以Google AMP为例的网页加速技术研究

Research on Web Page Acceleration Technology: A Case Study of Google AMP

关键词:  AMP  网页加载时间  Mahimahi  Selenium  流行度

Key words:  AMP  page loading time  Mahimahi  Selenium  popularity

摘要:

网页的加载速度与用户体验紧密相关，决定着用户的去留。本研究以Google AMP网页加速技术为例，对网页加速技术展开研究，并借助其关键技术，对腾讯新闻、CNBC新闻和VOX等网站进行改造，提出基于流行度分布和链路带宽的网页加载时间（Page Load Time，PLT）分析模型。结果表明，Google AMP能有效降低网页加载时间，相对于原版网页，最高可降低4.17%，即使在总体网络环境为最小往返时间（Minimum Round Trip Time，min_RTT）和丢包率较高、网络带宽较低的情况下，网页加载时间最高仍能降低92.46%；同时在启用流行度缓存机制的情况下，最高可再降低44%。本研究对进一步研究其他网页加速技术具有参考价值。

Abstract:

The loading speed of the web page is closely related to the user's experience, which determines the user's staying and leaving. In this study, Google AMP was taken as an example to study the page acceleration technology. With the help of its key technologies, Tencent News, CNBC News, VOX and other websites were reformed, and a Page Load Time (PLT) analysis model based on popularity distribution and link bandwidth was proposed. The results show that compared with the original webpage, Google AMP can effectively reduce the loading time of the webpage by up to 94.17%. Even with the overall network environment of high minimum RTT, high loss rate of the packet, and low network bandwidth, page load time can still be reduced by 92.46%. At the same time, when the popularity caching mechanism is enabled, AMP can be reduced by up to 44%. This study has a reference value for further research on other web acceleration technologies.