Just last week, Google announced plans to remove SPDY support from its open source Chromium project early next year, and it would be replaced by the not-yet-official HTTP/2 protocol. Today, the Internet Engineering Steering Group (IESG), the managing component of the Internet Engineering Task Force (IETF), announced that the HTTP/2 and HPACK specs have been formally approved and are on the way to becoming official standards.

For those who may not already know, HTTP/2 (spec) is a network protocol generally used by web browsers for transferring the HTML, images, and other resources that make up web pages – but it is frequently used by countless other types of apps for communication, as well. This will be the first update to the HTTP protocol since v1.1, which was finalized in 1999. Version 2.0 brings a number of advantages including data compression, lower latency, push, request and response multiplexing, prioritization, and many others.

HPACK (spec) was created to remove redundant header fields and compress data for HTTP/2 transmissions. This is meant to cut down on large, repetitive bundles of headers. In older versions of the HTTP spec, each request from a client (e.g. web browsers) required retransmission of each header field in plain text, which increased data usage and often resulted in lag.

There are a few final formalities remaining for the IETF –basically just editing and paperwork– but the spec itself is done and will not see any further functional changes. While drafts of the spec haven't gone through any major changes in months, this marks the stage where developers can confidently write code without concern that future versions of the protocol will introduce incompatibilities.

Most modern web browsers already provide complete implementations of HTTP/2 based on older versions of the draft. Quite a few web servers also offer partial support, and soon we'll see support rolling out that will make it easier for apps and web frameworks to take advantage of everything HTTP/2 has to offer. While the networking layer will remain completely invisible to regular users, the benefits will be significant.

Source: IETF, Mark Nottingham's Blog