How HTTP 2.0 Will Change The Face of The Internet

Almost everything you visit on the web at one point or another uses a special protocol known as the Hyptertext Transfer Protocol (HTTP). Ever since the year 1999, you’ve been using HTTP version 1.1. This has been the ongoing standard for many years until Google made an announcement on February 10, 2015 that its browser will be adding full support of what is now known as HTTP/2. This sounds like utter gibberish to some, but that’s because there’s no description of what HTTP/2 does differently. To understand this, we need to explore exactly what this new protocol version does, and how it is similar to the version of HTTP we’ve been using for nearly two decades.

Whenever a new protocol version is developed, it needs actual concrete goals. The most obvious goal is backward compatible with its predecessor, HTTP 1.1. Without that ability, every server in the world will have to switch to HTTP/2 for you to be able to browse their websites.

While maintaining compatibility with the older version, this new protocol will make use of advanced techniques as measures against latency, making pages load faster. This is the primary goal, the problem that HTTP/2 plans to address most aggressively.

Other improvements include added security and compatibility with reverse proxies.

In the big scheme of things, HTTP/2 is not going to be that much different from HTTP 1.1. As you surf the internet, the strongest effect you will feel is that webpages will load significantly faster as long as they support the new version.

http2-fiber

To say that “HTTP/2 makes everything faster” is a disservice to the amount of work that actually takes place behind the scenes to accomplish this. The HTTP 1.1 protocol is riddled with a series of issues that were acceptable in the earliest years of the 21st century but no longer make sense to continue to live with in a time where bandwidth is cheaper and servers are expected to load pages at much faster rates.

The chief way in which HTTP/2 plans to address page loading times is by compressing the header (a piece of data sent by your client to request that a server give you the data inside of a webpage you’re visiting). This minimizes the amount of time that your computer “shakes hands” with the destination server by reducing the amount of data that has to be sent. Nowadays, processors are powerful enough to handle millions of decompressions in a short amount of time. It makes more sense to do this now.

While the above will only take care of the latency in the initial request, there are also ways that HTTP/2 plans to take care of your entire interaction with a website. It will directly implement server push technologies, which allow servers to be more active in the communications process. Until recently, you had to send requests periodically to the server, making it interprets the headers you churn out every time you ask for information. With HTTP/2, the server will send you new data when it appears.

Lastly, HTTP/2 will do something called “multiplexing” when you send requests. In HTTP 1.1, there was a problem: Every new packet took precedence over the last one. All of them were processed in a linear fashion, leading to a problem called “head-of-line blocking”. Basically, a server’s performance was limited by the fact that it would have to process the first packet that comes to it while leaving the rest in a queue. If the packet took a long time to process, all the other packets had to wait in line for their turn. With HTTP/2, multiple packets will be processed at the same time.

With this combination of different “cures”, HTTP/2 will do everything it can to avoid slowdowns due to HTTP-specific problems. This will be particularly advantageous for websites with smaller servers that aren’t connected to as much bandwidth as the ones running Facebook and Google.

If you have questions or ideas, be sure to leave a comment with your thoughts!

6 comments

  1. I would had to your first paragraph the fact that not only are the pages expected to load pages at much faster rates, there is so much more media elements that need to load. Unlike when 1.1 was introduced webpages today have far more data to push to the client machine. Factor in the new HTML5 and CSS3 to the equation and we have much higher data streams that need to reach the end point.

    This is a welcome protocol and I wish the world wide web was not so slow to adopt the new tech that is making our lives easier. I am finding the adoption of IPv6 painfully slow and it is another tech that is needed badly on the web.

  2. While I can respect your enthusiasm on this matter. I would prefer to walk on the side of caution on this. IPV6 while indeed a needed feature of technology is not without it’s flaws regarding security, and if this version 2 of HTTP is still in its infancy, then we should make NO moves towards adopting it until it has been vetted, put through its paces and ensured that it is security hardened to the ultimate degree. Granted……there will be some whiz-kig-geniuis who may figure out a way to get around those security protocols, but to just throw this out there in it’s current state and “hope for the best”…would be the equivalent of following the pirate’s map to the huge pile of Fool’s Gold!

    • Eddie, there’s been a discussion in the HTTP/2 community about forcing TLS. The consensus was unclear, but security is a much higher priority with HTTP/2 than it was with HTTP 1.1.

      Your concern, nonetheless, remains valid.

  3. Many words but no information. Author doesn’t say the most important thing: how much faster the program will run.
    Author writes what the program ‘HTTP/2 will do’ but there is no download possibility.
    How ‘good’ or useful it will be, how much it will cost, when and where it will be (or is) available.
    Contemporary noise.

    • It’s a protocol whose draft was finished just yesterday. No pure implementation has been tested, but we are aware that it will load pages faster because we know what compression and push technology have done already. Making it a part of the overall HTTP ecosystem would make handshakes happen more quickly.

      As for cost, for the most part, it will be zero for the consumer. For web hosts, it may cost them a penny or two (to upgrade their CPanel and other stuff), but it will not impact them much anyway.

      The closest thing we have to a study on web loading speed is an implementation of a similar protocol (SPDY). Microsoft’s research and development division tested it (http://research.microsoft.com/apps/pubs/?id=170059) and found that SPDY does increase page loading speed, but smaller pages might load a bit slower.

      HTTP/2 is somewhat different from SPDY, so we don’t know exactly to what extent page load times will be impacted, but we know they’ll definitely be faster. By how much depends entirely on the web server and the connecting computer’s internet connection (because things like these are variables we cannot predict).

    • Lets go over this comment for how many things wrong on it.
      “Many words but no information. Author doesn’t say the most important thing: how much faster the program will run.”
      HTTP/2 isn’t a program. Its a protocol for internet browsing. We’ve been using the protocol HTTP 1.1 for over 15 years currently.

      “Author writes what the program ‘HTTP/2 will do’ but there is no download possibility.”
      Again, it isn’t a program and the author does state this at the beginning saying its about a protocol version not a program.

      “How ‘good’ or useful it will be, how much it will cost, when and where it will be (or is) available.”
      Well he states about how useful it will be by addressing what the protocol updates. It will cost nothing because its a protocol update not a program.

      While HTTP/2 is still going through some steps for final release, the when is addressed as it will start seeing acceptance now that its finalized in what it has, “Google made an announcement on February 10, 2015 that its browser will be adding full support of what is now known as HTTP/2” from the article.

Comments are closed.

Sponsored Stories