Distributed architecture method applied to visual data pushing platform

Abstract

The invention discloses a distributed architecture method applied to a visual data pushing platform. The distributed architecture method comprises the following steps of: (1) carrying out basic deployment; (2) carrying out distributed deployment; (3) carrying out data organization by using a visual grouping method; (4) establishing a main control node, and arranging at least two data nodes under the main control node; (5) acquiring one data node with the minimum load to process; (6) after the operation is successful, establishing mapping between the data node and a sub-picture, and establishing mapping between the sub-picture and a client. According to the method disclosed by the invention, a front-end request is processed by a plurality of the data nodes, a concurrence pressure can be shared, and the waiting phenomenon due to concurrence is effectively alleviated; due to automatic expansion and disaster recovery capacities, a dynamic adjustment can be carried out by means of newly adding or closing the data nodes when a system load is changed, the whole adjustment does not need stop, a user cannot feel that, and the operation is safe and risk-free.

Description

technical field [0001] The present invention relates to the application field of data visualization, and specifically refers to a distributed architecture method applied to a visualization data push platform. Background technique [0002] In the application field of data visualization, the front-end for data display and the back-end push service for providing data mostly adopt common C / S structure, and data transmission is based on TCP / IP protocol. There are many performance bottlenecks in this method, which can be supported The display front-end is relatively limited, and there is a single point of failure. Once the performance reaches the limit during the application process, automatic expansion cannot be achieved, and the expansion can only be done through high-cost means such as upgrading hardware. Even if the asynchronous non-blocking multi-threaded IO mode like Netty is adopted, it cannot cope with the impact of large concurrency and large data transmission on the syst...

AI Technical Summary

Problems solved by technology

[0002] In the application field of data visualization, the front-end for data display and the back-end push service for providing data mostly adopt common C / S structure, and data transmission is based on TCP / IP protocol. There are many performance bottlenecks in this method, which can be supported The display front end is relatively limited, and there is a single point of failure. Once the performance reaches the limit during the application process, automatic expansion cannot be achieved, and the expansion can only be performed by high-cost means such as upgrading hardware.

Even if the asynchronous non-blocking multi-threaded IO mode like Netty is adopted, it cannot cope with the impact of large concurrency and large data transmission on the system due to factors such as hardware and network of single-node architecture.

[0003] At present, the use of asynchronous non-blocking multi-threading for data exchange can effectively improve the concurrency capability of the system, but the support capability of the system is ultimately limited by the throughput of the network and network card in the single-node mode. In a large concurrency scenario, the system response Severe slowdown and poor performance

At the same time, it is limited by the hardware resources of the server where it is located. If the performance cannot meet the needs of the business scenario, the vertical expansion method of upgrading the hardware can only be adopted. Adversely affect the user's use

Images