We bring you another part about communication between the page and the server . This time we’ll look at loading web pages over the new HTTP/2 protocol, which will hopefully soon replace the 15-year-old HTTP protocol.
The communication between the website and the server begins as soon as the browser accesses the source code of the page and requests its external parts (fonts, scripts, images, font sets, etc.). This process is notorious, and there would be no reason to delve into it had it not been for the news that the IETF has approv the final version of the new HTTP protocol at the time of writing this series. The significance of this news probably does not ne to be particularly emphasiz, since HTTP is one of the pillars of Internet communication, no website would be normally accessible without this protocol.
The news of the completion of HTTP
Come like a bolt from the blue, it could be said that it was eagerly await. It has been a long 15 years since the release of the last version, and it was clear that the old protocol was no longer suitable. Nowadays, many external components are attach to the text content of the page, which the server could send directly with the HTML, thus saving time and a heap of packets. However, HTTP/1.1 rules do not allow this, each file must be sent separately and on request. For each new request within an existing TCP connection, it is necessary to wait for the processing of the previous request, which causes a non-negligible delay. Browsers therefore establish up to six parallel connections, which in turn are not particularly enthusiastic about the
Solution by merging
Of course, everyone wants their website russia phone number data to load as quickly as possible, so
The realization of the intention does not always mean the achievement of the set goal. And in this case it really is a Pyrrhic victory. The ruction in the number of requests is reem in blood by the fact that no part of the page, thus mix into the HTML, can be cach, so that the complete volume of data is unnecessarily download for each page of the site; browsing such a website is therefore not very fast. However, it is still better to download a small number of large files than a bunch of small ones. In practice, therefore, small files are strategies are not mutually exclusive merg into larger ones – fonts are merg, scripts are merg, images can also be merg (so-call CSS sprites are creat ).
Workaround using prefetching
Recently, a translat article by Santiago Valdarrama was publish here, which discuss the so-call prictive browsing – the idea was simple: to have the next page ready before the bw lists user clicks on the link, even before he even thinks about such a thing.