Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Real-Time AI Service Economy: A Framework for Agentic Computing Across the Continuum

Lauri Lovén, Alaa Saleh, Reza Farahani, Ilir Murturi, Miguel Bordallo López, Praveen Kumar Donta, Schahram Dustdar

TL;DR

This article shows that the structure of service-dependency graphs, modelled as DAGs whose nodes represent compute stages and whose edges encode execution ordering, is a primary determinant of whether decentralised, price-based resource allocation can work reliably at scale.

Abstract

Real-time AI services increasingly operate across the device-edge-cloud continuum, where autonomous AI agents generate latency-sensitive workloads, orchestrate multi-stage processing pipelines, and compete for shared resources under policy and governance constraints. This article shows that the structure of service-dependency graphs, modelled as DAGs whose nodes represent compute stages and whose edges encode execution ordering, is a primary determinant of whether decentralised, price-based resource allocation can work reliably at scale. When dependency graphs are hierarchical (tree or series-parallel), prices converge to stable equilibria, optimal allocations can be computed efficiently, and under appropriate mechanism design (with quasilinear utilities and discrete slice items), agents have no incentive to misreport their valuations within each decision epoch. When dependencies are more complex, with cross-cutting ties between pipeline stages, prices oscillate, allocation quality degrades, and the system becomes difficult to manage. To bridge this gap, we propose a hybrid management architecture in which cross-domain integrators encapsulate complex sub-graphs into resource slices that present a simpler, well-structured interface to the rest of the market. A systematic ablation study across six experiments (1,620 runs, 10 seeds each) confirms that (i) dependency-graph topology is a first-order determinant of price stability and scalability,(ii) the hybrid architecture reduces price volatility by up to 70-75% without sacrificing throughput, (iii) governance constraints create quantifiable efficiency-compliance trade-offs that depend jointly on topology and load, and (iv) under truthful bidding the decentralised market matches a centralised value-optimal baseline, confirming that decentralised coordination can replicate centralised allocation quality.

Real-Time AI Service Economy: A Framework for Agentic Computing Across the Continuum

TL;DR

This article shows that the structure of service-dependency graphs, modelled as DAGs whose nodes represent compute stages and whose edges encode execution ordering, is a primary determinant of whether decentralised, price-based resource allocation can work reliably at scale.

Abstract

Real-time AI services increasingly operate across the device-edge-cloud continuum, where autonomous AI agents generate latency-sensitive workloads, orchestrate multi-stage processing pipelines, and compete for shared resources under policy and governance constraints. This article shows that the structure of service-dependency graphs, modelled as DAGs whose nodes represent compute stages and whose edges encode execution ordering, is a primary determinant of whether decentralised, price-based resource allocation can work reliably at scale. When dependency graphs are hierarchical (tree or series-parallel), prices converge to stable equilibria, optimal allocations can be computed efficiently, and under appropriate mechanism design (with quasilinear utilities and discrete slice items), agents have no incentive to misreport their valuations within each decision epoch. When dependencies are more complex, with cross-cutting ties between pipeline stages, prices oscillate, allocation quality degrades, and the system becomes difficult to manage. To bridge this gap, we propose a hybrid management architecture in which cross-domain integrators encapsulate complex sub-graphs into resource slices that present a simpler, well-structured interface to the rest of the market. A systematic ablation study across six experiments (1,620 runs, 10 seeds each) confirms that (i) dependency-graph topology is a first-order determinant of price stability and scalability,(ii) the hybrid architecture reduces price volatility by up to 70-75% without sacrificing throughput, (iii) governance constraints create quantifiable efficiency-compliance trade-offs that depend jointly on topology and load, and (iv) under truthful bidding the decentralised market matches a centralised value-optimal baseline, confirming that decentralised coordination can replicate centralised allocation quality.
Paper Structure (55 sections, 8 theorems, 10 equations, 11 figures, 14 tables)

This paper contains 55 sections, 8 theorems, 10 equations, 11 figures, 14 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Service-Dependency Model
  5. System State
  6. Social Welfare and Objective
  7. Governance, Trust, and Feasible Allocations
  8. Messages, Mechanisms, and Utilities
  9. Mechanism Design
  10. Message Spaces and Direct Mechanisms
  11. Incentive Properties
  12. Structural Regimes for Efficient Allocation
  13. From Per-Task Valuations to Gross Substitutes
  14. Instability in General Dependency Graphs
  15. Restoring Tractability via Architectural Encapsulation
  16. ...and 40 more sections

Key Result

Proposition 1

Let $G_{\mathrm{res}}$ be a rooted tree or two-terminal series--parallel network (under the leaf-block semantics of def:feasibility-region) with node capacities $C_v > 0$. Then:

Figures (11)

  • Figure 1: Model overview: agentic layer, latency-aware valuations, resource dependencies, governance constraints, mechanism, and state evolution.
  • Figure 2: Hybrid architecture. Service-consuming agents trade over integrator-exposed slices (\ref{['prop:encapsulation']}); integrators bid for capacity at local marketplaces, which coordinate service and resource offerings from provider agents; governance flows through integrators (cross-domain) and marketplaces (local). For simple single-domain services, agents may also interact with a local marketplace directly.
  • Figure 3: Structural ablation (Experiment 1): DAG topology $\times$ load. Polymatroidal topologies (linear, tree) maintain zero price volatility; entangled DAGs degrade sharply under load.
  • Figure S1: Experiment 1: Effect of DAG topology and load on median latency (top-left), drop rate (top-right), utilisation (bottom-left), and price volatility (bottom-right). Polymatroidal topologies (linear, tree) maintain stable prices; entangled DAGs degrade sharply under load.
  • Figure S2: Experiment 3: Effect of governance policy (none/moderate/strict) on performance for tree and entangled DAGs under medium and high load. Panels: median latency, drop rate, service coverage, price volatility.
  • ...and 6 more figures

Theorems & Definitions (21)

  • Definition 1: Service-Feasibility Region
  • Proposition 1: Polymatroidal Structure under Tree/SP DAGs
  • proof
  • Lemma 1: GS Valuations under Slice Encapsulation
  • proof
  • Proposition 2: Efficient Mechanism Design under Structured DAGs
  • proof
  • Proposition 3: Polymatroidal Encapsulation of Arbitrary DAGs
  • proof
  • Definition 2: Polymatroid
  • ...and 11 more