Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PREAMBLE and IMRECEIVING for Improved Large Message Handling in libp2p GossipSub

Muhammad Umar Farooq, Daniel Kaiser

TL;DR

The paper analyzes large-message handling in GossipSub and shows that ignoring message size in forwarding and IWANT decisions causes high bandwidth use and slow dissemination. It introduces PREAMBLE, IMRECEIVING, IWANT improvements, and a reduced-forwarding strategy, all activated for large messages exceeding a threshold, and validates them with the shadow simulator. The results demonstrate substantial improvements, including up to 61% bandwidth reduction and up to 35% latency reduction for large messages, along with reduced duplicates. A GossipSub v1.4 prototype combining these techniques outperforms prior versions, offering practical gains for scaling pub/sub in libp2p networks. Future work includes assessing robustness under adversarial conditions and evaluating on broader test networks such as Ethereum testnets.

Abstract

Large message transmissions in libp2p GossipSub lead to longer than expected network-wide message dissemination times and very high bandwidth utilization. This article identifies key issues responsible for this behavior and proposes modifications to the protocol for transmitting large messages. These modifications preserve the GossipSub resilience and fit well into the current algorithm. The proposed changes are rigorously evaluated for performance using the shadow simulator. Results reveal that the suggested changes reduce bandwidth utilization by up to 61% and message dissemination time by up to 35% under different traffic conditions.

PREAMBLE and IMRECEIVING for Improved Large Message Handling in libp2p GossipSub

TL;DR

The paper analyzes large-message handling in GossipSub and shows that ignoring message size in forwarding and IWANT decisions causes high bandwidth use and slow dissemination. It introduces PREAMBLE, IMRECEIVING, IWANT improvements, and a reduced-forwarding strategy, all activated for large messages exceeding a threshold, and validates them with the shadow simulator. The results demonstrate substantial improvements, including up to 61% bandwidth reduction and up to 35% latency reduction for large messages, along with reduced duplicates. A GossipSub v1.4 prototype combining these techniques outperforms prior versions, offering practical gains for scaling pub/sub in libp2p networks. Future work includes assessing robustness under adversarial conditions and evaluating on broader test networks such as Ethereum testnets.

Abstract

Large message transmissions in libp2p GossipSub lead to longer than expected network-wide message dissemination times and very high bandwidth utilization. This article identifies key issues responsible for this behavior and proposes modifications to the protocol for transmitting large messages. These modifications preserve the GossipSub resilience and fit well into the current algorithm. The proposed changes are rigorously evaluated for performance using the shadow simulator. Results reveal that the suggested changes reduce bandwidth utilization by up to 61% and message dissemination time by up to 35% under different traffic conditions.

Paper Structure

This paper contains 9 sections, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Problem Description and Background
  3. Proposed Modifications
  4. PREAMBLE for Quick Identification of Messages
  5. IMRECEIVING Message to Minimize Duplicates
  6. IWANT Message Improvements
  7. Reduced Forwarding for a Manageable Workload
  8. Results and Discussions
  9. Conclusion

Figures (4)

  • Figure 1: Temporal representation of message spread with 100-millisecond intervals
  • Figure 2: Average number of duplicates ($\bar{d}$): impact of IDONTWANT and IWANT-reply messages
  • Figure 3: Performance against increasing message size: latency ($L_{cov}^{100}$), bandwidth ($B_N$), average duplicates ($\bar{d}$)
  • Figure 4: Increasing number of publishers (a,c,e) and increasing number of nodes (b,d,f): latency ($L_{cov}^{100}$), bandwidth ($B_N$), average duplicates ($\bar{d}$)