# Persistent BitTorrent Trackers

Crypto.world

> Private BitTorrent trackers keep communities fair by requiring users to maintain a good upload-to-download ratio, but the ratio lives in one central database: it cannot move to another tracker, the…

**arXiv**: 2511.17260

**Published**: 2025-11-21

**Authors**: François-Xavier Wicht, Zhengwei Tong, Shunfan Zhou, Hang Yin, Aviv Yaish

## AI summary

## Overview

**Research area:** Computer security and peer-to-peer systems (cs.CR), specifically incentive and reputation design for private BitTorrent trackers, combining applied cryptography, smart contracts, distributed hash tables, and trusted execution environments.

**Technical level:** Intermediate. The paper is written for readers who already understand how BitTorrent trackers and swarms work and who are comfortable with signature schemes, blockchains, and threat models; the cryptographic definitions are formal but the motivating narrative is accessible.

**Scope:** The paper specifies, constructs, and evaluates a "Persistent BitTorrent Tracker Scheme" (PBTS) that replaces self-reported upload statistics with cryptographically signed transfer receipts and stores reputation in smart contracts so it survives tracker shutdowns.

## What This Paper Is About

Private BitTorrent trackers keep communities fair by requiring users to maintain a good upload-to-download ratio, but the ratio lives in one central database: it cannot move to another tracker, the server is a single point of failure, and the statistics are self-reported and therefore forgeable. The paper redesigns the tracker so that reputation is recorded in smart contracts on a blockchain, uploads are proven by receiver-signed cryptographic receipts aggregated with BLS signatures, and peers can still discover each other through an authenticated DHT when the tracker is offline. The goal is a tracker whose reputation outlives the tracker itself and whose statistics are verifiable rather than trusted.

## Key Contributions

1. **A formal Persistent BitTorrent Tracker Scheme (PBTS)** — eight algorithms (`Setup`, `KeyGen`, `Register`, `Announce`, `Report`, `Attest`, `Ver`, `Migrate`) with explicit security requirements, plus proofs of four security properties under standard cryptographic assumptions.
2. **Verifiable per-piece attestation for BitTorrent** — receivers sign receipts binding the infohash, the sender's public key, the piece hash, the piece index, and an epoch timestamp, so upload credit is checked cryptographically instead of self-reported. The tracker aggregates receipts and credits both the uploader's upload counter and each receipt signer's download counter.
3. **Portable reputation via factory-deployed smart contracts** — a factory contract deploys per-tracker reputation contracts, and single-hop migration lets a successor tracker inherit state from a predecessor, so a shutdown does not erase contribution history.
4. **Tracker-less operation and privacy mitigations** — on-chain reputation acts as a public key infrastructure (PKI) and allow-list for an authenticated DHT fallback; an optional TEE mode (Intel TDX / AMD SEV) adds operator-integrity and peer-confidentiality guarantees; ephemeral session keys, zero-knowledge membership proofs, and confidential reputation via homomorphic commitments address linkability on a public ledger.

## Main Findings

- **Overhead stays small:** measured on a prototype, transfer receipts add less than 5% end-to-end overhead with typical piece sizes (the abstract also phrases this as less than 5% throughput overhead for typical workloads).
- **Hybrid ECDSA/BLS signing is much cheaper for clients:** ECDSA signs individual piece receipts at transfer time for low per-operation latency, while BLS is the overarching scheme that aggregates many receipts into one proof at report time; this "reduces client-side signing cost by an order of magnitude compared to using BLS throughout."
- **Reputation becomes portable:** because contributions are written to a factory-deployed contract, a new tracker instance can inherit a predecessor's state through single-hop migration, so users do not lose standing when a community disappears. The paper's motivating example is the 2007 takedown of the OiNK tracker, after which displaced members could not carry over their ratios.
- **The tracker is no longer a single point of failure for peer discovery:** when the tracker is unavailable, peers fall back to an authenticated DHT where stored reputation serves as the PKI, preserving access control without the server.
- **Collusion with real data yields no net gain:** peers cannot forge receipts, and although colluding peers could exchange genuine data to produce legitimate receipts, the downloader always records a reputation loss, while accountable sponsorship (as in the RED tracker's interview/sponsorship model) makes creating accounts costly.
- **TEE mode is optional and trades a hardware assumption for two guarantees:** integrity (the operator cannot selectively drop valid attestations to disadvantage specific peers) and confidentiality (peer IP addresses and activity patterns are hidden from the operator). The paper explicitly notes TEEs do not guarantee liveness or availability, since the host controls scheduling, initialization, and system calls.
- **The blockchain layer is abstracted, not assumed:** any "bulletin board" satisfying liveness, write integrity, and censorship resistance can replace a smart contract — a permissioned ledger run by a tracker federation is named as an alternative with different trust and cost trade-offs.

## Methodology in Plain English

The authors start from a diagnosis of three structural weaknesses in private trackers — non-portable ratios, centralization, and unverifiable statistics — and then rebuild each layer of the tracker to remove them.

For statistics, they flip the reporting direction: instead of an uploader claiming what it sent, the downloader signs a receipt for each piece it actually received, after checking the piece against the torrent's published hash. Uploaders collect these receipts, and at report time the tracker checks one aggregated signature covering the whole batch, then updates both parties' counters. To keep signing cheap on high-bandwidth transfers, they use two signature schemes: a fast per-piece scheme (ECDSA) during transfer and an aggregatable scheme (BLS) for compressing many receipts into a single proof.

For persistence, reputation counters live in smart contracts deployed by a factory contract, so a successor tracker can be spun up and inherit the state of a dead one. For availability, on-chain registration records double as a public key directory, letting peers authenticate each other in a DHT when the tracker is gone. For distrust of the operator, the tracker software can run inside a TEE, which attests that the code ran correctly and hides peer addresses from the operator.

They then formalize the scheme, state the threat model (blockchain trusted for contract execution; peers and DHT nodes untrusted; operator either trusted or TEE-constrained), prove four security properties, analyze privacy leakage on a public ledger, and build and measure a prototype. Note that the supplied paper text is truncated partway through the attestation algorithm, so the detailed theorem statements in Section V, the full privacy analysis in Section VI, and the per-benchmark numbers in Section VII are not available here; only the headline figures quoted above are reported.

## Why This Matters

Reputation systems that live in one server die with that server, and statistics that users report about themselves are only as trustworthy as the users. This paper shows a practical way to make contribution history a durable, verifiable asset, and it does so without abandoning the tit-for-tat economics that make private trackers work.

**Impact on research:** it connects peer-to-peer incentive design with on-chain reputation, aggregatable signatures, and TEEs, and it adds a takedown-resistance threat model that earlier BitTorrent work largely did not consider (prior work focused mostly on network-level interference and proxy rotation).

**Real-world applications:**

- Private tracker communities recovering from shutdowns or seizures without stranding members' ratios.
- Federated or invitation-based content communities that want to recognize standing earned elsewhere.
- Any peer-to-peer contribution system needing auditable proof of transfer (mirrors, volunteer compute, archival projects).
- Accountability tooling where receipts, rather than moderator audits, settle disputes over reported contributions.

**Industry relevance:** the design speaks directly to blockchain infrastructure providers, confidential-computing vendors (the paper targets Intel TDX and AMD SEV), and operators of permissioned file-sharing services who want verifiability without giving up access control.

## Future Directions

- **Quantifying the TEE mode in practice:** the paper positions TEE deployment as optional, and the trust placed in the hardware manufacturer, plus the fact that TEEs do not guarantee liveness, leaves open how much the untrusted-operator variant actually costs in performance and operational complexity.
- **Deploying and measuring the privacy mitigations:** ephemeral session keys, zero-knowledge membership proofs for DHT participation, and homomorphic-commitment reputation are proposed; the supplied text does not report measured costs for these mechanisms, so their overhead and usability remain an empirical question.
- **On-chain cost and scale:** the design writes reputation state to a blockchain per registration and per report; how this behaves for large swarms and whether a permissioned bulletin board is the better trade-off is left as a design axis rather than a settled answer.
- **Adversarial behavior beyond forgery:** the analysis argues collusion with real data yields no net benefit and relies on Sybil-resistant registration such as interviews or accountable sponsorship; how robust that assumption is against determined, well-resourced attackers is an open question the scheme inherits from existing private-tracker practice.

## Target Audience

Researchers and graduate students in security, applied cryptography, and peer-to-peer or blockchain systems; protocol designers building reputation or incentive layers; and technically sophisticated operators of private trackers or permissioned content communities who want to understand what verifiable, portable reputation would require. Readers without a BitTorrent or cryptography background will find the formal sections demanding, but the problem framing and the architectural argument are readable more broadly.

## Authors’ abstract

Private BitTorrent trackers enforce upload-to-download ratios to prevent free-riding, but suffer from three critical weaknesses: reputation cannot move between trackers, centralized servers create single points of failure, and upload statistics are self-reported and unverifiable. When a tracker shuts down, users lose their contribution history and cannot prove their standing to new communities. We address these problems by storing reputation in smart contracts and replacing…

[The full abstract on arXiv](https://arxiv.org/abs/2511.17260)

[Read the original paper](https://arxiv.org/abs/2511.17260)

---

Canonical: https://crypto.world/paper/persistent-bittorrent-trackers-2511.17260
