AI-Paging: Lease-Based Execution Anchoring for Network-Exposed AI-as-a-Service

TL;DR

AI-Paging addresses the challenge of providing network-exposed AIaaS with strong execution guarantees by introducing a lease-based control-plane primitive that separates a stable service identity from its execution anchors. It defines a minimal artifact set—AISI, AIST, COMMIT, and EVI—and enforces two invariants: lease-gated steering and make-before-break relocation, ensuring continuity under mobility and dynamic conditions. The authors prototype the design on existing 5G exposure, QoS, and steering mechanisms, mapping artifacts to control/user-plane operations, and demonstrate favorable latency, relocation continuity, and auditable evidence generation. The work offers a practical path toward standardized, auditable, and policy-compliant AIaaS as a network service, with potential adoption through AIMLE CAMARA CAPIF profiles and associated APIs.

Abstract

With AI-as-a-Service (AIaaS) now deployed across multiple providers and model tiers, selecting the appropriate model instance at run time is increasingly outside the end user's knowledge and operational control. Accordingly, the 6G service providers are envisioned to play a crucial role in exposing AIaaS in a setting where users submit only an intent while the network helps in the intent-to-model matching (resolution) and execution placement under policy, trust, and Quality of Service (QoS) constraints. The network role becomes to discover candidate execution endpoints and selects a suitable model/anchor under policy and QoS constraints in a process referred here to as AI-paging (by analogy to cellular call paging). In the proposed architecture, AI-paging is a control-plane transaction that resolves an intent into an AI service identity (AISI), a scoped session token (AIST), and an expiring admission lease (COMMIT) that authorizes user-plane steering to a selected AI execution anchor (AEXF) under a QoS binding. AI-Paging enforces two invariants: (i) lease-gated steering (without COMMIT, no steering state is installed) and (ii) make-before-break anchoring to support continuity and reliability of AIaaS services under dynamic network conditions. We prototype AI-Paging using existing control- and user-plane mechanisms (service-based control, QoS flows, and policy-based steering) with no new packet headers, ensuring compatibility with existing 3GPP-based exposure and management architectures, and evaluate transaction latency, relocation interruption, enforcement correctness under lease expiry, and audit-evidence overhead under mobility and failures.

Figures (6)

• Figure 1: AI Paging with network-exposed service binding.
• Figure 2: System architecture mapping for AI-Paging. The exposure API terminates the intent and returns a stable handle (AISI/AIST). The AI-Paging controller, locating in the service layer (SEAL-like), adjacent to the exposure gateway (CAPIF/NEF-facing), or as part of the operator AIaaS orchestrator. derives ASP, selects candidates, and acquires a time-bounded COMMIT. The user plane installs steering and QoS state only when backed by an active COMMIT and forwards traffic to the admitted execution anchor. Telemetry and anchor events are bound into EVI records for audit-ready attribution.
• Figure 3: Intent-to-serving transaction time across designs.
• Figure 4: Relocation continuity metrics under mobility/churn conditions.
• Figure 5: Recovery success probability versus stress level (offered load/churn).