[Paper Review] A Market-Based Framework for Multi-Resource Allocation in Fog Computing

Duong Tung Nguyen, Long Bao Le+1 more|arXiv (Cornell University)|Jul 25, 2018

IoT and Edge/Fog Computing48 references4 citations

TL;DR

This paper proposes a market-based framework for multi-resource allocation in fog computing, modeling services as budget-constrained buyers and fog resources as divisible goods to achieve a fair and efficient market equilibrium (ME). The framework ensures non-wasteful, frugal, and fair allocations through a convex optimization formulation and a privacy-preserving distributed algorithm that protects sensitive data without sacrificing accuracy.

ABSTRACT

Fog computing is transforming the network edge into an intelligent platform by bringing storage, computing, control, and networking functions closer to end-users, things, and sensors. How to allocate multiple resource types (e.g., CPU, memory, bandwidth) of capacity-limited heterogeneous fog nodes to competing services with diverse requirements and preferences in a fair and efficient manner is a challenging task. To this end, we propose a novel market-based resource allocation framework in which the services act as buyers and fog resources act as divisible goods in the market. The proposed framework aims to compute a market equilibrium (ME) solution at which every service obtains its favorite resource bundle under the budget constraint while the system achieves high resource utilization. This work extends the General Equilibrium literature by considering a practical case of satiated utility functions. Also, we introduce the notions of non-wastefulness and frugality for equilibrium selection, and rigorously demonstrate that all the non-wasteful and frugal ME are the optimal solutions to a convex program. Furthermore, the proposed equilibrium is shown to possess salient fairness properties including envy-freeness, sharing-incentive, and proportionality. Another major contribution of this work is to develop a privacy-preserving distributed algorithm, which is of independent interest, for computing an ME while allowing market participants to obfuscate their private information. Finally, extensive performance evaluation is conducted to verify our theoretical analyses.

Motivation & Objective

• To address the challenge of allocating multiple, heterogeneous resources (CPU, memory, bandwidth) across capacity-limited, distributed fog nodes to competing services with diverse preferences and priorities.
• To design a resource allocation mechanism that ensures high system utilization while maintaining fairness and efficiency, avoiding the pitfalls of welfare maximization and max-min fairness.
• To develop a privacy-preserving distributed algorithm that enables market participants to conceal private information—such as utility functions and budgets—without compromising solution accuracy.
• To establish theoretical guarantees for the market equilibrium, including envy-freeness, sharing-incentive, and proportionality, under realistic satiated utility functions.
• To extend General Equilibrium theory to practical fog computing settings with multiple nodes, multiple resource types, and finite demand constraints.

Proposed method

• Formulates the multi-resource allocation problem as a Fisher market with divisible goods, where services act as buyers with fixed budgets and fog nodes supply multiple resource types.
• Introduces a convex optimization program whose optimal solutions correspond exactly to non-wasteful and frugal market equilibria, ensuring fairness and efficiency.
• Employs a distributed, parallel algorithm based on dual decomposition and subgradient methods to compute the equilibrium without requiring agents to reveal private data.
• Designs a non-cryptographic privacy-preserving mechanism that obfuscates private inputs (e.g., utility functions and budgets) during iterative computation.
• Uses a dual ascent method with perturbed gradients to prevent inference of private data while maintaining convergence to the optimal solution.
• Validates the theoretical results through extensive simulations, demonstrating high resource utilization and fairness under diverse workloads.

Experimental results

Research questions

• RQ1 Can a market equilibrium model be effectively applied to multi-resource allocation in heterogeneous, capacity-limited fog computing environments?
• RQ2 How can fairness and efficiency be simultaneously achieved in a multi-resource, multi-node fog computing setting with finite service demands?
• RQ3 What properties does the resulting market equilibrium possess in terms of envy-freeness, proportionality, and sharing-incentive?
• RQ4 Can a distributed algorithm compute the market equilibrium while preserving the privacy of service-specific data such as utility functions and budgets?
• RQ5 How does the proposed framework compare to existing approaches like DRF, auctions, and welfare maximization in terms of fairness and utilization?

Key findings

• The proposed market equilibrium is non-wasteful and frugal, ensuring that no resources are left unused and no service is allocated more than its demand.
• All non-wasteful and frugal market equilibria are optimal solutions to a convex program, guaranteeing global optimality and convergence.
• The equilibrium exhibits strong fairness properties: it is envy-free (no service prefers another’s allocation), satisfies sharing-incentive (services prefer joining the market), and ensures proportionality (each service receives at least its fair share).
• The distributed algorithm converges to the optimal market equilibrium without requiring agents to disclose private information, offering a practical solution for real-world deployment.
• Performance evaluation confirms high resource utilization and fairness across diverse workloads, outperforming traditional methods like welfare maximization and max-min fairness in balanced efficiency and equity.
• The privacy-preserving mechanism is non-cryptographic and computationally efficient, avoiding the accuracy-privacy trade-offs of differential privacy and the high costs of homomorphic encryption or garbled circuits.