5G standalone deployment is changing the boundary between host-network access and MVNO service control. The immediate issue is not radio coverage alone. It is whether contracts, BSS/OSS interfaces and operating procedures can carry differentiated QoS, IMS dependencies, roaming policy and charging events across a 5G core. For wholesale teams, network slicing has become a practical interworking question rather than a future product claim.
SA migration turns wholesale access into a contract redesign
Wholesale access agreements often encode the network architecture that existed when they were signed. An MVNO annex may specify APN structures, PCRF rules, LTE bearer treatment and legacy online charging. The host may now be moving policy, session management and charging into a 5GC, with PCF and converged charging replacing dependencies built around the EPC and OCS. Unless the agreement translates those constructs, technical migration can alter service behaviour without establishing who accepted the change or who pays to correct it.
The central commercial question is what the tenant is entitled to receive. The answer could be a standard data profile, differentiated 5QI treatment, access through a dedicated DNN, exposure to a defined slice, or service aligned only with the host’s retail baseline. These are materially different products. They carry different assurance obligations, interface requirements and potential wholesale charges.
Technical capability does not establish wholesale entitlement. An MNO can run SA, VoNR and production slices for its retail base while withholding some or all of those functions from MVNO tenants. Retail availability may show that the core is ready. It does not settle tenant priority, operational support, performance evidence or commercial access.
Migration mechanics therefore belong in the contract. During dual NSA/SA operation, the parties need customer eligibility rules, device certification criteria and clear dependencies on SIM or eSIM profiles. They also need to allocate remediation costs when an accepted migration cohort loses data, voice or roaming continuity. Emergency calling, lawful-intercept responsibilities, accessibility obligations and existing service levels should not be assumed to transfer automatically to the new core arrangement.
At minimum, an SA migration annex should state:
the access scope, including eligible customer, tariff, device and geographic cohorts;
the QoS treatment and measurable performance commitments for each service class;
the DNN and IMS parameters, ownership boundaries and provisioning dependencies;
the charging records, delivery intervals and reconciliation process;
the domestic and international roaming treatment, including fallback behaviour;
the acceptance criteria, evidence required and rollback thresholds;
the change-control notice periods and allocation of remediation costs.
The issue is already operational. Tier-2 MNO, Western Europe, ~8M subscribers illustrates the contract pattern: a host can advance its retail SA programme while its wholesale estate remains governed by older access schedules. The gap is not resolved by changing a network roadmap. It requires amendments that connect technical delivery to tenant obligations, testing and liability.
QoS, charging and IMS expose the real integration gap
Policy and charging are where architecture changes become revenue discrepancies. Existing wholesale allowance rules may depend on PCRF decisions and real-time OCS balance checks. Moving those controls to PCF and converged charging requires precise mapping of counters, event timing, zero-rating logic, spend controls and out-of-bundle treatment. A session can attach successfully while the wrong party absorbs usage, a customer receives an incorrect allowance decision, or wholesale and retail records diverge.
5QI mapping adds another commercial layer. The MVNO needs defined treatment for default data, IMS signalling, voice media and any differentiated enterprise or FWA service. Generic language promising 5G availability is not sufficient. KPIs should measure successful registration, policy application, latency or throughput where contracted, voice setup performance and charging completeness. Otherwise, the host controls both service treatment and the evidence used to assess it.
IMS and VoNR readiness should run as a separate workstream rather than as a line item in radio acceptance. The host must specify whether the tenant uses host IMS, retains an external IMS dependency, falls back through EPS, or operates under a device-specific eligibility regime. Each model changes fault ownership. It also affects emergency routing, supplementary services, number presentation and continuity when customers move between SA and LTE coverage.
Identity and provisioning create a second class of failure. IMSI ranges must be recognised consistently across subscriber, policy and charging domains. Migration from HLR/HSS dependencies towards UDM/AuC must preserve authentication and service profiles. MNP lookups must remain aligned with routing. eSIM RSP lifecycle events and profile compatibility can determine whether an otherwise eligible handset reaches SA. To the customer and the MVNO care operation, these faults appear as failed activation, missing voice or intermittent service rather than as core migration defects.
Material migration should follow end-to-end evidence across provisioning, data registration, PCF policy application, charging, IMS registration, VoNR calls, emergency-call handling and ticket ownership. Testing a host reference device on a standard retail profile does not validate the tenant chain. The accepted configuration must include the MVNO’s commercial plans, identifiers, SIM estate, interfaces and exception paths.
Observability is consequently a wholesale requirement, even when the host retains operational control of the core. Tenants need usable event records, timestamps that can be correlated across domains, incident classifications and agreed access to performance evidence. Without that visibility, the MVNO cannot distinguish a provisioning error from a policy rejection or charging fault. It also cannot challenge service credits, reconcile usage or give enterprise customers a credible incident account.
Roaming and partner economics will determine the migration timetable
Domestic SA availability does not establish SA roaming availability. Inter-PLMN arrangements, visited-network support, steering rules and settlement may remain anchored in LTE-era models. Some routes will depend on Diameter interworking, while others require support for 5G signalling and new operational procedures. The result is an uneven footprint in which domestic SA, outbound roaming and inbound roaming progress on different schedules.
Multi-IMSI and sponsor arrangements increase that variance. An MVNO using more than one host footprint needs a market-level view of which IMSI, PLMN and roaming-policy combinations support SA, VoNR and emergency calling. A successful configuration on one sponsor does not establish compatibility on another. Steering decisions can also move a customer from an SA-capable path to an LTE-only path without any change visible in the tariff or customer proposition.
The device base narrows the executable population further. SA access, VoNR and eSIM provisioning outcomes vary by handset firmware, carrier bundle, certification status and local emergency-service configuration. A broad launch statement can therefore conceal a much smaller eligible base. Wholesale planning should use actual device and profile distributions, not only the percentage of population covered by the host’s SA radio footprint.
Commercial migration triggers should reflect that operational reality. The parties should agree customer cohorts, minimum successful-registration rates, charging reconciliation tolerances, roaming exception handling and a rollback process. Rollback must return customers to a known policy and charging domain; it cannot simply disable SA while leaving stale provisioning or conflicting eligibility flags in connected systems.
Capital pressure may encourage hosts to present SA migration as a fixed network milestone. Wholesale directors should separate that narrative from an executable tenant plan. A credible roadmap identifies which tenants are in scope, which interfaces must change, when device and SIM certification will finish, how roaming will be staged and what commercial terms apply to differentiated QoS or slice-related services. Dates without those dependencies are planning assumptions, not commitments.
MVNE servicing 12+ tenants in EMEA shows the structural problem. Its hosts may expose SA on different dates, apply different VoNR eligibility rules and offer different levels of policy or slice control. The MVNE must then decide whether to maintain a consistent tenant proposition at the lowest common capability, segment features by host, or fund abstraction and exception handling across its platform. Each choice affects product consistency, support cost and margin.
The SA transition will reward host MNOs that publish a tenant-specific migration path, not simply a retail 5G coverage update. MVNOs should secure contractual clarity on policy, charging, IMS and roaming before SA becomes the default access mode for their customer base. The decisive measure is not whether the core can deliver the service. It is whether the wholesale agreement, integration estate and operating model can deliver it repeatedly, reconcile it accurately and recover it when a dependency fails.