Academic Research, LSC1 Cryptography & Zero-Egress RDM Hub

1. Official Institutional & Archival Identifiers

To support reproducible systems research in applied cryptography, browser security, and high-throughput scientific data transfer, all LimeShare protocol specifications and benchmark datasets are published under Creative Commons Attribution 4.0 International (CC BY 4.0) and indexed across permanent institutional repositories:

• Harvard University Dataverse Collection: https://dataverse.harvard.edu/dataverse/limeshare
• Harvard University Dataverse Primary Dataset DOI: doi:10.7910/DVN/M6TLOS (Version 2.0 Released)
• CERN Zenodo Primary LSC1 Protocol DOI (Vol. 1): doi:10.5281/zenodo.23671752
• CERN Zenodo Academic RDM & Scientific Datasets DOI (Vol. 2): doi:10.5281/zenodo.23671754
• CERN Zenodo WebCrypto vs. WebRTC P2P Benchmark DOI (Vol. 3): doi:10.5281/zenodo.23671756
• CERN Zenodo Verifiable Cryptographic Erasure DOI (Vol. 4): doi:10.5281/zenodo.23671758
• CERN Zenodo 2026 Cloud Egress & Attachment Limits DOI (Vol. 5): doi:10.5281/zenodo.23671762
• Machine-Readable Citations & Schema: BibTeX (/cite.bib) · Citation File Format (/CITATION.cff) · Frictionless Data Package (/datapackage.json)

2. Formal Specification of the LSC1 Streaming AEAD Container

Standard web-based encrypted file sharing implementations fail on multi-gigabyte scientific datasets because they accumulate the entire ciphertext into a single in-memory Blob or fail to authenticate chunk ordering. The LSC1 protocol implements the provably secure STREAM online authenticated encryption construction (Hoang, Reyhanitabar, Rogaway & Vizár, CRYPTO 2015) over WebCrypto AES-256-GCM:

• Key Hierarchy (HKDF-SHA256): A 256-bit master key K is generated via crypto.getRandomValues. Subkeys are derived via HKDF-SHA256(K, salt = transferId, info) with domain-separated labels for manifest, file:{fileId}, and share:{fileId}.
• 32-Byte Authenticated Header (AAD): Every file stream begins with magic = "LSC1" (4 B) || version = 0x01 (1 B) || chunkSize = 4,194,304 (4 B BE) || noncePrefix (7 B) || plaintextSize (8 B BE) || reserved (8 B). This 32-byte header is passed as Additional Authenticated Data (AAD) to every frame’s AES-GCM invocation.
• Deterministic 96-Bit STREAM Nonce: Frame i uses the 12-byte IV noncePrefix (7 B) || counter_i (4 B BE) || isLastFrame (1 B: 0x01 on final frame, 0x00 otherwise). Any attempt by an untrusted network or storage node to truncate the tail, reorder frames, or splice frames across files causes immediate GCM authentication tag failure before plaintext is emitted.
• O(1) Random-Access Byte Offset: The total ciphertext length is deterministic: 32 + ceil(size / 4194304) × 16 + size. Frame i starts at exact byte offset 32 + i × 4,194,320, enabling HTTP Range seeking for in-browser 4K video review and parallel multipart downloading.

3. University Research Data Management (RDM) & Lab Workflows

University laboratories, core sequencing facilities, and interdisciplinary consortia use LimeShare’s zero-egress architecture to transfer large scientific payloads without AWS S3 / GCP egress fees or institutional VPN bottlenecks:

@dataset{limeshare_harvard_2026,
  author       = {{LimeShare Cryptography and Distributed Systems Lab}},
  title        = {{LimeShare LSC1 Protocol: Authenticated Streaming Encryption (AES-256-GCM STREAM), WebRTC P2P \& Zero-Egress File Transfer Benchmark Dataset (2026)}},
  year         = {2026},
  publisher    = {Harvard Dataverse},
  version      = {V2},
  doi          = {10.7910/DVN/M6TLOS},
  url          = {https://doi.org/10.7910/DVN/M6TLOS}
}