Slashdot Mirror
HTTP/2 Finalized
An anonymous reader writes: Mark Nottingham, chair of the IETF HTTP working group, has announced that the HTTP/2 specification is done. It's on its way to the RFC Editor, along with the HPACK specification, where it'll be cleaned up and published. "The new standard brings a number of benefits to one of the Web's core technologies, such as faster page loads, longer-lived connections, more items arriving sooner and server push. HTTP/2 uses the same HTTP APIs that developers are familiar with, but offers a number of new features they can adopt. One notable change is that HTTP requests will be 'cheaper' to make. ... With HTTP/2, a new multiplexing feature allows lots of requests to be delivered at the same time, so the page load isn't blocked." Here's the HTTP/2 FAQ, and we recently talked about some common criticisms of the spec.
MS will only support this in Spartan on Windows 10. Which means 7 and XP are already out of support.
http://saveie6.com/
The post even has counter-arguments and links to stuff. Are the 90's back again? Do I hear rave music, modem hand-shake tones and browser wars?
I'm really going to miss being able to telnet to a server and troubleshoot using plain text. Feels like a lot of simple has disappeared from the internet
Peace, or Not?
Web pages being slow has nothing to do with HTTP and all to do with bloat eating every bit of performance that is available to the marketers who are in charge of the web now. A binary protocol to save a couple of bytes in times where every web page loads at least three "analytics" scripts weighing dozens of kbytes each. Marketing rots the brain.
As the author of an open source webserver, I must say that I'm not really happy with HTTP/2. It adds a lot of extra complexity to the server side of the protocol. And all sorts of ugly and nasty things in HTTP/1 (too much work to go into that right now) have not been fixed.
What I have experienced is that SPDY (and therefor also HTTP/2) will only offer more speed if you are Google or are like Google. Multiplexing doesn't offer that much speed increase as some people would like you to believe. Often, the content of a website is located on multiple systems (pictures, advertisements, etc), which still requires that the browser uses more than one connection, even with HTTP/2. Also, HTTP/1 already allows a browser to send multiple requests without waiting for the response of the previous request. This is called request pipelining, but is turned off by default in most browsers. What I also often see is that a browser makes a first request (often for a CGI script) and the following requests (for the images, JS, CSS, etc) are never made due to browser caching. So, to me HTTP/2 adds a lot of complexity with almost no benefits in return.
Then why do we have HTTP/2? Well, because it's good for Google. They have all the content for their websites on their own servers. Because IETF failed to come up with a HTTP/2 proposal, a commercial company (Google in this case) used that to take control. HTTP/2 is in fact a protocol by Google, for Google.
In my experience, you are far better off with smart caching. With that, you will be able to get far better speed-increase results than HTTP/2 will ever offer. Specially if you use a framework that communicates directly with the webserver about this (like I did with my PHP framework). You will be able to get hundreds to thousands requests per second for a CGI script instead of a few tens of requests. This is a speed increase that HTTP/2 will never offer.
I think this is a failed change to do it right. HTTP is just like SMTP and FTP one of those ancient protocols. In the last 20 years, a lot has changed. HTTP/1 worked fine for those years. But for where the internet is headed, we need something new. Something completely new and not a HTTP/1 patch.
It doesn't have to be like this. All we need to do is make sure we keep talking.
"HTTP/2... also introduces unsolicited push of representations from servers to clients."
Seriously? Do we need yet ANOTHER way for a server to push unwanted code and malware onto our client systems? This is the greatest gift we could POSSIBLY give to the cybercriminals who want to break into our systems.
How about we think of security somewhere in this process, instead of pretending it's someone elses's problem?
Sometimes the "writing on the wall" is blood spatter...
The main problem isn't the size of the stuff that gets loaded. What's dozens of kb these days? Even a megabyte takes a tenth of a second to transfer on my connection. The problem is latency: it takes more than 100ms for that megabyte of data to even start flowing, let alone get up to speed. That's what the multiplexing is intended to deal with.
That said, the root cause of all this is the sheer amount of unnecessary stuff on pages these days. Fancy multiplexing or not, no request can finish faster than the one that's never made.
While we're discussing HTTP, what's wrong with Slashdot lately?
I keep getting timeouts, signed out, nothing but the front page, everything else loaded from a CDN (that just gives me browser security warnings because it's not coming up as slashdot.org), etc.
Are you guys under attack, or just unable to keep a site with ~100 comments per article up?
So you don't having something like HTTP_REFERER again
It's a bandwidth vs latency issue. HTTP1/1.1 is more latency sensitive, HTTP2 helps high latency. You say that pages load slowly because they're so large, yet these large bloated web pages consume nearly no bandwidth. HTTP2 multiplexing allows async requests and preemptive pushing of data, which should allow better usage of bandwidth. Fewer connections will also allow for quicker TCP relieve window ramp up and reduce the thundering hurd issue that many connections over a congested link creates.
Existing standards that are "good enough" tend to be hard to replace.
"It'll" is a contraction, not slang, you fucking nitwit.
What's dozens of kb these days? Even a megabyte takes a tenth of a second to transfer on my connection.
Which means you could burn through a 5 GB/mo cap in less than 9 minutes.[1]
Cellular ISPs in the United States charge a cent or more for each megabyte. Satellite and rural DSL ISPs tend to charge half a cent per megabyte.[2] This means cost-conscious users will have to either install ad blockers or just do without sites that put this "sheer amount of unnecessary stuff on pages".
[1] 5000 MB/mo * 0.1 s/MB * 1 min/60 s = 8.3 min/mo
[2] Sources: major cell carriers' web sites; Exede.com; Slashdot/a
Natalie Portman, get your grits ready!
Most of that bloat you speak of is delivered via Javascript. A few weeks ago, I finally put my foot down and made my default browser Firefox with NoScript. I have a (very) small list of sites allowed to execute scripts, and most of the time I will browse a website in its broken state, since I can still see all the text and images. It even foils the news websites that use Javascript to "faux-paywall" their content behind canvases, as opposed to only sending part of the story content from the server (I'm still shaking my head as to who on the business side thought this was a good business stance since one can still read most of the articles for free, but that's another discussion entirely). But lo and behold, once I started a staunch NoScript policy, page loads completed much faster, cookie sprawl was reduced, and Firefox's memory usage stayed relatively low. I also started to learn which scripts from which servers were truly allowed for things like externally-served comment systems (disqus, etc.), and also noticed the way that some webpages end up triggering a cascading dependency of server connections due to scripts calling scripts on other servers, etc.
One of the worst instances of cascading Javascript sprawl that I've seen was a page from The Verge, with 33 domains allowed (including theverge.com) executing 133 scripts. The SBNation "eulogy for RadioShack" article had the "script counter" go over 160. Oh, and that's leaving out web fonts, which NoScript also blocks (which also reveals how often some websites use custom fonts to draw vector-based icons; you can see the Unicode codes for each character since the font isn't loaded). Vox seems to love abusing Javascript in their designs; it's most of the reason why I've abandoned reading Polygon (the other reasons being the banal editorial content, aside from a few notable articles). In comparison, Slashdot is currently asking for 21 scripts, but is running perfectly fine without any Javascript enabled (despite nagging here and there).
I've ended up moving YouTube browsing to Pale Moon, since it natively supports the HTML5 player without issues. I may end up moving all of my browsing to Pale Moon with NoScript, since it natively supports a status bar.
The whole situation with the Javascript bloat reminds me of the scene from Spaceballs where Lonestar tells Princess Vespa, "Take ONLY what you NEED to SURVIVE." We're stuck with a bunch of prima donna web designers who want to duct tape on more adverspamming and social spying avenues to their website, and not standing back and taking a look at how bad it's impacting the user experience, not to mention the bloat from the hundreds of extra scripts and objects loaded by the browser, as well as the tens of connections to third-party servers.
"Ancillary does not mean you get to rule the world." --U.S. Circuit Judge Harry Edwards, speaking to the FCC's lawyer
Which is how we'll know who to put up against the wall and shoot.
...to pay your $699 licensing fee you cock-smoking teabaggers.
Celebrity worship is a poor substitute for Deity worship and costs more to boot.
No... maybe. It depends.
Amdahl's law is in full force here. There comes a point where increasing the bandwidth of an internet connection doesn't make pages load faster, because the page load time is dominated by the time spent setting up connections and requests (i.e. the latency). Each TCP connection needs to do a TCP handshake (one round trip), and then each HTTP request adds another round trip. Also, all new connections need to go through TCP window scaling, which means the connection will be slow for a few more round trips. Keep-alive connections help a bit by keeping TCP connections alive, but 74% of HTTP connections only handle a single transaction, so they don't help a great deal.
Oh! by the way, not everybody's connection is like yours, specially over mobile networks.
Mobile networks (and, yes, satellite) tend to have high latency, so round-trips are even more of the problem there. Also... when people shop for internet connections, they tend to concentrate on the megabits, and not give a damn about any other quality metrics. So that's what ISPs tend to concentrate on too. You'll see them announce 5x faster speeds, XYZ megabits!!, yet they don't even monitor latency on their lines. And even if your ISP had 0ms latency, there's still the latency from them to the final server (Amdahl's law rearing its ugly head again).
Given all that, I think I'm justified in saying that the main problem with page loading times isn't the amount of data but the number of round-trips required to fetch it. Reducing the amount of data is less important than reducing the number of, or impact of, the round-trips involved. And that's the main problem that HTTP/2 is trying to address with its fancy binary multiplexing.
(Now, if your connection is a 56k modem with 2ms latency, then feel free to ignore me. HTTP/2 isn't going to help you much.)