Overcoming Data Availability Attacks in Blockchain Systems: Short Code-Length LDPC Code Design for Coded Merkle Tree

Debarnab Mitra,Lev Tauz,Lara Dolecek

DOI: https://doi.org/10.1109/TCOMM.2022.3194138

2022-08-28

Abstract:Light nodes are clients in blockchain systems that only store a small portion of the blockchain ledger. In certain blockchains, light nodes are vulnerable to a data availability (DA) attack where a malicious node makes the light nodes accept an invalid block by hiding the invalid portion of the block from the nodes in the system. Recently, a technique based on LDPC codes called Coded Merkle Tree was proposed by Yu et al. that enables light nodes to detect a DA attack by randomly requesting/sampling portions of the block from the malicious node. However, light nodes fail to detect a DA attack with high probability if a malicious node hides a small stopping set of the LDPC code. In this paper, we demonstrate that a suitable co-design of specialized LDPC codes and the light node sampling strategy leads to a high probability of detection of DA attacks. We consider different adversary models based on their computational capabilities of finding stopping sets. For the different adversary models, we provide new specialized LDPC code constructions and coupled light node sampling strategies and demonstrate that they lead to a higher probability of detection of DA attacks compared to approaches proposed in earlier literature.

Information Theory,Cryptography and Security

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how light nodes can effectively detect data availability attacks (DA attacks) in blockchain systems. Specifically:

1. **Background and Challenges**:
- Light nodes improve the scalability of the blockchain system by storing a small amount of blockchain data.
- Malicious nodes can initiate data availability attacks by hiding a part of an invalid block, causing light nodes to accept the invalid block.
- Existing solutions (such as Coded Merkle Tree, CMT) use Low - Density Parity - Check Codes (LDPC codes) to encode blocks to help light nodes detect DA attacks. However, if a malicious node hides a small stopping set in the LDPC code, the light node may not be able to detect the DA attack.

2. **Deficiencies of Existing Methods**:
- Although random LDPC codes can provide a relatively large minimum stopping set with long code lengths, they perform poorly with short code lengths, especially in low - latency systems and Internet - of - Things blockchains.
- When using random LDPC codes and random sampling strategies, if a malicious node hides a small stopping set, the probability that a light node can detect a DA attack will be significantly reduced.

3. **Main Contributions of the Paper**:
- Focusing on the design of LDPC codes with short code lengths, a new LDPC code construction algorithm (EC - PEG algorithm) and a greedy sampling strategy are proposed to increase the probability that light nodes can detect DA attacks.
- For stronger adversary models, a PEG algorithm with linear programming constraints (LC - PEG) and a linear programming sampling strategy (LP - sampling) are proposed, further increasing the detection probability.

4. **Specific Technical Solutions**:
- **Weak Adversary Model**: Use the EC - PEG algorithm to concentrate the stopping sets on a small number of variable nodes, and combine it with a greedy sampling strategy, which significantly reduces the probability of light node detection failure.
- **Medium and Strong Adversary Models**: Use the LC - PEG algorithm and LP - sampling strategy to further increase the detection probability by optimizing the structure of LDPC codes and the sampling strategy.

5. **Experimental Verification**:
- The effectiveness of the proposed LDPC codes and sampling strategies under different adversary models is verified by simulation results, showing their superiority over existing methods.

In conclusion, this paper aims to improve the ability of light nodes to detect data availability attacks under short code lengths by improving the design of LDPC codes and sampling strategies, thereby enhancing the security of blockchain systems.