README.md

Storage server

This server stores files in logical buckets. Each bucket maintains a Merkle tree, which enables on-demand Merkle proofs.

HTTP-based APIs

• File Upload POST /upload/:bucket_id/:file_name

  • Upload a file to a specific bucket

• Complete Upload POST /complete_upload/:bucket_id

  • Finalize a bucket upload. This instructs the server to generate the Merkle Tree for uploaded files in a specified bucket.

• File request GET /file/:bucket_id/:file_index

  • Retrieve a file by its index from a specified bucket.

• Proof request GET /proof/:bucket_id/:file_index

  • Retrieve a Merkle proof for a specific file in a bucket.

Merkle tree

build_merkle benchmark

• Build a merkle tree with leaves count 10_000, 100_000, 1000_000
• 14th Gen Intel(R) Core(TM) i9-14900k

 build_merkle_10000      time:   [832.36 µs 832.70 µs 833.02 µs]
 build_merkle_100000     time:   [8.3247 ms 8.3281 ms 8.3315 ms]
 build_merkle_1000000    time:   [85.651 ms 85.723 ms 85.800 ms]

verify_merkle_proof benchmark

• Generate and verify a proof for a leaf in 100_000 leaves tree
• 14th Gen Intel(R) Core(TM) i9-14900k

verify_merkle_proof_100000 time:   [1.2044 µs 1.2055 µs 1.2069 µs]

A HTTP-based Client with async-upload

Features:

• Upload concurrently all files from a source folder to the server in encrypted form.
• Maintain a Merkle root of the successfully uploaded files.
• Request both a file and its Merkle proof from the server.
• If the proof is valid, the client decrypts the file and stores it locally.
• Simple UI prompt

How to run

docker-compose build

# Bootstrap the server and run client with example files
docker-compose run client
docker-compose logs server -f

Run unit tests and benches

cargo test
cargo bench