]> The Matrix.org Foundation C.I.C.

travisr@matrix.org

Applications and Real-Time More Instant Messaging Interoperability mimi terminology mimi definitions mimi words This document introduces a set of terminology to use when discussing or describing concepts within MIMI. About This Document The latest revision of this draft can be found at . Status information for this document may be found at . Discussion of this document takes place on the More Instant Messaging Interoperability Working Group mailing list (), which is archived at . Subscribe at . Source for this draft and an issue tracker can be found at .

Introduction The More Instant Messaging Interoperability (MIMI) working group is chartered to specify the minimal set of mechanisms to make modern instant messaging applications interoperable. Through prior discussions and upcoming documents, it's become important to have a shared understanding for what is being discussed specifically. This document expands upon MLS's terminology by defining terms specific to MIMI's purpose. Document authors SHOULD prefix terms from the MLS specification with "MLS" to denote that the term is specifically coming from the MLS specification. For example, "MLS Group" versus "Group". This document defines terms which are non-conflicting to help ensure clarity when the MLS prefix is missing, however. Documents within the MIMI working group may introduce their own terminology to explain concepts within their context. Those documents SHOULD NOT override or change the terminology described in this document or from MLS.

Terminology MIMI defines: Messaging Provider or Provider: A service offering instant messaging to users. Typically a provider will have a server to route events between users (or clients, specifically). User: A (normally) human operator of a client. Users have a User ID to identify them canonically within the system. User Identity: A mapping of User ID to external identifier, such as a phone number. In some cases, the user identity can be synonymous with the user ID, however the default assumption if not clarified is that they are different concepts. Client: A user interface for messaging, performing encryption as needed. Presents chats to the user to interact with. Synonymous with MLS Client. Clients have a Client ID to canonically identify them among a user's other clients. Clients can additionally be called Devices to differentiate them from a named application. Chat: The place where users communicate. This is semantically distinct from an MLS Group: an MLS Group is responsible for handling client keys while a chat is simply the user-facing construct for communication, however systems can declare a chat to be the same as an MLS Group depending on their design. Chats have a Chat ID to canonically identify them within the system. Chats can additionally be called Rooms, Conversations, and Channels. A chat can have any one of many different characterizations/behaviours, called Chat Types:

• DM: A direct message chat between exactly two users. DMs are typically created when a user searches for another user to message, rather than creating a group chat. All users in the chat have the same permission capabilities under the access control semantics. The chat name is that of the opponent user in the DM. DMs are typically canonical: exactly one chat with the opponent exists at a time.
• Group DM: A subtype of DMs where there are more than two users. The chat name consists of the opponent users in the DM. Inherited from DMs, Group DMs are also canonical: creating a new Group DM with the exact same users ultimately redirects to the existing chat instead of creating a new one. These may also be called Ad-hoc Chats in some scenarios.
• Group Chat: A chat which requires an invite to be able to participate, and can have zero or more users. A user-provided chat name is defined for the chat. Typically, the creator has admin permissions while other designated users have either admin permissions or moderator permissions. Most users have default permissions which allow them to send events to the chat. Unlike (Group) DMs, multiple chats with the same set of users can exist at the same time, even with the same chat name.
• Public Chat: A group chat which does not require an invite to participate. Usually these types of chats are discovered through a directory or website. Chats which require a request to join, or knock, are considered public chats. Sometimes these will be referred to as Public Chatrooms.
• Auditorium Chat: Either a group chat or public chat where most users are unable to send events and cannot be seen as chat members by other users. When an event is sent, it may appear to be sent by the chat itself rather than the specific user who sent it. Sometimes these are called Auditorium Rooms.

Depending on the system, a chat's type can be mutable. For example, a user may be permitted to introduce new users to a group DM to implicitly convert it to a group chat, or they simply may be unable to implicitly or explicity change the chat type. Chat Name: The title or human-focused textually distinguishing factor for the chat. It may be automatically generated based on the chat members. Chat Avatar: A picture or graphic associated with the chat, usually in combination with a chat name. Chat avatars can be automatically generated based on the chat members. Chat Member: A joined user in the chat. Note that this may have implications on the associated MLS Members in the MLS Group for the user's devices. Event: The container for an encrypted MLS Message, sent over the wire between servers and clients (through their local servers). Events have an Event ID to canonically identify them at least within the chat. Chat Property: Information stored in the chat, such as the name, topic, avatar, chat members, etc. This may be in the shape of an event. Note that chat properties are different from what is needed to construct an MLS Group Context. Message: Synonymous with an MLS Message. Messages have a Message ID to canonically identify them at least within the group. These are essentially what a user would call a "message", though specifically the unencrypted portion. When encrypted, they are called events. Server: Synonymous with an MLS Delivery Service (DS). Responsible for routing events to other servers and local clients. Note that the role of a server can be accomplished in the same logical place as a client (i.e.: in peer-to-peer environments), however the default assumption if not clarified is that the client and server are two different entities. Client-Server API: The interface between a client and server. This may be nothing more than a function call if the client and server are the same logical entity. Server-Server API: The interface between a server and another server. Transport: The method and format for moving information between entities. For example, a Client-Server API might describe HTTPS and JSON as its transport. For added clarity, documents should clarify which API surface they are defining a transport for ("Server-Server Transport", for example). Message Format: The specific format that clients use within the encrypted body of an MLS Message. Sometimes this will also be called the Content Format. Access Control: The set of algorithms which determine whether an event or MLS Message is permitted in the chat/group. For instance, this may define whether an MLS Proposal is accepted or whether the user is able to become a chat member. Admin Permissions: Typically at least the user who created the chat, these permissions grant the associated users an ability to change the permissions of other users in the chat. The set of users with these permissions in a chat are called Admins. Admin permissions inherit moderator permissions. Moderator Permissions: These permissions grant the associated users an ability to kick, ban, or otherwise restrict another user's ability to use the chat. For example, deleting a spammer's events. The set of users with these permissions in a chat are called Moderators. Note that with both moderator permissions and admin permissions a system may have finer granularity, such as a set of users being able to kick but not ban. Documents with these semantics should clarify this case. Invite: An action taken by a user in a chat to encourage another user to become a chat member (or joined to the chat). This can be explicit through the server-server API, or implicit with an invite code. Invite Code: An out-of-band invite for a user to join the chat, such as a text string shared with the target user. The sharing mechanism can additionally be a URL or QR code. Direct Invite: An invite to a DM or Group DM. These invites might appear in a different section of the client's UI to denote their semantic difference from a non-DM invite. Kick: An action taken by a user in a chat to remove another user, assuming the sender has moderator permissions. The affected user can re-join the chat with either another invite or by simply joining, depending on the chat type. Ban: Similar to kicking, a user is kicked from the room and not allowed to re-join until unbanned by a moderator. If a user is banned, they are typically not able to knock to get back into the chat either. Knock: A request from a user (who is not currently a chat member) to participate in the chat. Upon acceptance, the requesting user may receive an invite or be directly joined to the chat. Upon rejection, the requesting user can be banned or otherwise declined. Status: Temporarily displayed text or images associated with a user's profile. Instagram Stories are an example of a user's status.

Diagram Reference In the simplest possible form, interoperable messaging between two end users looks as such:

Typical, simplified, architecture for interoperability | Provider 1 |<----->| Provider 2 |<----->| Client 2 | | | | | | | | | +----------+ +------------+ +------------+ +----------+ ^ ^ | D | +-------------------------------------------------------------+ ]]>

In this figure, Client 1 is directly connected to Provider 1 over A. A is a provider-specific Client-Server API and transport. Similarly, Client 2 is directly connected to Provider 2 over C. C is also a provider-specific Client-Server API and transport. Provider 1 and 2 talk to each other with a Server-Server API over a transport. This is B in the figure. Clients additionally have an implicit method of communication (D in the figure) where they use a shared Message Format. There is no transport for D in this figure: the figure is denoting that servers/providers are unable to assist with a format conversion due to the event's content (an MLS message) being encrypted.

IANA Considerations This document has no IANA actions.

Normative References Cisco Inria & Mozilla Phoenix R&D Meta Platforms Google University of Oxford Messaging applications are increasingly making use of end-to-end security mechanisms to ensure that messages are only accessible to the communicating endpoints, and not to any servers involved in delivering messages. Establishing keys to provide such protections is challenging for group chat settings, in which more than two clients need to agree on a key but may not be online at the same time. In this document, we specify a key establishment protocol that provides efficient asynchronous group key establishment with forward secrecy and post-compromise security for groups in size ranging from two to thousands. Inria & Mozilla Mozilla Google Twitter Wire The Messaging Layer Security (MLS) protocol (I-D.ietf-mls-protocol) specification has the role of defining a Group Key Agreement protocol, including all the cryptographic operations and serialization/deserialization functions necessary for scalable and secure group messaging. The MLS protocol is meant to protect against eavesdropping, tampering, message forgery, and provide further properties such as Forward Secrecy (FS) and Post-Compromise Security (PCS) in the case of past or future device compromises. This document describes a general secure group messaging infrastructure and its security goals. It provides guidance on building a group messaging system and discusses security and privacy tradeoffs offered by multiple security mechanisms that are part of the MLS protocol (e.g., frequency of public encryption key rotation). The document also provides guidance for parts of the infrastructure that are not standardized by the MLS Protocol document and left to the application or the infrastructure architects to design. While the recommendations of this document are not mandatory to follow in order to interoperate at the protocol level, they affect the overall security guarantees that are achieved by a messaging application. This is especially true in case of active adversaries that are able to compromise clients, the delivery service, or the authentication service.