data volume to 5million,and change the number of threads,batches,and partitions to test the message producers and consumers.The producer throughput rate is 74MB/s,and the sending volume is up to More than 150,000 messages per second;consumer throughput is close to 120MB/s,and processing message capacity is more than 240,000 messages per second.Using A gradient test scheme to increase the number of concurrent read and write requests to test the impact on the maximum response time and average response time of the system.When the number of concurrent requests reaches 2000,the maximum response time for a system write request is 503 ms,the maximum response time for a query request is 557 ms,and the average response time is 461 ms.The results show that this system can not only complete the centralized management of the LED display screen,but also has a high system throughput,a short system maximum response time and an average response time in the case of large data volumes and high concurrent numbers.