It will still be more than two years before we can expect to see the final revision of HTTP 2.0, the successor to the HTTP 1.1 standard which dictates how virtually every application on the web communicates today. Given how much the web has changed since HTTP 1.1 became the law of the land in 1999, it’s long been clear that a major revision would be necessary. A number of organizations, including Microsoft and Google, submitted proposals for this update over the last few months. Last week, the IETF HTTPbis Working Group responsible for this new version met in Vancouver, Canada to discuss the proposed changes and start the process of defining HTTP 2.0. Judging from the state of these current discussions, HTTP 2.0 will adopt Google’s increasingly popular SPDY protocol as the basis for its own standard.
The new standard is, of course, still a moving target, but there now seems to be consensus about some of the areas the working group will focus on, including compression, multiplexing, mandatory TLS, client pull/server push, flow control and WebSockets. The current HTTP standard has served the web well over the years, bit HTTP’s one-request-at-a-time model means that all those small objects that now make up many modern sites can’t always be fetched in parallel. This means HTTP traffic is also often slowed down because latency has increased now that we often rely on mobile networks. Despite the fact that we now have more bandwidth than ever before, some recent research from Google has shown that latency holds back even the fastest network connections when it comes to surfing the web.
Despite the fact that SPDY will form the basis for the HTTP 2.0 standard, though, the working group’s chair Mark Nottingham stressed that “it’s important to understand that SPDY isn’t being adopted as HTTP/2.0; rather, that it’s the starting point of our discussion, to avoid a laborious start from scratch.”
Microsoft also proposed its own set of changes – the “HTTP Speed+Mobility” proposal – earlier this year. Even Microsoft uses SPDY as the basis for its work, but also takes into account the recent work done around WebSockets. The other main departures from SPDY, said Microsoft’s Henrik Frystyk Nielsen, Gabriel Montenegro and Rob Trace yesterday, “are to address the needs of mobile devices and applications.”
What’s interesting here is that Microsoft’s team notes how the approach to defining the new standard has to be data driven and how its own research shows that SPDY often isn’t any faster than HTTP 1.1 with “all the known optimizations” (there is also some research that actually shows that SPDY can slow down some sites).
It’ll be interesting to watch how these different proposals will play out until the end of 2014 when the new standard is expected to be released. Given the changing nature of the web, HTTP is obviously due for an update and despite the fact that there seem to be some disagreements brewing in the background, the new standard will be a major boon for all of us who use the web every day.