TL;DR
This paper tackles the bottleneck of slow block propagation in Bitcoin, which limits its transaction throughput. The authors propose a novel "cut-through" forwarding mechanism and enhance it with rateless erasure codes to significantly speed up block relaying, enabling a potential 100x increase in transaction processing capacity without compromising security.
What Problem Does This Paper Address?
The core problem this paper addresses is the low transaction throughput of current blockchain networks, exemplified by Bitcoin. While blockchains offer robust security and decentralization, their transaction processing capacity (Transactions Per Second or TPS) is severely limited. Bitcoin, for instance, handles only about 5-7 TPS, a stark contrast to traditional payment systems like Visa (1700 TPS). A common approach to increasing TPS is to enlarge block sizes, allowing more transactions per block. However, larger blocks take longer to propagate across the peer-to-peer (P2P) network. This increased propagation delay can compromise blockchain security by leading to more forks (multiple valid chains being created simultaneously) and chain reorganizations, ultimately hindering widespread adoption. Existing solutions often involve complex changes to consensus protocols or new network infrastructures, which are difficult to integrate.
Methods
The paper proposes a two-pronged approach to accelerate block propagation:
- Cut-Through Compact-Block Relaying (Cut-through CBR): This method improves upon the existing "compact-block relaying" (CBR) protocol. In CBR, a node receives a compressed block (containing transaction hashes) and then requests individual transactions only if it's missing them. However, CBR still uses a "store-and-forward" mechanism, meaning a node must receive the entire compact block before forwarding it. The proposed Cut-through CBR breaks down the compact block into smaller "chunks" and allows a node to start forwarding these chunks to its peers before it has received the entire block. This "cut-through" or pipelined forwarding significantly reduces latency, especially for large blocks.
- Coded Cut-through CBR (Rateless Erasure Codes): To further enhance efficiency, the authors integrate rateless erasure codes (specifically fountain codes) into the Cut-through CBR. Instead of sending distinct chunks, the miner encodes these chunks into "coded symbols." This allows any peer to reconstruct the original chunks (and thus the block) by receiving a sufficient subset of these coded symbols from multiple sources. This approach leverages the P2P network's upload bandwidth more effectively, as peers can receive data from multiple sources simultaneously, rather than being limited to a single sender for each chunk.
Datasets
Not discussed in the paper. The authors implemented a Bitcoin network simulator on NS3 for their experiments.
Key Insights and Ideas
- Decoupling Block Reception and Forwarding: The core innovation of "cut-through" forwarding is realizing that a node doesn't need to wait for the entire block to arrive before it can start relaying parts of it. This parallel processing is key to reducing latency.
- Leveraging P2P Redundancy with Coding: Rateless erasure codes are a powerful tool for P2P networks. They allow for robust data recovery even when some data is lost or when downloaders have limited bandwidth, by enabling them to fetch data from multiple sources.
- Security vs. Throughput Trade-off is Manageable: The paper demonstrates that increasing block size (and thus TPS) doesn't have to come at the cost of increased propagation delay and compromised security. Their methods effectively mitigate this trade-off.
- Compatibility is Key: A major advantage of their approach is that it modifies only the networking layer, making it compatible with existing Bitcoin infrastructure and consensus mechanisms, facilitating easier adoption.
Key Results
The paper's simulations and analysis show significant improvements:
- 100x TPS Increase: The proposed protocols can increase the TPS capacity of the Bitcoin blockchain by up to 100x.
- Reduced Propagation Latency: For large blocks (e.g., > 25 MB), Cut-through CBR and Coded Cut-through CBR reduce median and tail propagation latency by up to three times compared to standard CBR, and over a hundred times compared to the original SR protocol.
- Suppressed Propagation Divergence: The propagation divergence factor (a measure of asynchrony among nodes) increases significantly with block size for traditional protocols. The proposed methods, especially Coded Cut-through CBR, effectively suppress this increase, keeping it at levels comparable to small block sizes even for large blocks. This directly translates to a lower stale-block rate.
- Coded CBR Outperforms Cut-through CBR: The coded version of the protocol offers further speedups, reducing latency by up to two times compared to the uncoded cut-through version.
Limitations and Future Work
- Linear Network Analysis: While the theoretical analysis uses a simple linear network, the practical evaluation is done on a simulated network with a more realistic topology.
- Optimal k for Coded CBR: The paper mentions finding the optimal
k(number of chunks) for Coded Cut-through CBR through experiments, implying that a theoretical derivation for the absolute optimalkmight be complex or context-dependent. - Future Work: The authors suggest that their protocols can be seamlessly incorporated into existing Bitcoin-like blockchains, implying future work could involve real-world deployment and testing. The paper also hints at further performance analysis and optimization.
Summary updated