Design a Real-Time Shared To-Do List

Scenario

Design the backend architecture for a real-time collaborative task list application (such as shared party checklists or team sprint boards). Multiple users can view, add, check off, edit, and reorder tasks within the same list concurrently. Edits must be durable, immediately acknowledged to the author, and propagated to other active collaborators with sub-second latency.

Requirements

Functional Requirements

• Shared Lists: Users can create lists, invite collaborators, and view real-time changes.
• Task Operations: Support creating, updating (title, status, assignee), deleting, and reordering tasks.
• Real-Time Synchronization: Active collaborators on a list must see changes pushed to them in near real-time.
• Durable Writes: A user modifying a task must receive an explicit, synchronous acknowledgement once the update is committed to persistent storage.
• Conflict Resolution: Explicitly handle concurrent edits (e.g., two users modifying the same item simultaneously or reordering overlapping items).
• Reconnect & Catch-up: Clients that temporarily lose connection must be able to catch up without re-downloading the entire list state.

Non-Functional Requirements

• Scale: 10 million daily active users (DAU); 1 million concurrent active lists at peak.
• Group Size: Typically 2–10 active users per list; up to 100 peak collaborators per list.
• Latency: Write durability acknowledgement < 100ms; end-to-end fan-out latency to collaborators < 250ms.
• Availability & Consistency: Authoritative state must maintain per-list linearizability or strong causal consistency; no silent data loss.

What Your Design Should Cover

1. API & Data Model: Core data models, task ordering strategy (e.g., fractional indexing / LexoRank), and versioning mechanism.
2. Architecture & Data Flow: Clear separation between the synchronous authoritative write path (HTTP/WebSocket to storage) and the asynchronous fan-out path (Pub/Sub to connected clients).
3. Concurrency & Conflicts: Strategy for handling conflicting updates and concurrent reordering (e.g., optimistic concurrency control with monotonic version numbers).
4. Connection Management & Catch-up: WebSocket connection tier, subscription tracking, and delta-based state synchronization upon client reconnection.