A charger can be physically installed, powered and visible to motorists yet still fail commercially because it cannot reliably communicate. This EV charging connectivity example follows a realistic multi-site deployment and shows why connectivity is not a minor technical line item. It is the operational path between a charge point, the charging management platform, payment services, support teams and, in some cases, energy systems.
Consider an operator deploying 250 AC and DC charge points across retail car parks, council locations and a fleet depot. The business case assumes high availability, remote fault resolution, tariff changes without site visits, card and app payments, and data accurate enough for customer support and revenue reconciliation. Every one of those outcomes depends on decisions made before the first charger is commissioned.
EV charging connectivity example: the operating model
In this example, each charge point has a cellular modem and an embedded SIM or eSIM. The charger establishes an outbound encrypted connection over 4G to a charge point management system using OCPP, typically OCPP 1.6J today, with OCPP 2.0.1 considered where the charger, platform and intended functions genuinely support it.
The charge point sends status updates, meter values, transaction records and fault notifications. The management system sends commands back: start or stop a session, change a tariff, reset a controller, apply a configuration change or trigger a firmware update. Payment terminals, if fitted, may use separate connectivity, although some designs route them through the charge point controller. That distinction matters for both PCI scope and fault isolation.
The apparent simplicity hides several dependencies. The charger manufacturer must implement OCPP properly. The cellular provider must offer coverage at the actual cabinet or parking bay, not just a favourable map. The management system must maintain persistent sessions at scale. The SIM profile, APN, IP addressing model and firewall rules must allow the intended traffic without making remote access unnecessarily exposed.
If any one of these pieces is treated as somebody else’s problem, the operator inherits expensive ambiguity when chargers go offline.
Start with the service, not the SIM
A common procurement mistake is asking for “connectivity for chargers” before defining what the charger estate must do. A low-utilisation destination charging network has different requirements from a motorway rapid-charging operation, an electric bus depot or a private workplace estate.
For publicly accessible charging, the usual requirements include remote monitoring, transactions, tariffs, customer support evidence and roaming or payment integration. At a depot, smart charging and energy management may be more important than payment acceptance. A failed connection at 2am could prevent vehicle scheduling, even though no member of the public ever sees the charger.
The right design begins with practical questions. How long can a charger be offline before it creates a material service failure? Must charging continue if the central platform is unavailable? Which commands need near-real-time delivery? Who needs remote access, and for what purpose? Is the operator prepared to send an engineer for a firmware update, or must updates be controlled remotely?
These answers determine the connectivity architecture. They also stop a project from buying costly capabilities that have no operational value.
Cellular is usually the default, but not automatically the answer
Cellular connectivity is often the most workable option because charge points are distributed, installation programmes need pace, and fixed-line lead times are unpredictable. 4G is sufficient for the vast majority of charger telemetry and OCPP traffic. 5G can be relevant at major hubs, for high-bandwidth local services or where it is part of a wider private-network design, but it is rarely needed simply to exchange charger messages.
Coverage must be tested at the installed equipment location. A signal survey from the car park entrance is not enough. Metal cabinets, underground plant rooms, concrete structures and the charger enclosure itself can materially alter radio performance. For critical sites, an external antenna, a different network or a dual-SIM design may be justified.
Dual-SIM is not a universal cure. It increases hardware, contract, management and diagnostic complexity. It makes sense where downtime has a high operational cost, where a site is known to have weak single-network coverage, or where resilience is an explicit customer commitment. For a modest AC estate with tolerant service levels, a well-selected single-network design can be the sensible commercial decision.
Public internet, private APN and IP design
Many charge points work perfectly well using an outbound encrypted connection over the public internet. This is often the quickest and least expensive approach. Since the charge point initiates the connection, it does not necessarily need a publicly reachable IP address.
A private APN can provide more control over traffic routing and separation from general internet use. It may support fixed private addressing, controlled access to a cloud environment or connection into an enterprise network. That can be appropriate for a fleet depot, critical infrastructure estate or a large operator with clear security governance.
However, a private APN does not make a poor design secure by itself. It can also create unnecessary cost and operational friction if the charge point management system is cloud-based and the integration path has not been agreed in detail. The decision should be based on data flows, threat model, support model and contractual responsibility – not on a generic preference for “private”.
The same discipline applies to static IP addresses. They are useful when a platform or security policy requires predictable addressing, but they enlarge the attack surface if remote access is enabled carelessly. In most cases, remote vendor support should be tightly controlled, logged and time-limited rather than permanently open to every installed device.
The overlooked layer: SIM and lifecycle management
A SIM is not just a route onto a mobile network. It is an asset that needs ownership, inventory control, activation rules, tariff visibility and a defined process for faults, replacement and end of life.
In a 250-charger estate, poor SIM management soon becomes an operational problem. Teams lose sight of which ICCID is in which charger, a replacement controller inherits the wrong subscription, data limits create unexpected restrictions, or a terminated contract silently takes a site offline. Where a charging manufacturer supplies connectivity as part of a bundled offer, the operator should still establish who owns the data, how usage is reported, whether SIMs can be transferred, and what happens if the manufacturer relationship changes.
eSIM can improve manufacturing logistics and allow a profile to be selected later, but it is not automatically more flexible in practice. The practical value depends on the eSIM platform, the available profiles, commercial agreements and whether remote profile changes are supported in the deployed hardware. For a fixed charger with a long expected life, contractual portability and operational control can matter more than the form factor.
Commissioning is where good designs are lost
A charger can pass an installation test and still be badly commissioned. The field process must verify more than signal bars. It should confirm the correct SIM identity, network registration, APN settings, secure connection to the intended platform, OCPP boot notification, accurate charger serial number, time synchronisation and a successful test transaction.
Remote configuration must be treated carefully. A firmware update or parameter change can repair a recurring defect, but a failed update can remove a charger from service. Staged deployment, rollback capability, maintenance windows and clear approval ownership are far more valuable than a theoretical ability to update every device at once.
This is also why the charging platform, charger supplier, connectivity provider and installer need a shared fault workflow. When a charger is offline, the support team needs to distinguish power loss, modem failure, radio coverage, SIM status, DNS or certificate issues, platform connection faults and charger-controller errors. Without agreed diagnostics, each supplier points elsewhere and the site remains unavailable.
Design for recovery, not just launch day
The strongest connectivity programmes assume that devices will fail, networks will change and suppliers will be replaced. They retain accurate records, make monitoring actionable and avoid dependencies that cannot be unwound.
For this EV charging connectivity example, that means setting measurable service levels: expected connection availability, alert thresholds, time to diagnose, time to restore, and a clear escalation route across all parties. It means collecting enough telemetry to spot recurring issues by location, charger model, firmware version and mobile network. It also means reconciling charging sessions against payment and meter data, rather than assuming a completed OCPP message equals collected revenue.
Commercially, the operator should examine more than the monthly connectivity price. A cheaper SIM estate can become expensive if it leaves no route for migration, provides poor visibility or creates long support queues during an outage. Conversely, an over-engineered private solution can delay rollout and consume budget that would be better spent on site power, support capability or spare parts.
The useful question is not whether a charger has connectivity. It is whether the organisation can see, control, support and recover that charger at the pace its customers and operations require. Get that answer right early, and connectivity becomes a managed part of the charging service rather than the reason it fails.
