Doppelganger in Bitcoin Mining Pools: An Analysis of the Duplication Share Attack

Yujin Potter, Yujin Kwon, Dohyun Kim, Yunmok Son, Jaeyeong Choi, Yongdae KimWisa 2016 · DOI

TL;DR

This paper analyzes the "Duplicate Share Attack" (DSA) in Bitcoin mining pools, demonstrating how attackers can exploit a vulnerability in how pools validate shares to earn unfair profits. The authors provide a mathematical framework to derive optimal attack strategies, showing that even with minimal computational power, attackers can achieve significant gains, particularly by targeting multiple smaller, unpatched pools.

What Problem Does This Paper Address?

Bitcoin's decentralized nature relies on miners to validate transactions and secure the network through a "proof-of-work" system. As mining difficulty increased, individual miners began pooling their resources in "mining pools" to increase their chances of earning rewards. These pools are typically managed by a central entity that distributes rewards proportionally to the "shares" submitted by individual miners, representing their computational effort.

The problem arises from a vulnerability known as the "Duplicate Share Attack" (DSA). This attack exploits a flaw in how some mining pools validate submitted shares. Specifically, certain pools failed to distinguish between shares with identical data but different capitalization in hexadecimal characters. This allowed an attacker to submit the same valid share multiple times, effectively claiming credit for more work than they actually performed, thereby receiving an unfair share of the mining rewards. The paper aims to mathematically analyze this attack, quantify its potential profit, and explore optimal strategies for attackers.

Methods

The authors employ a mathematical and game-theoretic approach to analyze the Duplicate Share Attack. They model the profit of miners and mining pools based on their computational power.

Key methodological contributions include:

  • Mathematical Modeling of DSA: The paper derives equations to represent an attacker's profit when targeting a single mining pool. This involves defining variables for the pool's power, the attacker's power, the ratio of attacker's power contributed to the pool, and the number of duplicate shares submitted.
  • Optimal Strategy Derivation: Based on the mathematical models, the authors derive the optimal strategy for an attacker. This strategy involves maximizing the number of duplicate shares submitted (k) and carefully choosing the ratio of their computational power allocated to the pool (γ) to maximize profit while avoiding suspicion.
  • Extension to Multiple Pools: The analysis is extended to scenarios where an attacker targets multiple mining pools simultaneously. This involves summing the profits from individual pool attacks and optimizing the power distribution across these pools.
  • Feasibility Analysis: The paper discusses the practical feasibility of the attack by identifying that while major pools have patched the vulnerability, smaller, less popular pools may still be susceptible.

Datasets

Not discussed in the paper. The analysis is theoretical and relies on mathematical modeling rather than empirical data from specific mining pools.

Key Insights and Ideas

  • Exploitable Vulnerability: The core insight is that a seemingly minor implementation detail (case-insensitivity in share validation) can lead to significant financial gains for attackers in Bitcoin mining pools.
  • Optimal Attack Strategy: The paper clearly defines an optimal strategy for attackers: submit as many duplicate shares as technically possible (limited by network latency and the exploit's nature) and carefully balance their contribution to the pool to maximize profit without raising suspicion.
  • Power of Small Pools: A critical finding is that targeting multiple small, unpatched mining pools can be more profitable than attacking a single large pool, even if the total computational power of the targeted pools is the same. This is because the vulnerability is more prevalent in less scrutinized pools.
  • Efficiency of DSA: The paper demonstrates that DSA is a highly efficient attack, capable of yielding substantial extra profits with minimal computational power compared to other known attacks like block withholding.

Key Results

  • Single Pool Attack: An attacker can achieve a global maximum profit of 12.49% of the total block subsidy when targeting a single pool with 25% of the total network's computational power, using the optimal strategy.
  • Multiple Pool Attack: By targeting multiple small, vulnerable pools, an attacker can earn 4.997% of the total block subsidy with a minimal computational power of 0.24%. This highlights the effectiveness of spreading the attack.
  • Vulnerability Persistence: The authors found that many small mining pools remain vulnerable to DSA, indicating that the threat is still present despite fixes in larger pools.
  • Comparison to Other Attacks: DSA is shown to be more efficient than previous attacks like block withholding, offering a higher profit margin for a given computational power.

Limitations and Future Work

  • Assumptions: The analysis relies on several simplifying assumptions, such as honest miners and managers (except for the attacker), and a fixed block subsidy. Real-world scenarios might involve more complex dynamics.
  • Practical Limits on k: While the theoretical maximum for duplicate shares (k) is high, practical limitations due to network latency and the specific exploit's constraints mean k is practically limited (e.g., to 2085).
  • Suspicion Threshold: The paper assumes an attacker can operate without raising suspicion by limiting their contribution to a pool to 50% of the total substantive power. Determining the exact threshold for suspicion in real-world pools is complex.
  • Future Work: The authors suggest that managers and developers of mining pools must prioritize patching the DSA vulnerability in their Stratum protocols to secure their operations. Further research could explore more sophisticated detection mechanisms or analyze the game-theoretic implications of such attacks on the broader Bitcoin ecosystem.

Summary updated

BibTeX
@inproceedings{potter2016doppelganger,
  author = {Yujin Potter and Yujin Kwon and Dohyun Kim and Yunmok Son and Jaeyeong Choi and Yongdae Kim},
  title = {{Doppelganger in Bitcoin Mining Pools: An Analysis of the Duplication Share Attack}},
  booktitle = {Wisa},
  year = {2016},
  doi = {10.1007/978-3-319-56549-1\_11},
}
Figure