Between reality and cyberspace
SPDY - a protocol to make the web faster?

Today I came across Google SPDY (via @powtac), an experimental protocol for faster datatransfer on the web, created in the context of Google’s “Let’s make the web faster”.

The protocol

Google looked into Protocols before (see their research regarding TCP Slow Start here) and now tries to tackle some problems HTTP has with today’s kind of web requests.

The goals

The goals alltogether sound like this could really be adopted quickly and give especially the - still a bit slow - mobile internet a boost. The key goals are:

  • To target a 50% reduction in page load time. Our preliminary results have come close to this target (see below).
  • To minimize deployment complexity. SPDY uses TCP as the underlying transport layer, so requires no changes to existing networking infrastructure.  
  • To avoid the need for any changes to content by website authors. The only changes required to support SPDY are in the client user agent and web server applications.
  • To bring together like-minded parties interested in exploring protocols as a way of solving the latency problem. We hope to develop this new protocol in partnership with the open-source community and industry specialists.
The 50% boost in terms of speed has to be evaluated “in the wild” (the blogpost says something about a peak gain of 64% under lab conditions - the results stated in the blogpost vary from 27,93% to 63.53%),
but as its part of the goals to not require changes on the content- or network-side, I guess (and hope) that the protocol will not only turn out to be as effective as stated in the blog, but also become widely-adopted quickly.

As of now
Indicators for a pretty quick adoption and a low bareer of implementation might be, that Firefox 11 and Chromium already have experimental support for SPDY.

This, of course, depends a bit upon adoption on the server-side, too.
But also there the hurdles seem not to big, as there is an Apache-Module and an experimental stand-alone SPDY-Webserver already plus a bunch of other stuff:
Alltogether this seems very interesting and if the reality backs the results Google found in their experimental setup, we can hope for overcoming the HTTP bottleneck.
I wonder why this hasn’t caught my attention much earlier?