Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

On the Bandwidth Consumption of Blockchains

Andrei Lebedev, Vincent Gramoli

TL;DR

This work addresses the practical problem of bandwidth consumption in blockchains by delivering the first empirical, cross-protocol comparison across five layer-1 networks using a controlled 25-node testbed. It shows that transport mechanisms (polling vs WebSockets) and block propagation strategies (full blocks vs hashes) are the primary drivers of traffic, rather than raw transaction size, and that traffic patterns vary dramatically with network scale and validator composition. Solana exhibits node-dependent, high-bandwidth growth with network size, while Algorand and Redbelly scale mainly with the validator set; Avalanche and Aptos show hybrid scaling, with nuanced propagation behaviors. The findings have direct implications for deployment cost and robustness, highlighting design choices that can mitigate egress charges and improve resilience in large-scale blockchain deployments.

Abstract

With the advent of blockchain technology, the number of proposals has boomed. The network traffic imposed by these blockchain proposals increases the cost of hosting nodes. Unfortunately, as of today, we are not aware of any comparative study of the bandwidth consumption of blockchains. In this paper, we propose the first empirical comparison of blockchain bandwidth consumption. To this end, we measure the network traffic of blockchain network nodes of five blockchain protocols: Algorand, Aptos, Avalanche, Redbelly and Solana. We study the variation over time, differentiate the receiving and sending traffic and analyze how this traffic varies with the number of nodes and validators. We conclude that the transport protocol is the main factor impacting the network traffic, segregating node roles helps reduce traffic and different blockchains are differently impacted by the network size.

On the Bandwidth Consumption of Blockchains

TL;DR

This work addresses the practical problem of bandwidth consumption in blockchains by delivering the first empirical, cross-protocol comparison across five layer-1 networks using a controlled 25-node testbed. It shows that transport mechanisms (polling vs WebSockets) and block propagation strategies (full blocks vs hashes) are the primary drivers of traffic, rather than raw transaction size, and that traffic patterns vary dramatically with network scale and validator composition. Solana exhibits node-dependent, high-bandwidth growth with network size, while Algorand and Redbelly scale mainly with the validator set; Avalanche and Aptos show hybrid scaling, with nuanced propagation behaviors. The findings have direct implications for deployment cost and robustness, highlighting design choices that can mitigate egress charges and improve resilience in large-scale blockchain deployments.

Abstract

With the advent of blockchain technology, the number of proposals has boomed. The network traffic imposed by these blockchain proposals increases the cost of hosting nodes. Unfortunately, as of today, we are not aware of any comparative study of the bandwidth consumption of blockchains. In this paper, we propose the first empirical comparison of blockchain bandwidth consumption. To this end, we measure the network traffic of blockchain network nodes of five blockchain protocols: Algorand, Aptos, Avalanche, Redbelly and Solana. We study the variation over time, differentiate the receiving and sending traffic and analyze how this traffic varies with the number of nodes and validators. We conclude that the transport protocol is the main factor impacting the network traffic, segregating node roles helps reduce traffic and different blockchains are differently impacted by the network size.
Paper Structure (33 sections, 6 figures)

This paper contains 33 sections, 6 figures.

Table of Contents

  1. Introduction
  2. Background and Blockchain Networks
  3. The Algorand network
  4. The Avalanche network
  5. The Aptos network
  6. The Redbelly network
  7. The Solana network
  8. Experimental Settings
  9. Distributed system setup
  10. Measuring peer-to-peer bandwidth
  11. Varying Consumption
  12. Bandwidth consumptions as heatmaps
  13. Transaction sizes
  14. Consumption Skewness over Routes
  15. Some nodes exchange more data than others
  16. ...and 18 more sections

Figures (6)

  • Figure 1: Heatmaps $M_{i,j}$ of the bandwidth used in between sending node $n_i$ and receiving node $n_j$ for each blockchain, 20 nodes, 20 validators.
  • Figure 2: Transaction size per blockchain.
  • Figure 3: Client-side network traffic (TX and RX) stacked by experimental phase. Y-axis scales differ across subplots to accommodate variations in magnitude.
  • Figure 4: Heatmaps $M_{i,j}$ of the bandwidth used in between sending node $n_i$ and receiving node $n_j$ for each blockchain.
  • Figure 5: Total bandwidth consumption () for blockchains under test. The top row plots bandwidth against the number of nodes (grouped by validator count), while the bottom row plots bandwidth against the number of validators (grouped by node count).
  • ...and 1 more figures