What is HTTP/2

The internet is in a constant process of de­vel­op­ment that is char­ac­ter­ised by advanced tech­no­lo­gies, new content, and growing user demands. The simpler days of simpler websites have now been replaced by complex web presences that are perfectly tailored to the visitor and their devices. Access time and data transfer play an ever-in­creas­ing role regarding the usability and speed at which modern websites display their content. It is therefore not sur­pris­ing that HTTP/2 is a modern, more advanced version of the old internet protocol, HTTP/1.1, and should decrease the loading time. 

When looked at from a technical per­spect­ive, every com­mu­nic­a­tion unit is comprised of two areas: the message header, also known as HTTP header, and the message body. The header delivers all relevant in­form­a­tion that needs to be in­ter­preted, such as the message body that contains the user data.

Just like the two corner­stones of the World Wide Web – URL and HTML – the de­vel­op­ment of HTTP began in the European Or­gan­iz­a­tion for Nuclear Research (CERN) in Switzer­land - this took place in 1989. The first published spe­cific­a­tion, RFC 1945 (Request For Comments), issued in May 1996, became commonly known as HTTP/1.0. This first version was replaced in June 1999 by HTTP/1.1, which is still in use today. The official successor HTTP/2 was launched in May 2015 as RFC 7540 and the plan is for it to re­volu­tion­ise data transfer online.

One of the main char­ac­ter­ist­ics of HTTP/1.1 has already been mentioned: in order to transfer different page elements like image, JavaS­cript, or CSS files to the client/browser, a separate TCP/IP con­nec­tion needs to be opened and closed again. Just like its role model, SPDY, HTTP/2 uses the multiplex procedure in order for the whole website to be loaded with just a single con­nec­tion. The server also has the ability to pass on pre­dict­able follow-up data to the client without them having to ask (server push). With HTTP/2, data packets from websites can be sent in order of their priority; for example, elements that are re­spons­ible for building the site would appear first. The HTTP/2 header is sent heavily com­pressed in com­par­is­on to HTTP/1.1 because un­ne­ces­sary in­form­a­tion is omitted. An ad­di­tion­al im­prove­ment is the use of binary code (in place of text files) when com­mu­nic­at­ing, which sim­pli­fies the process and makes it less error-prone. Here is a summary of all the HTTP/2 im­prove­ments:

• con­nec­tion between a browser (client) and a website
• binary-coded message exchange
• heavily com­pressed HTTP message header
• server push with fore­see­able requests (only for Apache and not for Nginx)
• pri­or­it­isa­tion of important site elements

The back­ground of the new HTTP/2  features is il­lus­trated in this video by Akami Tech­no­lo­gies:

Since the release of HTTP/2, a lot of work has been done. Numerous web browsers and servers meanwhile support the new internet protocol. Mozilla Firefox has even had the protocol im­ple­men­ted since version 36 (February 2015). All versions until now have been loaded with HTTP/2 over TLS, which is the encrypted variant of the transfer protocol.

The number of websites that already use the new transfer protocol is re­l­at­ively low compared to sup­port­ing browsers. Nev­er­the­less, the number increased within one year to almost 10% from 1.2% at the end of September 2016. In March 2017, at least 12.7% of all website were using HTTP/2. An overview of the current state of usage can be found here. So far website operators with very high viewing numbers, like Google and Twitter, have converted their web presence to HTTP/2. Con­sid­er­ing the ad­vant­ages and the strongly in­creas­ing support through all current browsers, it will only be a question of time until the majority of websites use HTTP/2.

This overview shows which browsers support HTTP/2:

*partially supported