P2PDocs is a peer-to-peer collaborative text editor developed for the Distributed Systems course at the University of Udine. The project explores how decentralized systems can handle real-time collaboration with consistency, conflict resolution, and offline-first operation.
Project Overview#
The goal of P2PDocs was to implement a fully decentralized editor where multiple peers can simultaneously edit documents. Core distributed algorithms were implemented and evaluated, including:
- Conflict-Free Replicated Data Types (CRDTs) for automatic conflict resolution
- Causal Broadcast for ordered message delivery
- Offline-first collaboration to allow peers to work without a central server
The system provides a hands-on demonstration of real-world challenges in distributed systems: network partitions, concurrency, and fault tolerance.
Technical Contributions#
1. LSEQ CRDT Implementation#
We adopted the LSEQ strategy to preserve user intention across peers:
- Position Vectors: Rational numbers for dense ordering
- Peer-ID Tagging: Resolve concurrent edits
- AVL Order-Statistic Tree: Efficient O(log n) insertions and deletions
2. Network Layer#
- Echo-Wave Protocol for reliable operation propagation
- Neighbor Handshake for joining peer synchronization
- Crash Recovery with ETS (Erlang Term Storage) for persistence
3. Testing & Validation#
- Unit Tests with ExUnit
- Multi-Node Testing using Docker to simulate network partitions and node failures:
./start_n_instances_for_tests.sh 5 # Launch 5 peers
./disconnect.sh 2 # Simulate node failureTeam#
Distributed Systems Course Project
Stack: Elixir, Node.js, Docker
Future Work#
Based on our final report, future improvements could include:
- Markdown support for rich text editing
- Real-time cursor tracking
- Optimized vector clock size for performance
Learn More#
For a technical deep dive, including full CRDT implementation details, network protocol design, and system validation, see the Project Report.
Explore the Code: GitHub
