- The difference between in-transit and end-to-end encryption
- Who has access to your information when not using an encryption
- Who has access to your information when using an in-transit encryption
- Who has access to your information when using an end-to-end encryption
The best way to protect your online communications is through encryption. But not all encryption is equally protective. We will focus on the concepts that distinguish between the degrees of protection.
Encrypted or Not
The most basic version of encrypted communications is in-transit encryption, where your information is encrypted between your computer and a server. In the context of browsing the web, this is the best you can do to protect the content (but not the metadata) of your communications from an adversary. Most web browsers indicate whether your browsing is encrypted by the URL, as illustrated below.
In the top example, the information is transmitted unencrypted. The full URL in this case is http://whenisgood.net, where http indicates accessing a web page without encryption. This browser (Firefox) emphasizes this point with a struck-through lock. In the bottom example, the information is encrypted: https indicates accessing a web page with encryption, and the s stands for “secure.” The keys used for this encryption are exchanged between your computer and the whenisgood servers using the Diffie-Hellman key exchange, as described in the chapter “Exchanging Keys for Encryption.”
Using http, every entity on the path between you and the website pictured below can access the content of your web browsing (such as the pictures being loaded and any information you might type into a web form). Further, anyone snooping on the communications between the entities on these paths (such as between Comcast’s network and the internet backbone) may also have access to your browsing content. We qualify this with may because the communications between two entities on this path may be encrypted. For example, the communications between a cell phone and a cell tower are encrypted in most cases.
By contrast, when you use https, only you and the website (technically, the servers that are hosting the website) have access to the content of your browsing. We specify content here because certain metadata would still be known by entities on the path between you and the website, such as the basic URL of the website, the amount of time you spend browsing the website, and the amount of information you are downloading from the website.
When we communicate with another person by email, instant messaging, or video chat, those communications are (most often) routed through the communication provider (e.g., Google servers for email or Microsoft servers for Skype calls), as pictured below. Nowadays, those communications are usually encrypted but most often only encrypted between you and the communication provider. That is, while the entities and eavesdroppers along the path between Assata and the communication provider’s servers (center) and between the communication provider’s servers and Bobby do not have access to the content of your communications (but can glean metadata), the communication provider does have access to the content of your communications.
We call this in-transit encryption because the content is encrypted while it is in transit between Assata and the communication provider and between the communication provider and Bobby. The in-transit encryption keys are generated separately for each part of the path between Assata and Bobby, pictured below. The provider (center) performs a Diffie-Hellman key exchange with Assata, generating a shared key, and performs a separate Diffie-Hellman key exchange with Bobby, generating a different shared key. When Assata sends a message to Bobby through the provider, the message is first encrypted with the key Assata shares with the provider, and then the message is transmitted to the provider. The provider decrypts the message with the key that Assata and the provider share. Then the provider re-encrypts the message with the key that the provider shares with Bobby before transmitting the encrypted message to Bobby. Bobby can then decrypt the message. Therefore, the message only exists in a decrypted state on Assata’s and Bobby’s devices and the provider’s servers; the message is encrypted when it is in transit between these entities.
While in-transit encryption protects your communications from many potential adversaries (such as your internet service provider, the Wi-Fi hotspot, or a snoop along the communication channels), the communication provider still has access to all that information. Even if the provider is not a direct adversary, they may share that information with an adversary (such as through a subpoena or warrant). End-to-end encryption (E2EE) will protect your communications from even the communication provider.
For E2EE (pictured below), Assata and Bobby exchange keys (using a Diffie-Hellman key exchange or a similar procedure). While their communications are routed through the communication provider, so long as the provider isn’t mounting a man-in-the-middle attack, the communications through the provider are encrypted with a key that only Bobby and Assata have access to. That is, the message only exists in a decrypted state on Assata’s and Bobby’s devices. Assata’s and Bobby’s devices are the endpoints of the communication (hence end-to-end encryption).
While E2EE is the gold standard, there are further considerations. As mentioned above, end-to-end encryption is only established if the communication provider (or another third-party member) does not mount a man-in-the-middle attack starting at the time of key exchange. However, as we covered in the chapter “The Man in the Middle,” if Assata and Bobby verify their keys through independent channels, they can determine whether or not a man-in-the-middle attack has occurred and so whether their communications are truly end-to-end encrypted.
While many apps or services claiming E2EE provide the ability to verify keys, many do not, providing little guarantee to trust the claims of E2EE. Further, for those E2EE apps that do provide the ability to verify keys, most operate on a Trust on First Use (TOFU) basis. That is, communications may begin without verifying keys first. However, while actually doing key verification is the only way to guarantee E2EE, the existence of the ability to verify keys is protective against automated man-in-the-middle attacks, as even a small fraction of users verifying keys would catch widespread man-in-the-middle attacks.
And, of course, E2EE only protects the communication between devices—it does not protect the data that is on the device. E2EE apps should be combined with strong passwords to protect the account or device.
In Context: Multiparty Video Chatting
There are many apps and services for video chatting between two or more people, with varying degrees of security. Here are three illustrative examples:
- Wire provides the gold standard of E2EE. Each user has an account that can be accessed from multiple devices (e.g., laptop and smartphone). There is a public key for each device that is used to establish an encryption key for a session (e.g., a video call), and the fingerprints of these keys can be compared to verify true E2EE. Wire allows E2EE video calls for groups of up to twelve users.
- Zoom allows video calls for much larger groups and does provide E2EE in that a video stream is encrypted and decrypted by the users with the same key. However, this key is established and distributed by the Zoom servers. Since Zoom has access to the encryption key, this cannot be considered true E2EE. Further, there is no mechanism for users to verify the encryption keys. As of summer 2020, Zoom had a proposal for establishing keys for true E2EE, but it has not yet implemented it.
- Jitsi Meet also provides large-group video conferencing, but only using in-transit encryption. However, Jitsi Meet is available to be hosted on any server (including your own, if you are so inclined). There is an instance of Jitsi Meet hosted by May First, a nonprofit that provides technical solutions to social movements and is a trusted third party to many groups. Even though May First has access to these communications, some would prefer to trust May First over a profit-driven solution such as Zoom.
- Blum, Josh, Simon Booth, Oded Gal, Maxwell Krohn, Julia Len, Karan Lyons, Antonio Marcedone, et al. “E2E Encryption for Zoom Meetings.” Zoom Video Communications, December 15, 2020.
- http-example © OSU OERU is licensed under a CC BY-NC (Attribution NonCommercial) license
- https-example © OSU OERU is licensed under a CC BY-NC (Attribution NonCommercial) license
- who-has-access-to-your-data © OSU OERU is licensed under a CC BY-NC (Attribution NonCommercial) license
- data-access-communication © OSU OERU is licensed under a CC BY-NC (Attribution NonCommercial) license
- notE2EE © OSU OERU is licensed under a CC BY-NC (Attribution NonCommercial) license
- E2EE © OSU OERU is licensed under a CC BY-NC (Attribution NonCommercial) license