One-dimensional vs. Multi-dimensional Pricing in Blockchain Protocols

Aggelos Kiayias, Elias Koutsoupias, Giorgos Panagiotakos, Kyriaki Zioga · 2025

Read One-dimensional vs. Multi-dimensional Pricing in Blockchain Protocols by Aggelos Kiayias, Elias Koutsoupias, Giorgos Panagiotakos, Kyriaki Zioga (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-13271.txt.

Introduction

Aggelos Kiayias, Elias Koutsoupias, Giorgos Panagiotakos, and Kyriaki Zioga compare two ways of charging for blockchain resources in a 2025 paper. A transaction eats several resources at once: storage, computing, and bandwidth. Most chains charge one combined price, which can leave capacity idle when one resource runs out first.

Pricing each resource on its own promises better use at steady state, and the authors prove it wins under stable conditions. In shifting states, though, the single price catches up faster and is far simpler to compute. That flip side matters before any chain switches over. The paper ends with mechanisms meant to soften the transition costs of multi-part pricing.

Catalog record

CREATORAggelos Kiayias, Elias Koutsoupias, Giorgos Panagiotakos, Kyriaki Zioga
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-13271

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

One-dimensional vs. Multi-dimensional Pricing in Blockchain Protocols Aggelos Kiayias, Elias Koutsoupias, Giorgos Panagiotakos, Kyriaki Zioga Abstract Blockchain transactions consume diverse resources, foremost among them storage, but also computation, communication, and others. Efficiently charging for these resources is crucial for effective system resource allocation and long-term economic viability. The prevailing approach, one-dimensional pricing, sets a single price for a linear combination of resources. However, this often leads to under-utilization when resource capacities are limited. Multi-dimensional pricing, which independently prices each resource, offers an alternative but presents challenges in price discovery. This work focuses on the welfare achieved by these two schemes. We prove that multi-dimensional pricing is superior under stable blockchain conditions. Conversely, we show that one-dimensional pricing outperforms its multi-dimensional counterpart in transient states, exhibiting faster convergence and greater computational tractability. These results highlight a critical trade-off: while multi-dimensional pricing offers efficiency gains at equilibrium, its implementation incurs costs associated with system transitions. Our findings underscore the necessity for a deeper theoretical understanding of these transient effects before widespread adoption. Finally, we propose mechanisms that aim to mitigate some of these issues, paving the way for future research. 1 Introduction Blockchain ledgers, by their very nature, operate under strict resource constraints. The decentralized and replicated design of these systems means every transaction included in a block must be processed, validated, and stored by all participating nodes. This distributed overhead, encompassing storage, computation, and communication , inherently limits the number of transactions that can be effectively processed within a given timeframe. Consequently, blockchain protocols impose restrictions on how blocks are assembled, leading to a setting where only a fraction of pending transactions can be admitted into the ledger at any given time . To ensure efficient resource allocation and prioritize the most economically valuable transactions, they must be suitably priced to offset the costs incurred by their processing and inclusion. In simpler blockchain architectures, such as Bitcoin, this pricing mechanism is largely single-resource based . Transaction inclusion is primarily determined by a “feerate”, defined as the fee paid per byte (or virtual byte) of transaction data. Transactions offering a higher feerate are preferentially included, as Bitcoin blocks have an upper bound on “virtual bytes”. This effectively treats storage as the primary, if not sole, bottleneck resource. However, as blockchain platforms evolve to support more complex functionalities, the single-resource pricing model becomes increasingly inadequate. Blockchains offering extensive smart contract capabilities, like Ethereum, must account for a wider array of resources beyond mere data storage, including computational steps and memory access. Ethereum, for instance, introduced “gas” as a global unit of account, pricing all operations of its virtual machine in terms of the total gas they consume. While a step towards recognizing diverse resource demands, this still aggregates all resource consumption into a single “gas” dimension. This prevailing approach, which sets a single price for a linear combination of resources, can be characterized as “ one-dimensional pricing ”. Yet, it is increasingly evident that transactions on sophisticated virtual machines may have distinct and non-linearly related requirements across multiple resource types, such as storage, computational steps, and network bandwidth. This inherent multi-dimensionality of resource consumption naturally suggests that a “ multi-dimensional ” fee approach, where each resource is independently priced, might be more appropriate. Recent developments in systems like Ethereum, such as EIP-4844, further underscore this shift by introducing new distinct resource types, like “blobs”, in addition to the existing “gas” mechanism. Blobs are designed for temporary, high-throughput data storage, distinct from the more permanent storage associated with typical transaction data, highlighting the growing recognition of different resource types with unique pricing considerations. The intuitive appeal of multi-dimensional pricing lies in its potential for more efficient resource allocation, allowing for finer-grained control over demand for specific, constrained resources. However, despite these theoretical advantages, one-dimensional pricing remains the dominant paradigm in current blockchain implementations. This discrepancy points to a critical trade-off that remains not fully understood, despite the recent expansion of literature focusing on improving multi-dimensional pricing [ 8 , 1 , 7 , 10 , 11 ] . This paper focuses on this fundamental trade-off, comparing the welfare achieved by one-dimensional and multi-dimensional pricing schemes for blockchain transactions. We demonstrate that, under stable blockchain conditions and at equilibrium, multi-dimensional pricing theoretically outperforms its one-dimensional counterpart, leading to superior resource utilization and overall welfare. Conversely, we show that in transient states, one-dimensional pricing exhibits faster convergence and greater computational tractability, making it more robust during periods of rapid change. Furthermore, fast convergence may lead to higher welfare as it can reduce the loss of welfare during the suboptimal pricing of the transient phase. These findings highlight a crucial dilemma: while multi-dimensional pricing offers clear efficiency gains at equilibrium, its widespread adoption necessitates a deeper theoretical understanding of the costs associated with system transitions and the challenges of price discovery in a dynamic environment. Our Results. We conduct a comparative investigation of one-dimensional and multi-dimensional transaction fee systems, highlighting their performance in welfare optimization and dynamic behavior. • Welfare Optimization. We prove that under stable blockchain conditions and at equilibrium, multi-dimensional pricing achieves superior welfare compared to its one-dimensional counterpart (Theorem 1 ). This rigorous finding stems from its ability to more efficiently accommodate resource consumption within the same space. • Stabilization Time. Conversely, we demonstrate that in transient states, one-dimensional pricing significantly outperforms multi-dimensional pricing in terms of convergence. Specifically, we quantify its faster convergence to the steady state (Theorem 2 ). • Computational Tractability. In the context of transient states, we also show that one-dimensional pricing offers greater computational tractability for block producers when the system is under congestion. We establish this by demonstrating that the relevant revenue maximization problem in that case is considerably more complex under multi-dimensional pricing, reducing a multi-dimensional variant of the knapsack problem to highlight its (fine-grained) intractability (Theorem 4 ). Our work thus rigorously verifies and quantifies key tradeoffs between these pricing schemes. These results underscore the need for further research into alleviating these tradeoffs, specifically through improved price discovery algorithms for multi-dimensional systems or more adaptive weighting mechanisms for one-dimensional schemes (as discussed further in the Conclusions section). Related work. In [ 8 ] , the authors describe a novel price-update procedure for the multidimensional mechanism, and show that it maximizes a utility function that takes in account both the users and the miners as well as other network/protocol related objectives. Further, in a follow-up [ 1 ] , they show that this price-update procedure achieves similar performance to choosing optimal fixed prices offline, i.e., after seeing future traffic. Both works focus on analyzing only multidimensional mechanisms. In [ 7 ] , the authors develop a practical framework for dynamically pricing multiple blockchain resources using linear-quadratic control theory. Further, [ 10 ] , provides a statistical analysis of Ethereum’s multidimensional fee market since EIP-4844’s implementation, i.e., the introduction of a second synthetic resource, blobs. The concurrent work of Lavee et al. [ 11 ] is the one that is most related to ours. In this paper, the authors study approximately how much more throughput is needed by the one-dimensional mechanism in order to be able to accommodate any allocation that is feasible in the multi-dimensional mechanism. They

Read the complete text