Design a Slack-like messaging system

Design a Slack-like real-time team messaging system. Requirements: - Users can: - Create and join workspaces. - Create public and private channels within a workspace. - Send messages in channels and direct messages (DMs). - See message history and search messages by keyword. - Messages should appear in near real-time to all online participants in a channel or DM. - Support presence indicators (online/away) at a basic level. - System should support: - 10 million registered users. - 1 million daily active users. - Peak 200k concurrent connections. Tasks: 1. Clarify and state any additional assumptions you need (e.g., file sharing scope, message size limits, retention policy). 2. Propose a high-level architecture: - How clients connect (e.g., WebSockets, polling). - Major backend services and their responsibilities (e.g., gateway, messaging service, user service, search service). - Data storage for users, workspaces, channels, and messages. 3. Design the data model for core entities: - Users, workspaces, channels, memberships, and messages. - Explain how you would partition/shard message data. 4. Describe how real-time messaging works end-to-end: - From a user sending a message to it being delivered and persisted. - Handling fan-out to many subscribers in a busy channel. 5. Discuss scalability: - How to handle 200k+ concurrent connections. - Horizontal scaling strategies for message processing. - Handling hot channels with very high message throughput. 6. Address reliability and consistency: - Ensuring messages are not lost (at-least-once vs exactly-once semantics). - Ordering guarantees within a channel. 7. Describe how you would implement search over message history: - Indexing strategy. - High-level read/write path for search. 8. Mention key trade-offs and potential future extensions (e.g., threading, reactions, file attachments).