When software controls hardware: platforms that unlock devices on site

Key takeaways

• A platform can control devices at many locations: after booking and payment, it unlocks the device for exactly the booked slot. A small on-site computer acts as the bridge and talks to the device's controller.
• The computer on site opens the connection to the platform itself and collects its jobs there. So the location's network needs no inbound firewall rules.
• A well-planned system keeps working during short internet outages. The computer buffers operating data and delivers it once the connection is back.
• Remote updates arrive signed, are rolled out device by device and automatically revert to the last working version if something goes wrong. A dashboard shows whether bookings and device data match.

As of September 2026

You operate devices at several locations, such as treatment units in partner studios, car wash bays or fitness equipment. Your customers should book and pay online, and the device on site should run exactly when paid for. This guide explains the building blocks between the click on "Pay" and the running device, and shows how they work together reliably.

From booking to a running device

The process has four steps:

1. Book and pay: The customer chooses a location and a time slot and pays online.
2. Create the release: The platform turns this into a job for exactly this device and time slot.
3. Collect the job: A small computer next to the device collects the job and passes it on to the device's controller.
4. Run and report: The device starts, and the computer reports back what actually happened.

What a PLC is and what role it plays

Many machines and devices are run by a programmable logic controller, or PLC. Siemens describes the principle in its introduction to the SIMATIC S7-1200 controller: once the program is loaded, the CPU, meaning the controller's processor, monitors the inputs and changes the outputs according to the logic of the user program. Inputs include switches, for example, and outputs include switches for motors (motor contactors) or valves.

For a platform, the division of labor is key: the PLC governs how the device works. The platform decides whether and when it may run for a customer.

The computer on site as a bridge

A PLC is built for the reliable cycle at the machine, not for bookings, accounts and payments on the internet. That is why a small computer on site, for example a mini PC, sits between the device and the platform. Experts call it an edge device: it works at the edge of the network, right next to the device.

Microsoft's documentation on Azure IoT, its services for connected devices (Internet of Things, IoT), calls this the edge-connected pattern. According to Microsoft, it suits industrial protocols, fast on-site processing and environments where security requirements keep devices from connecting directly to the internet, among others. The computer maintains the connection to the devices, receives their data and forwards commands from the cloud to the right device. Older devices without their own encryption or internet access can also be connected securely through such a computer (gateway), Microsoft says.

Why the computer on site collects its own jobs

A device in a studio or hotel sits in someone else's network, behind the location's router and firewall. For the platform to reach it from outside, the local IT would have to allow inbound connections, at every location. The reverse approach is simpler and more secure: the computer on site opens the connection to the platform itself, much like a browser opens a website.

Microsoft gives exactly this reasoning for its network sensors in industrial environments (Defender for IoT): the sensor initiates all connections itself. This protects devices in the internal network from unsolicited inbound connections, and no inbound firewall rules are needed. AWS likewise notes for its remote access to IoT devices that existing inbound firewall rules can stay unchanged.

A platform therefore places every job in a queue for the respective device. With Microsoft's IoT Hub, a message stays there until the device collects it and confirms it as completed. According to Microsoft, this protects against connectivity and device failures, and every message also carries an expiration time. This fits bookings well: a release applies to a specific time slot and expires afterwards.

Telemetry: what the device reports back

Telemetry refers to the operating data a device continuously sends to the platform, for example whether the device is running, how long it ran and whether it reports a fault. Microsoft lists receiving sensor data, sending commands to individual devices and monitoring their activity as typical tasks of IoT services.

For a booking platform, telemetry shows whether a paid service actually took place. Faults become visible early, and maintenance can be planned around actual running time.

What happens when the internet goes down?

Short connection outages can be planned for from the start. Microsoft shows how for its edge software Azure IoT Edge: after an initial synchronization, the computer and the connected devices keep working without internet. Operating data is stored locally and delivered in the same order once the connection is back. How long messages may wait is configurable, and the default is two hours.

In terms of the process, one question needs answering first: what should apply without a connection? One option is to transfer releases to the computer on site well before the booked slot. Microsoft explicitly describes that a device can first store a job locally and then confirm it as completed. So an already booked slot runs even during a short outage, and the data follows later.

Remote updates: signed, step by step, with fallback

The software on the computer on site keeps evolving. Three principles make sure updates arrive reliably, without anyone having to visit every location:

• Signed: An update carries a digital signature, a kind of tamper-proof seal, that the device checks before installation. Secure Boot works on the same principle: when the system starts, the device's built-in base software (firmware) checks the signature of every piece of boot software and only boots if it is valid. Microsoft's Device Update service additionally sends the device a signed list with the checksums of all files (update manifest), so the device can check every download against it. Checksums are a kind of fingerprint of each file.
• Step by step: An update reaches devices gradually, not all at once. AWS IoT Jobs supports such a staged rollout and can be configured to stop it as soon as a defined share of devices reports failures.
• With fallback: If the new version fails to start properly, the device automatically returns to the previous one. Android does this for system updates and only makes an update permanent after a successful boot.

The dashboard: bookings and devices at a glance

The dashboard brings bookings, payments and device data together. For a reliable view of every device, Microsoft describes the so-called heartbeat pattern as one option: the device checks in at fixed intervals, even when it has nothing new to report. The platform tracks each device's last heartbeat, and if none arrives in time, it assumes a problem.

The greatest benefit comes from matching the data. Does every paid booking match a release and running time on the device? A good dashboard answers this question at a glance and flags discrepancies on its own.

In practice

For Wasserplasma Experience, we developed a booking and settlement platform that connects booking, payment and the devices in the partner studios. A mini PC at every device talks directly to its Siemens S7 PLC. It releases the device for the booked slot and reports operating data back to the platform.

The software on the mini PCs receives signed updates that are rolled out device by device. If something does not work after an update, it automatically reverts to the last working version. Our article Real-time on platforms shows how the back-office live dashboard reveals discrepancies between booking and device. We tested the entire chain, from the booking through the PLC to the payout to the partner studios, on a real device.

David Sroka of njudev puts it this way: "Booking is the easy part. The real work starts where software meets hardware and money: a device that only runs when it's paid for." Our guide Paying out partners explains how the platform then distributes the money to partners.

Frequently asked questions

Does the location have to rebuild its network?
The computer on site needs internet access at the location. Inbound firewall rules are not necessary, because it opens the connection to the platform itself. In tightly secured corporate networks, the IT team adds the platform's address to the list of allowed destinations, as Microsoft also provides for corporate firewalls and intermediary proxy servers.

Can existing devices be connected?
That depends on the controller. What matters is whether it can exchange signals for release and feedback with a computer on site. Microsoft describes how a gateway can securely connect even older devices without internet capability of their own. Talk to the manufacturer of your device early, because they know the controller's program best.

Do I need a separate computer for every device?
Not necessarily. One computer can serve several devices on the same network, and Microsoft's Azure IoT Edge allows up to 100 child devices per gateway by default. At Wasserplasma, each device has its own mini PC. Which option fits depends on how many devices share a location and how independently of each other they should run.

Can I start without connecting the devices?
Yes. Booking and payment can be built as a first stage, and until the devices are connected, the team on site can unlock them as usual. The devices follow as a separate stage. How to start with a lean MVP is covered in our guide What does your own platform cost?

Conclusion

A platform that unlocks devices on site connects three worlds: booking and payment on the internet, a small computer at the location and the device's controller. Clear principles make this interplay reliable. The computer collects its jobs itself, keeps working during short outages and receives signed updates with automatic fallback. Telemetry and the dashboard show at any time whether every paid booking has actually reached the device.

Want to connect devices to your platform?

Describe your idea in our AI project estimator. You will receive a free estimate by email, and we will get back to you within 24 hours.

Learn more about our services for platforms and marketplaces and SaaS products. Or simply write to us at our email address.

Sources

1. Siemens AG: "SIMATIC S7-1200. Erste Schritte mit der S7-1200" (Getting Started, German edition), edition 11/2009, A5E02486792-01. https://cache.industry.siemens.com/dl/files/875/39644875/att_76193/v1/s71200_getting_started_de-DE_de-DE.pdf
2. Microsoft Learn: "Introduction to Azure IoT", April 15, 2026. https://learn.microsoft.com/en-us/azure/iot/iot-introduction
3. Microsoft Learn: "Secure your IoT solutions", July 7, 2026. https://learn.microsoft.com/en-us/azure/iot/iot-overview-security
4. Microsoft Learn: "Methods for connecting sensors to Azure", Microsoft Defender for IoT, February 23, 2023. https://learn.microsoft.com/en-us/azure/defender-for-iot/organizations/architecture-connections
5. Amazon Web Services: "AWS IoT secure tunneling", AWS IoT Core Developer Guide. https://docs.aws.amazon.com/iot/latest/developerguide/secure-tunneling.html
6. Microsoft Learn: "Understand Azure IoT Hub cloud-to-device messaging", February 11, 2026. https://learn.microsoft.com/en-us/azure/iot-hub/iot-hub-devguide-messages-c2d
7. Microsoft Learn: "Operate Azure IoT Edge devices offline", March 2, 2026. https://learn.microsoft.com/en-us/azure/iot-edge/offline-capabilities
8. Microsoft Learn: "Secure boot", March 10, 2026. https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/oem-secure-boot
9. Microsoft Learn: "Security for Device Update for Azure IoT Hub", August 19, 2022. https://learn.microsoft.com/en-us/azure/iot-hub-device-update/device-update-security
10. Amazon Web Services: "How job configurations work", AWS IoT Core Developer Guide. https://docs.aws.amazon.com/iot/latest/developerguide/jobs-configurations-details.html
11. Android Open Source Project: "Virtual A/B overview". https://source.android.com/docs/core/ota/virtual_ab
12. Microsoft Learn: "Monitor device status", Azure IoT Hub, January 5, 2024. https://learn.microsoft.com/en-us/azure/iot-hub/monitor-device-connection-state