Design And Implementation Of Http Transport Optimization Prototype System

In recent years, the popularity of smartphones and rich mobile applications jointly promote the rapid development of mobile Internet, and bring data tsunami to the existing wireless network. The original HTTP/1.1protocol can no longer satisfy people’s needs due to the low efficiency of data transmission, caused by interaction mechanisms and message design. And the serial concurrent restriction of HTTP can easily lead to frequent reconnection at data transport layer. In addition, a large number of small-volume data transmission and other application behaviors may cause mobile network resource abuse. Therefore, a new network mechanism needs to be researched urgently to improve the HTTP protocol transmission efficiency, air interface resource utilization efficiency and the user experience.This paper gives the solutions of HTTP transport optimization prototype system. Based on HTTP/2.0protocol framework, we integrate Network-Friendly header compression algorithm to optimize the existing HTTP protocol. While maintaining the code transparency to the existing business server and client implementation, develop implementation of protocol conversion gateway server and client optimized protocol middleware. Between the access network and the wireless terminal, the system can provide connection multiplexing, traffic compression, priority scheduling and other transmission optimization for different HTTP applications on the same terminal. We establish HTTP traffic model to simulate user behaviors of typical mobile applications, and develop traffic generator and simulation server for simulation test under various mobile environment. Through data analysis and display module, we can get test results for different statistical indicators and visual presentation. Then the optimization effect of the system can be evaluated comprehensively and objectively.Experimental results show that for the five categories of typical mobile applications, this system has different levels of optimization effects for the four evaluating indicator:throughput, delay, loss rate and power consumption. Among them, application-layer latency’s optimal ratio can reach70%, indicating that the HTTP protocol transmission efficiency can be greatly improved after optimization. And according to the evaluation of the system optimization configuration, when the number of multiplexed TCP connections is3, the optimization effect of each application is better. And for header compression algorithm, NF algorithm’s performance is slightly better than Gzip algorithm, but there is not much difference.