The state and rate of HTTP/2 adoption

Tuesday, March 31st, 2015

http2 logoThe protocol HTTP/2 as defined in the draft-17 was approved by the IESG and is being implemented and deployed widely on the Internet today, even before it has turned up as an actual RFC. Back in February, already upwards 5% or maybe even more of the web traffic was using HTTP/2.

My prediction: We’ll see >10% usage by the end of the year, possibly as much as 20-30% a little depending on how fast some of the major and most popular platforms will switch (Facebook, Instagram, Tumblr, Yahoo and others). In 2016 we might see HTTP/2 serve a majority of all HTTP requests – done by browsers at least.

Counted how? Yeah the second I mention a rate I know you guys will start throwing me hard questions like exactly what do I mean. What is Internet and how would I count this? Let me express it loosely: the share of HTTP requests (by volume of requests, not by bandwidth of data and not just counting browsers). I don’t know how to measure it and we can debate the numbers in December and I guess we can all end up being right depending on what we think is the right way to count!

Who am I to tell? I’m just a person deeply interested in protocols and HTTP/2, so I’ve been involved in the HTTP work group for years and I also work on several HTTP/2 implementations. You can guess as well as I, but this just happens to be my blog!

The HTTP/2 Implementations wiki page currently lists 36 different implementations. Let’s take a closer look at the current situation and prospects in some areas.


Firefox and Chome have solid support since a while back. Just use a recent version and you’re good.

Internet Explorer has been shown in a tech preview that spoke HTTP/2 fine. So, run that or wait for it to ship in a public version soon.

There are no news about this from Apple regarding support in Safari. Give up on them and switch over to a browser that keeps up!

Other browsers? Ask them what they do, or replace them with a browser that supports HTTP/2 already.

My estimate: By the end of 2015 the leading browsers with a market share way over 50% combined will support HTTP/2.

Server software

Apache HTTPd is still the most popular web server software on the planet. mod_h2 is a recent module for it that can speak HTTP/2 – still in “alpha” state. Give it time and help out in other ways and it will pay off.

Nginx has told the world they’ll ship HTTP/2 support by the end of 2015.

IIS was showing off HTTP/2 in the Windows 10 tech preview.

H2O is a newcomer on the market with focus on performance and they ship with HTTP/2 support since a while back already.

nghttp2 offers a HTTP/2 => HTTP/1.1 proxy (and lots more) to front your old server with and can then help you deploy HTTP/2 at once.

Apache Traffic Server supports HTTP/2 fine. Will show up in a release soon.

Also, netty, jetty and others are already on board.

HTTPS initiatives like Let’s Encrypt, helps to make it even easier to deploy and run HTTPS on your own sites which will smooth the way for HTTP/2 deployments on smaller sites as well. Getting sites onto the TLS train will remain a hurdle and will be perhaps the single biggest obstacle to get even more adoption.

My estimate: By the end of 2015 the leading HTTP server products with a market share of more than 80% of the server market will support HTTP/2.


Squid works on HTTP/2 support.

HAproxy? I haven’t gotten a straight answer from that team, but Willy Tarreau has been actively participating in the HTTP/2 work all the time so I expect them to have work in progress.

While very critical to the protocol, PHK of the Varnish project has said that Varnish will support it if it gets traction.

My estimate: By the end of 2015, the leading proxy software projects will start to have or are already shipping HTTP/2 support.


Google (including Youtube and other sites in the Google family) and Twitter have ran HTTP/2 enabled for months already.

Lots of existing services offer SPDY today and I would imagine most of them are considering and pondering on how to switch to HTTP/2 as Chrome has already announced them going to drop SPDY during 2016 and Firefox will also abandon SPDY at some point.

My estimate: By the end of 2015 lots of the top sites of the world will be serving HTTP/2 or will be working on doing it.

Content Delivery Networks

Akamai plans to ship HTTP/2 by the end of the year. Cloudflare have stated that they “will support HTTP/2 once NGINX with it becomes available“.

Amazon has not given any response publicly that I can find for when they will support HTTP/2 on their services.

Not a totally bright situation but I also believe (or hope) that as soon as one or two of the bigger CDN players start to offer HTTP/2 the others might feel a bigger pressure to follow suit.

Non-browser clients

curl and libcurl support HTTP/2 since months back, and the HTTP/2 implementations page lists available implementations for just about all major languages now. Like node-http2 for javascript, http2-perl, http2 for Go, Hyper for Python, OkHttp for Java, http-2 for Ruby and more. If you do HTTP today, you should be able to switch over to HTTP/2 relatively easy.


I’m sure I’ve forgotten a few obvious points but I might update this as we go as soon as my dear readers point out my faults and mistakes!

How long is HTTP/1.1 going to be around?

My estimate: HTTP 1.1 will be around for many years to come. There is going to be a double-digit percentage share of the existing sites on the Internet (and who knows how many that aren’t even accessible from the Internet) for the foreseeable future. For technical reasons, for philosophical reasons and for good old we’ll-never-touch-it-again reasons.

The survey

Finally, I asked friends on twitter, G+ and Facebook what they think the HTTP/2 share would be by the end of 2015 with the help of a little poll. This does of course not make it into any sound or statistically safe number but is still just a collection of what a set of random people guessed. A quick poll to get a rough feel. This is how the 64 responses I received were distributed:

http2 share at end of 2015

Evidently, if you take a median out of these results you can see that the middle point is between 5-10 and 10-15. I’ll make it easy and say that the poll showed a group estimate on 10%. Ten percent of the total HTTP traffic to be HTTP/2 at the end of 2015.

I didn’t vote here but I would’ve checked the 15-20 choice, thus a fair bit over the median but only slightly into the top quarter..

In plain numbers this was the distribution of the guesses:

0-5% 29.1% (19)
5-10% 21.8% (13)
10-15% 14.5% (10)
15-20% 10.9% (7)
20-25% 9.1% (6)
25-30% 3.6% (2)
30-40% 3.6% (3)
40-50% 3.6% (2)
more than 50% 3.6% (2)

Firefox and partial content

Monday, June 16th, 2014

Update: parts of the change mentioned in this blog post has subsequently been reverted since clearly I had a too positive view of the Internet.

Firefox BallOne of the first bugs that fell into my lap when I started working for Mozilla not a very long time ago, was bug 237623. Anyone involved in Mozilla knows a bug in that range is fairly old (we just recently passed one million filed bugs). This particular bug was filed in March 2004 and there are (right now) 26 other bugs marked as duplicates of this. Today, the fix for this problem has landed.

The core of the problem is that when a HTTP server sends contents back to a client, it can send a header along indicating the size of the data in the response. The header is called “Content-Length:”. If the connection gets broken during transfer for whatever reason and the browser hasn’t received as much data as was initially claimed to be delivered, that’s a very good hint that something is wrong and the transfer was incomplete.

The perhaps most annoying way this could be seen is when you download a huge DVD image or something and for some reason the connection gets cut off after only a short time, way before the entire file is downloaded, but Firefox just silently accept that as the end of the transfer and think everything was fine and dandy.

What complicates the issue is the eternal problem: not everything abides to the protocol. This said, if there are frequent violators of the protocol we can’t strictly fail on each case of problem we detect but we must instead do our best to handle it anyway.

Is Content-Length a frequently violated HTTP response header?

Let’s see…

  1. Back in the HTTP 1.0 days, the Content-Length header was not very important as the connection was mostly shut down after each response anyway. Alas, clients/browsers would swiftly learn to just wait for the disconnect anyway.
  2. Back in the old days, there were cases of problems with “large files” (files larger than 2 or 4GB) which every now and then caused the Content-Length: header to turn into negative or otherwise confused values when it wrapped. That’s not really happening these days anymore.
  3. With HTTP 1.1 and its persuasive use of persistent connections it is important to get the size right, as otherwise the chain of requests get messed up and we end up with tears and sad faces
  4. In curl’s HTTP parser we’ve always been strictly abiding to this header and we’ve bailed out hard on mismatches. This is a very rare error for users to get and based on this (admittedly unscientific data) I believe that there is not a widespread use of servers sending bad Content-Length headers.
  5. It seems Chrome at least in some aspects is already much more strict about this header.

My fix for this problem takes a slightly careful approach and only enforces the strictness for HTTP 1.1 or later servers. But then as a bonus, it has grown to also signal failure if a chunked encoded transfer ends without the ending trailer or if a SPDY or http2 transfer gets prematurely stopped.

This is basically a 6-line patch at its core. The rest is fixing up old test cases, added new tests etc.

As a counter-point, Eric Lawrence apparently worked on adding stricter checks in IE9 three years ago as he wrote about in Content-Length in the Real World. They apparently subsequently added the check again in IE10 which seems to have caused some problems for them. It remains to be seen how this change affects Firefox users out in the real world. I believe it’ll be fine.

This patch also introduces the error code for a few other similar network situations when the connection is closed prematurely and we know there are outstanding data that never arrived, and I got the opportunity to improve how Firefox behaves when downloading an image and it gets an error before the complete image has been transferred. Previously (when a partial transfer wasn’t an error), it would always throw away the image on an error and instead show the “image not found” picture. That really doesn’t make sense I believe, as a partial image is better than that default one – especially when a large portion of the image has been downloaded already.

Follow-up effects

Other effects of this change that possibly might be discovered and cause some new fun reports: prematurely cut off transfers of javascript or CSS will discard the entire javascript/CSS file. Previously the partial file would be used.

Of course, I doubt that these are the files that are as commonly cut off as many other file types but still on a very slow and bad connection it may still happen and the new behavior will make Firefox act as if the file wasn’t loaded at all, instead of previously when it would happily used the portions of the files that it had actually received. Partial CSS and partial javascript of course could lead to some “fun” effects of brokenness.

Testing 2-digit year numbers in cookies

Wednesday, September 1st, 2010

In the current work of the IETF http-state working group, we’re documenting how cookies work. The question came up how browsers and clients treat years in ‘expires’ strings if the year is only specified with two digits. And more precisely, is 69 in the future or in the past?

I decided to figure that out. I setup a little CGI that can be used to check what your browser thinks:

It sends a single cookie header that looks like:

Set-Cookie: testme=yesyes; expires=Wed Sep  1 22:01:55 69;

The CGI script looks like this:

print "Content-Type: text/plain\n";
print "Set-Cookie: testme=yesyes; expires=Wed Sep  1 22:01:55 69;\n";
print "\nempty?\n";
print $ENV{'HTTP_COOKIE'};

You see that it prints the Cookie: header, so if you reload that URL you should see “testme=yesyes” being output if the cookie is still there. If the cookie is still there, your browser of choice treats the date above as a date in the future.

So, what browsers think 69 is in the future and what think 69 is in the past? Feel free to try out more browsers and tell me the results, this is the list we have so far:


Firefox v3 and v4 (year 2069)
curl (year 2038)
IE 7 (year 2069)
Opera (year 2036)
Konqueror 4.5


Chrome (both v4 and v5)
Gnome Epiphany-Webkit

Thanks to my friends in #rockbox-community that helped me out!

(this info was originally posted to the httpstate mailing list)

Beyond just “69″

(this section was added after my first post)

After having done the above basic tests, I proceeded and wrote a slightly more involved test that sets 100 cookies in this format:

Set-Cookie: test$yy=set; expires=Wed Oct  1 22:01:55 $yy;

When the user reloads this page, the page prints all “test$yy” cookies that get sent to the server. The results with the various browsers is very interesting. These are the ranges different browsers think are future:

  • Firefox: 21 – 69 (Safari and Fennec and MicroB on n900) [*]
  • Chrome: 10 – 68
  • Konqueror: 00 – 99 (and IE3, Links, Netsurf, Voyager)
  • curl: 10 – 70
  • Opera: 41 – 69 (and Opera Mobile) [*]
  • IE8: 31 – 79 (and slimbrowser)
  • IE4: 61 – 79 (and IE5, IE6)
  • Midori: 10 – 69 (and IBrowse)
  • w3m: 10 – 37
  • AWeb: 10 – 77
  • Nokia 6300: [none]

[*] = Firefox has a default limit of 50 cookies per host which is the explanation to this funny range. When I changed the config ‘network.cookie.maxPerHost’ to 200 instead (thanks to Dan Witte), I got the more sensible and expected range 10 – 69. Opera has the similar thing, it has a limit of 30 cookies by default which explains the 41-69 limit in this case. It would otherwise get 10-69 as well. (thanks to Stanislaw Adrabinski). I guess that the IE8 range is similarly restricted due to it using a limit of 50 cookies per host and an epoch at 1980.

I couldn’t help myself from trying to parse what this means. The ranges can roughly be summarized like this:

0-9: mostly in the past
10-20: almost always in the future except Firefox
21-30: even more likely to be in the future except IE8
31-37: everyone but opera thinks this is the future
38-40: now w3m and opera think this is the past
41-68: everyone but w3m thinks this is the future
69: Chrome and w3m say past
70: curl, IE8, Konqueror say future
71-79: IE8 and Konqueror say future, everyone else say past
80-99: Konqueror say future, everyone else say past

How to test a browser near you:

  1. goto
  2. reload once
  3. the numbers shown on the screen is the year numbers the browsers consider
    to be the future as described above

This is the type and I mean it

Wednesday, July 2nd, 2008

So someone pointed out this IEBlog entry for me, and I find it so hilarious I felt a need to share the fun. See the “MIME-Handling: Sniffing Opt-Out” paragraph towards the end.

Apparently Internet Explorer 7 and earlier just don’t care much for the Content-Type: header that servers reply, but they instead scan the body and guess what type it is. Thus “knowing better” than the content provider what content it truly is.

So in IE8 they’re (according to that blog entry) introducing a new attribute to the Content-Type header. If the site also sets “authoritative=true” it means it really means the type and the browser will then actually believe the site. I can’t stop giggling.

And yeah, some of the other craziness on that page is also good reading and they truly make you wonder what they are smoking during their brain storm meetings.