A robust, scalable, and fault-tolerant peer-to-peer (P2P) file sharing framework implemented in C/C++. The system uses TCP-based socket communication, multi-threading, and system-level constructs to enable high-throughput, verified, and distributed file exchange among peers.
The architecture is designed around multi-tracker coordination, chunk-level parallelism, and integrity validation, ensuring consistent availability and strong resilience to node and tracker failures.
This project presents a hybrid P2P model combining centralized metadata management with decentralized file distribution. Trackers maintain global metadata, while peers directly exchange file segments. The system supports:
- Secure user and group authentication
- Multi-source parallel downloads
- SHA1-based integrity validation
- On-demand serving of file pieces
- Multi-tracker metadata synchronization
The result is a high-availability distributed system suitable for real-world file sharing scenarios.
- User credentials hashed using SHA1.
- Tracker-managed session lifecycle.
- Prevents unauthorized access to shared resources.
- Logical separation of shared content via groups.
- Group admins manage membership requests.
- Strong permission enforcement for file visibility and sharing.
- Files split into 512 KB chunks for efficient distribution.
- Users upload file metadata (path + per-chunk hashes) to the tracker.
- Peers download chunks concurrently from multiple sources.
- Immediate re-sharing of downloaded chunks promotes decentralization.
App-level guarantees include:
✔ High transfer speeds ✔ Load distribution across peers ✔ Successful reconstruction with verified integrity
- Dynamically computes peer load.
- Prioritizes chunk download from least-loaded peers.
- Avoids network hotspots and increases overall throughput.
- Multiple trackers operate in parallel.
- Continuous metadata synchronization through TCP.
- Auto-recovery and full resync when a tracker rejoins after downtime.
- Provides strong fault tolerance and high system availability.
Each client node internally spawns:
- A server thread to serve chunk requests.
- Multiple worker threads for parallel chunk downloads.
- Independent communication handlers for both peer and tracker modules.
This ensures scalability under heavy workloads and minimizes transfer latency.
make
./tracker tracker_info.txt <tracker_id>
# Example:
./tracker tracker_info.txt 1tracker_info.txt specifies the IP and port of each tracker.
make
./client <IP>:<PORT> tracker_info.txt
# Example:
./client 127.0.0.1:6000 tracker_info.txt| Command | Description |
|---|---|
create_user <user_id> <passwd> |
Create a new account |
login <user_id> <passwd> |
Authenticate user |
create_group <group_id> |
Create a sharing group |
join_group <group_id> |
Request to join group |
leave_group <group_id> |
Leave an existing group |
list_requests <group_id> |
View pending join requests |
accept_request <group_id> <user_id> |
Approve a join request |
reject_request <group_id> <user_id> |
Reject a membership request |
list_groups |
List all groups |
list_my_groups |
List groups the user belongs to |
list_files <group_id> |
List files shared within a group |
upload_file <filepath> <group_id> |
Upload file metadata |
download_file <group_id> <file_name> <destination> |
Download a file |
show_downloads |
Display all downloads |
stop_share <group_id> <file_name> |
Stop sharing a file |
logout |
Logout current user |
exit |
Terminate the client |
The system has been tested under:
- Multi-peer parallel downloads
- Sudden peer shutdowns
- Tracker failure and recovery
- Artificial latency and network disturbance
- Consistency checks across multiple trackers
Results confirm:
✔ Strong fault tolerance ✔ No metadata inconsistency ✔ Graceful degradation under partial failures
- TLS-based encrypted peer and tracker communication
- Leader election for tracker decentralization
- Visual dashboard for monitoring transfers and cluster health
- Adaptive chunking for dynamic network conditions
- Auto-resume support for interrupted downloads
Malay Damani Tech Stack: C++, TCP/IP Sockets, POSIX Threads, Distributed Systems 🔗 LinkedIn: https://www.linkedin.com/in/malay-damani9126hacks/