Rent a LoRaWAN network server
instead of running one.
S: from 129.90 € · M: from 159.80 € · L: from 219.80 € net per month with Essential. Higher service levels follow WZ-IT terms and include the first registered workload. Non-binding monthly indication for a standard environment in Germany. Setup, migration, additional integrations and vendor licences are quoted separately according to scope. Term and notice period are defined in the proposal.
- Dedicated instance, not a shared tenant
- Eight regions from Germany to Australia
- Gateways with no count limit
- contract term according to proposal
WZ-IT assesses your configuration and agrees provisioning and operations with you.
Non-binding monthly indication for a standard environment in Germany. Setup, migration, additional integrations and vendor licences are quoted separately according to scope. Term and notice period are defined in the proposal.
What a LoRaWAN network server does
- 01
A LoRaWAN network has three layers that regularly get confused with each other. At the bottom sit the gateways. They receive the radio signals from your sensors and pass them on. They decrypt nothing and decide nothing; they are antennas with an internet connection.
- 02
The network server is the layer above, and that is where the actual work happens. It takes in packets from all gateways, discards duplicates when the same signal was heard by several gateways, checks frame counters against replay, decrypts the network layer, manages join procedures and session keys, adjusts data rate and transmit power per device through adaptive data rate, and schedules downlinks into the narrow receive windows of the sensors. Without this layer, a gateway is just a device pushing encrypted noise onto the internet.
- 03
Only above that sits the application layer: dashboards, alarms, reports, storage. This separation is why a rented network server combines well with a freely chosen platform. Buying both as one package from one vendor means paying for a platform even when all you needed was the network layer.
Three operating models, compared honestly
Not every project needs its own network server. This table also says when renting one is the wrong call.
| Public community network | Self-operated | Rented network server | |
|---|---|---|---|
| Entry cost | None | Server plus setup | S: from 129.90 € · M: from 159.80 € · L: from 219.80 € net per month with Essential. Higher service levels follow WZ-IT terms and include the first registered workload. Non-binding monthly indication for a standard environment in Germany. Setup, migration, additional integrations and vendor licences are quoted separately according to scope. Term and notice period are defined in the proposal. |
| Ongoing operations | None | Your own team | With the provider |
| Coverage | Wherever the community runs gateways | Your own gateways | Your own gateways |
| Message volume | Fair-use rules of the network | Your own design | Compute S: 2 vCPU, 4 GB RAM, 80 GB storage. M: 4 vCPU, 8 GB RAM, 160 GB storage. L: 8 vCPU, 16 GB RAM, 240 GB storage. Each includes 5 TB of monthly traffic; additional traffic costs €1 per started TB. |
| Data ownership | With the network operator | Entirely yours | Own instance, export at any time |
| Availability commitment | None | Your own | 99.9%, independent of support tier |
| Fits | Prototypes and trials | Teams with their own ops crew | Production without an ops team |
Honestly: if you are building a prototype with five sensors within range of a community gateway, a public network serves you better. And if you already have a team that runs PostgreSQL, Redis and an MQTT broker in their sleep, self-operation beats any rental on paper. Renting pays off in between: production use with your own gateways, but without your own operations crew.
Which engine we operate
We run ChirpStack, the open source LoRaWAN network server under the MIT licence. It supports LoRaWAN 1.0.x and 1.1 as well as device classes A, B and C, plus adaptive data rate, tenant isolation, codec functions in JavaScript and location resolution through TDOA and RSSI.
Choosing an open source engine is a decision against lock-in, and it is the reason this contract does not trap you. Everything you configure here, device profiles, codecs, integrations, runs identically on an instance you operate yourself. A full export is available at any time. With a proprietary platform, the cost of leaving is part of the product, and it is rarely quantified.
chirphost is not an official ChirpStack project product and is not affiliated with it. The project awards no partner status, which is why nobody in this market can show one. What counts instead are verifiable operational facts: which version runs, how promptly updates are applied, and who actually answers when something breaks.
Eight regions, with certification profile
The region determines latency to your gateways and the legal framework for your data. Both are decisions, not defaults.
| Region | Data center certifications | Typical reason |
|---|---|---|
| Germany | ISO 27001, BSI C5 | Default region, EU868, public sector |
| Netherlands | ISO 27001, SOC 2 Type II | EU location, EU868, Benelux latency |
| USA | ISO 27001, SOC 2 Type II | US915, North American latency |
| Canada | ISO 27001, SOC 2 Type II | Canadian data residency |
| Singapore | ISO 27001, SOC 2 Type II | AS923, South East Asia |
| Japan | ISO 27001, SOC 2 Type II | Japanese data residency |
| Australia | ISO 27001, SOC 2 Type II | AU915, Oceania latency |
| India | ISO 27001, SOC 2 Type II | IN865, local data residency |
More on data locations and certifications is on the security and data locations page.
What operated actually means
This is exactly the list used elsewhere as an argument against open source operation. With a rented network server it sits on our side of the invoice.
Setup and sizing
Provision the instance, choose the region, match capacity to device count and sending behaviour. With a setup call on request, included on the Business support.
Engine updates
New ChirpStack versions and operating system patches are applied, security-relevant ones first. LoRaWAN spec changes arrive with the engine.
Database and broker
PostgreSQL, Redis and the MQTT broker are part of the job. In self-operation those three are reliably where it gets unpleasant.
Backups
Automatic backups, kept in the same region as the instance. No silent replication into another jurisdiction.
Round-the-clock monitoring
Availability and system metrics are monitored continuously. What is collected are operational metrics, not your sensor payloads.
Incident response
Essential: Initial technical response to critical incidents (P1) by the end of the next business day, Mon-Fri, 08:00-17:00 CET/CEST · Business: Initial technical response to critical incidents (P1) within 4 service hours, Mon-Fri, 08:00-18:00 CET/CEST · Priority: Initial technical response to critical incidents (P1) within 1 service hour, Mon-Fri, 08:00-18:00 CET/CEST · Production: Initial technical response to critical incidents (P1) within 60 minutes, 24/7, P1 around the clock · Critical: Initial technical response to critical incidents (P1) within 30 minutes, 24/7, including callback, P1 around the clock. Production and Critical require technical acceptance. Critical is capacity-limited. Response targets apply to P1 incidents through the agreed ticket system, not to every request or a guaranteed resolution time. The flexible operations allowance applies to planned work during regular service hours and expires at month end. Operations, monitoring, backups, maintenance and updates within the agreed scope do not reduce it. Additional and incident work costs €120 net per hour in 6-minute increments during regular service hours and €240 net per started hour outside them.
Interfaces and features
The network layer stays open upwards. Which platform runs above it is your call.
MQTT
Uplinks, join events and status messages are published over MQTT. Any MQTT client can subscribe, from Node-RED to your own application.
REST and gRPC
Complete APIs for devices, gateways, tenants and integrations. Anything the interface can do can also be automated.
Codec functions
Payload decoding in JavaScript, across vendors. Sensors from different manufacturers deliver one consistent data format upwards.
Tenant isolation
Isolated tenants with their own users, devices and gateways, inside your own instance. Built for system integrators and platform providers.
Adaptive data rate
The server adjusts data rate and transmit power per device based on link quality. That extends battery life and keeps the radio channel clear.
Location without GPS
TDOA and RSSI derive an approximate position from the network itself, with no GPS module in the device and none of the power draw it costs.
Which gateways work
Any gateway speaking the Semtech UDP packet forwarder or the Basics Station protocol is compatible. That covers the common manufacturers, among them Kerlink, Mikrotik, RAK Wireless, Dragino and Browan.
Basics Station is the newer option and in practice the better one: the connection runs over TLS with certificates rather than open UDP, and the gateway pulls its channel configuration from the server instead of having it maintained locally. During first setup that is exactly where things tend to stall, particularly around certificates and picking the right sub-band. We will look over your shoulder there on request.
Owning a gateway is not a precondition for the instance to run, by the way. If you already operate gateways, you simply point them at the new instance, regardless of which network server they were registered with before.
What operations cost
S: from 129.90 € · M: from 159.80 € · L: from 219.80 € net per month with Essential. Higher service levels follow WZ-IT terms and include the first registered workload. Non-binding monthly indication for a standard environment in Germany. Setup, migration, additional integrations and vendor licences are quoted separately according to scope. Term and notice period are defined in the proposal.
Compute S: 2 vCPU, 4 GB RAM, 80 GB storage. M: 4 vCPU, 8 GB RAM, 160 GB storage. L: 8 vCPU, 16 GB RAM, 240 GB storage. Each includes 5 TB of monthly traffic; additional traffic costs €1 per started TB.
Essential: Initial technical response to critical incidents (P1) by the end of the next business day, Mon-Fri, 08:00-17:00 CET/CEST · Business: Initial technical response to critical incidents (P1) within 4 service hours, Mon-Fri, 08:00-18:00 CET/CEST · Priority: Initial technical response to critical incidents (P1) within 1 service hour, Mon-Fri, 08:00-18:00 CET/CEST · Production: Initial technical response to critical incidents (P1) within 60 minutes, 24/7, P1 around the clock · Critical: Initial technical response to critical incidents (P1) within 30 minutes, 24/7, including callback, P1 around the clock. Production and Critical require technical acceptance. Critical is capacity-limited. Response targets apply to P1 incidents through the agreed ticket system, not to every request or a guaranteed resolution time. The flexible operations allowance applies to planned work during regular service hours and expires at month end. Operations, monitoring, backups, maintenance and updates within the agreed scope do not reduce it. Additional and incident work costs €120 net per hour in 6-minute increments during regular service hours and €240 net per started hour outside them.
Custom infrastructure, high availability and custom MQTT brokers are quoted separately. Setup according to agreed scope; contract term according to proposal.
Hosting by WZ-IT
One team for your hosting and everything running on it.
chirphost is a hosting service by WZ-IT. We connect infrastructure, software and ongoing operations. Timo and Robin discuss your requirements and agree who is responsible for what.
- Directly with Timo and Robin - no sales handoff
- Operations and support under the agreed WZ-IT service levels
- Commission hosting and additional project work separately
A WZ-IT team consultation to get acquainted and discuss your requirements.
More about WZ-IT
Read on
Security and data locations
Where data sits per region, which certifications apply and who actually holds them.
Building a private LoRaWAN network
The practical guide: gateway siting, coverage, frequency plan and the mistakes that cost the most time.
Managed or self-operated
The cost calculation in detail, including the operational tasks routinely underestimated when running it yourself.
Frequently asked questions about renting a network server
What does a rented LoRaWAN network server cost?+
S: from 129.90 € · M: from 159.80 € · L: from 219.80 € net per month with Essential. Higher service levels follow WZ-IT terms and include the first registered workload. Non-binding monthly indication for a standard environment in Germany. Setup, migration, additional integrations and vendor licences are quoted separately according to scope. Term and notice period are defined in the proposal.
Do I need my own gateways?+
For a private network, yes, because the network server only processes what your gateways receive. You can keep existing gateways as long as they support the Semtech UDP packet forwarder or Basics Station. Depending on the environment a gateway covers a few hundred metres in dense urban areas up to several kilometres in open terrain, which is why siting matters more for coverage than sheer numbers.
What is the difference between a network server and an IoT platform?+
The network server is the network layer: it manages devices, session keys, duplicates, data rates and downlinks, and hands decrypted payloads upwards. The IoT platform is the application layer above it: dashboards, alarms, rules, long-term storage. Many vendors sell both as one product. Keeping them separate means you can change platform without touching the network.
Can I migrate from TTN, Actility or another server?+
Yes. Devices, gateways and application integrations can be carried over, in many cases preserving the session so devices do not have to re-join. What definitely has to change is the gateway configuration and the integrations on your application side. For larger networks this can be staged.
Which regions can I run in?+
Eight: Germany, Netherlands, USA, Canada, Singapore, Japan, Australia and India. Germany is the default region. Inside the EU, Germany and the Netherlands avoid a third-country transfer entirely. The choice of region affects both latency to your gateways and the legal framework covering your data.
How are a pilot and contract start agreed?+
Provisioning, any pilot and contract terms are agreed in the proposal. There is no automatic trial subscription.
Rent the network layer, choose the platform freely
Provisioning, any pilot and contract terms are agreed in the proposal. There is no automatic trial subscription.
WZ-IT assesses your configuration and agrees provisioning and operations with you.
Non-binding monthly indication for a standard environment in Germany. Setup, migration, additional integrations and vendor licences are quoted separately according to scope. Term and notice period are defined in the proposal.