What is the HTTP/2 Protocol? Overview and Examples
The HTTP/1.1 protocol has been around for quite some time, but in recent years more web services have switched to the new HTTP/2 protocol. This article will define HTTP and the histories of HTTP/1.1 to HTTP/2, and you can use the links below to discover key differences and advantages of HTTP/2.
- Why is HTTP important?
- The history of HTTP/1.1 to HTTP/2
- What is HTTP/2?
- HTTP/1.1 vs HTTP/2: Main differences and benefits
- Disadvantages of HTTP/2
- How to start using HTTP/2 protocol
- How you can check if HTTP/2 is working
Why is HTTP important?
HTTP (Hypertext Transfer Protocol) is the protocol that powers most of the web today. As you opened this page, your browser made an HTTP request towards a web server that returned the text that you are currently reading.
HTTP is also used for things that you do not see with the naked eye. When you post something on social media, send a message to one of your friends, or open your favorite app, an HTTP request is most likely made in the background by the application you are using. These requests are made towards web services that might store your social media posts, messages, or app data.
Without HTTP, the web would simply stop working.
The history of HTTP/1.1 to HTTP/2
In 1989, Sir Timothy John Berners-Lee invented the HTTP protocol on a NeXTcube workstation with a 25 MHz CPU and several MBs of RAM. The protocol worked on networks with port connection speeds of 10 Mbits. Today, however, we have dramatically faster CPUs and thousands of MBs of RAM, but the main WWW protocol is still only HTTP/1.1. It was last updated in 1992, so why are we still using it?
While other protocols have been updated over the years (FTP became SFTP; POP3 evolved to IMAP, and telnet became SSH), HTTP/1.1 has not changed and has developed many issues with speed and security, and user-friendliness as a result.
Google was the first to investigate issues with HTTP/1.1. At the time, they were spending millions of dollars a year to support their data centers, and the HTTP/1.1 protocol simply cost too much in terms of CPU resources and internet connection capacity. They developed SPDY as an experimental alternative to HTTP/1.1—a protocol designed for better security and improved page load times.
The HTTP Working Group of the Internet Engineering Task Force (IETF) investigated Google’s SPDY protocol and Microsoft’s equivalent when designing the next version of HTTP. Facebook recommended HTTP/2 to be based on SPDY in July 2012.
Based on the research done by Google, Microsoft, and Facebook, the IETF released the HTTP/2 protocol in 2015. This became the second major version of the most useful internet protocol, HTTP.
What is HTTP/2?
HTTP/2 is the second version of the HTTP protocol aiming to make applications faster, simpler, and more robust by improving many of the drawbacks of the first HTTP version.
The primary goals of HTTP/2 are:
- Enable request and response multiplexing
- Header compression
- Compatibility with the methods, status codes, URIs, and header fields defined by the HTTP/1.1 standard
- Optimized prioritization of requests, making sure that loading for optimal user experience is as fast as possible
- Support for server-side push
- Server-side backwards compatibility, making sure servers can still serve clients only supporting HTTP/1.1 without any changes
- Transforming to a binary protocol from the text-based HTTP/1.1
Some key features of HTTP/2 include:
- Binary: Meaning commands use 1s and 0s and not text
- Multiplex: Permits multiple requests and responses to be sent at the same time
- Compression: Compresses headers that have been requested previously to make things more efficient
- Stream prioritization: This allows for the exchange of successive streams at one time
- Server push: The server can send additional information needed for a request before it is requested
- Increased security: HTTP/2 is supported through encrypted connections
HTTP/1.1 vs. HTTP/2: Main differences
As mentioned, the aim when designing HTTP/2 was to fix many of the drawbacks experienced with HTTP/1.1. In this section, we go through the differences between HTTP/2 and its predecessor.
HTTP/2 offers built-in compression of the request headers (HPACK). Modern web applications usually accept a range of different headers, such as authorization, caching directives, and client information. While compression of these might not make much of a difference for a single request, there is a lot of data sent over the network to be saved when compressing them in high-traffic applications. HTTP/1.1 does not compress headers by default.
Many of the new features in HTTP/2 are aimed at improving performance for the end-user. One example of this is how external resources can be preemptively pushed to the client's browser before they are explicitly requested. HTTP/1.1 does not have these advanced features.
HTTP/2 is binary instead of textual as HTTP/1.1. In practice, this means simplified implementation of commands that previously could be mixed up due to optional whitespace when using the text format. Browsers that support HTTP/2 will convert textual commands into binary before sending them over the network.
Because of the binary format used by HTTP/2, there is no longer a risk with so-called response splitting attacks that are possible with HTTP/1.1. This enables an attacker to manipulate the response headers by injecting whitespace into a textual response. This is no longer possible with the binary format of HTTP/2.
While the HTTP/1.1 protocol delivers responses based on a single request, HTTP/2 uses multiplexing and server push features to increase the delivery performance.
The buffer is the space used by the client and server that holds the requests that have not yet been processed. In HTTP/1.1, the flow control used to manage the available buffer space is implemented at the transport layer. In HTTP/2, the client and server can implement their own flow controls to communicate the available buffer space.
HTTP/2 enables full request and response multiplexing. In practice, this means a connection made to a web server from your browser can be used to send multiple requests and receive multiple responses. This gets rid of a lot of the additional time that it takes to establish a new connection for each request. HTTP/1.1 does not support multiplexing.
Faster encrypted connections
HTTP/2 uses the new ALPN extension, which allows for faster-encrypted connections since the application protocol is determined during the initial connection. Using HTTP/1.1 without ALPN needs additional round trips for the encryption handshake.
No need for HTTP/1.1 workarounds
In order to bypass some of the drawbacks with HTTP/1.1, multiple workarounds have been invented. Two examples of these are:
Domain sharding is a common performance workaround used with HTTP/1.1 to trick browsers into opening more simultaneous connections than would normally be allowed.
These are no longer needed with the built-in multiplexing in HTTP/2.
Benefits of HTTP/2
What are some pros to using HTTP/2?
- HTTP/2 is binary instead of textual.
- HTTP/2 is fully multiplexed. This means that HTTP/2 can send multiple requests for data in parallel over a single TCP connection. This is the most advanced feature of the HTTP/2 protocol because it allows you to download web files via ASync mode from one server. Most modern browsers limit TCP connections to one server.
- It uses header compression HPACK to reduce overhead.
- It allows servers to “push” responses proactively into client caches instead of waiting for a new request for each resource.
- It uses the new ALPN extension, which allows for faster-encrypted connections since the application protocol is determined during the initial connection.
- It reduces additional round trip times (RTT), making your website load faster without any optimization.
Domain sharding and asset concatenation are no longer needed with HTTP/2.
- It’s widely supported by browsers. As a basic internet technology, protocol HTTP/2 must be supported by the current version of your browser to work well.
As you can see, most modern browsers support HTTP/2 well (with Opera Mini just out of range).
Disadvantages of HTTP/2
First of all, there really isn’t an available alternative today that’s superior to HTTP/2. Browsers refused to support SPDY in favor of HTTP/2, with only Mozilla Firefox still supporting SPDY. But, as an IT professional, you should still know the protocol’s weak points. Some experts believe that these issues may be fixed in the future with the release of the HTTP/3 protocol, but for now, these are a few of the cons.
It’s not very fast or modern.
Most experts around the world expected a lot of new features in the HTTP/2 protocol, but these were not included in the final version. The main reason is simple: HTTP/2 must maintain backwards compatibility with the old HTTP/1.1 by using the same POST and GET requests, codes of status (200, 301, 404, and 500), etc. Also, several new features like compression page headers are vulnerable to BREACH and CRIME attacks.
Encryption is not required.
Years ago, data encryption was mostly limited to financial transactions. But hacks and even government surveillance are more viable threats, with implications on your private life and business.
Encryption may guard you against those types of threats, but disputes inside the team of HTTP/2 developers led to a decision to “leave encryption as is in HTTP/1.1.” This means that owners of a website may still choose a lower level of security, potentially putting users at risk. Developers of browsers can “fix” this issue in a way—currently, they turn on HTTP/2 only if SSL/TLS protocols are available.
Cookie security is still an issue.
One of the biggest security failures of HTTP/1.1 is cookies—.txt files containing client data that may be included in an HTTP request to a server (website). By using these files, the server can identify the client and, for example, keep the client session open, so it doesn’t have to ask for a login and password. Some ad providers like Google can store information about the user in cookies (e.g., gender, age, interests, etc.). That information is then used to target personalized ads to that user.
This may not sound too bad, but the reality is that cookies may be stolen or tampered with by hackers. This means that hackers can access your email, social networks, and other websites that contain your personal data, even without passwords. This is called a cross-site scripting (XSS) attack.
Experts hoped that cookies would be replaced with new technologies, but so far, they have not.
How to start using HTTP/2 protocol
If you’re a client-side user, all you need to do is update your browser to the newest version. Most browsers support HTTP/2. If you’re a server-side user, the process of upgrading to HTTP/2 is a little more complicated.
Here’s a quick guide for the most popular server OS and webserver: Apache running Ubuntu 14.04. Other configurations may require more steps and must be done by skilled DevOps engineers (like Microsoft IIS). For Apache, HTTP/2 just requires installing a module that can be easily enabled for any website.
First, install Apache version 2.4.17 or higher on your server. The installation of the HTTP/2 module for Apache takes just three steps:
- Add repository: sudo add-apt-repository ppa:ondrej/apache2
- Now, update Apache2 from this repository to get the latest modules: sudo apt-get upgrade apache2
- Enable HTTP/2 protocol as a module for Apache:
--CODE language-markup line-numbers--
sudo a2enmod http2
Now we must enable HTTP/2 for your virtual host. Open a .conf file for your virtual host. You’ll see something like this one:
--CODE language-markup line-numbers--
Protocols h2 http/1.1
CustomLog "/var/log/httpd/one-access.log" combined
<directory var="" www="" html=""></directory>
Allow from all
Add this line somewhere in the VirtualHost section:
--CODE language-markup line-numbers--
Protocols h2 http/1.1
Save this, close the file, and don’t forget to restart the Apache2 service:
--CODE language-markup line-numbers--
service apache2 restart
If you’ve done all of this correctly, the HTTP/2 protocol will start working.
HTTP/2 and the Nginx server
Nginx is a popular and very fast web server that works especially well with static content (like .css, .js, or .jpg files). Google promised a 40 to 50 percent increase in speed with the SPDY (HTTP/2) protocol. But in 2014, the Russian company Yandex tested this by experimenting with the SPDY protocol for their Yandex Mail service along with the Nginx web server, and the result wasn’t too impressive.
Yandex was very interested in improving the download speed of static content (.js, .png, .css, etc.). The calculated percentile (80%) and median (average speed of download) show that the download speed was still relatively poor.
The file load time was reduced by just 0.6 percent (and they checked it twice). So, this shows that using Nginx with HTTP/2 is not effective because this webserver is already very fast. Also, Nginx prioritizes server resources: for an inactive HTTP 10,000 to maintain 10,000 connections, it only needs 2.5 MB of RAM. SPDY, as an early version of HTTP/2, was designed as a module for Apache that fixed specific issues, but currently, it’s not effective for other web servers.
What SPDY does do well on Nginx is to reduce the number of TCP connections to servers and save some cost (processing power) by using a smaller CPU load.
How you can check if HTTP/2 is working
Here’s a guide for Google Chrome users:
1. Open Developer Tools and click on the Network tab.
2. Now, right-click on the header of the table and select the Protocol column to add it.
3. Do not close the Developer Tools and refresh the page.
The third column now shows resources that are loaded via the HTTP/2 protocol.
Whether you decide to start using HTTP/2 or not is definitely your own choice as a website owner, but it’s undeniably the future. If you have questions on how to shift to HTTP/2, speak to a web developer via Upwork.