The Bureaucracy of Speed: Structural Equivalence Between Memory Consistency Models and Multi-Agent Authorization Revocation

TL;DR

This paper introduces a new coherence model for identity and access management that significantly reduces unauthorized operations in high-velocity environments by leveraging a state-mapping approach and a novel RCC strategy, validated through extensive simulation.

Contribution

The paper proposes the Capability Coherence System (CCS) and a bounded-staleness semantics framework, providing a safety theorem and a new RCC strategy that outperforms traditional TTL-based revocation methods.

Findings

RCC reduces unauthorized operations by over 120 times in high-velocity scenarios.

Simulation confirms zero bound violations across all runs, ensuring safety.

The approach scales independently of agent velocity, unlike traditional TTL strategies.

Abstract

The temporal assumptions underpinning conventional Identity and Access Management collapse under agentic execution regimes. A sixty-second revocation window permits on the order of $6 \times 10^3$ unauthorized API calls at 100 ops/tick; at AWS Lambda scale, the figure approaches $6 \times 10^5$. This is a coherence problem, not merely a latency problem. We define a Capability Coherence System (CCS) and construct a state-mapping $\varphi : \Sigma_{\rm MESI} \to \Sigma_{\rm auth}$ preserving transition structure under bounded-staleness semantics. A safety theorem bounds unauthorized operations for the execution-count Release Consistency-directed Coherence (RCC) strategy at $D_{\rm rcc} \leq n$, independent of agent velocity $v$ -- a qualitative departure from the $O(v \cdot \mathrm{TTL})$ scaling of time-bounded strategies. Tick-based discrete event simulation across three business-contextualised scenarios (four strategies, ten deterministic seeds each) confirms: RCC achieves a $120\times$ reduction versus TTL-based lease in the high-velocity scenario (50 vs. 6,000 unauthorized operations), and $184\times$ under anomaly-triggered revocation. Zero bound violations across all 120 runs confirm the per-capability safety guarantee. Simulation code: https://github.com/hipvlady/prizm