On-premise server
Run the whole system inside your own network
The on-premise server runs Noccela's positioning engine and the Noccela Location Intelligence Platform on a server in your facility. Positions are computed and stored inside your network, and the system does not depend on an internet connection.
When to choose on-premise
Most sites start in Noccela's cloud. These are the situations where a server of your own is the better choice.
Your IT policy requires it
There is no internet
Safety-critical use
Local integration
What the on-premise server is
A compact server that runs the whole positioning engine, which turns the measurements of the base stations (UWB anchors) into positions, and the Location Intelligence Platform, with its map, dashboards, workflows, administration and APIs.
The base stations and tags are the same as in a cloud installation. The difference is where the computing happens and where the data lives.
Your network
- Tags
- Base stations and PoE switch
- On-premise server Positioning engine · Location Intelligence Platform · APIs
Users in a browser
MES, WMS, ERP
Internet: optional, not needed for positioning
Cloud or on-premise?
| Noccela cloud | On-premise server | |
|---|---|---|
| Where positions are computed | In Noccela’s cloud | On a server in your network |
| Internet connection | Needed between the site and the cloud | Not needed |
| Where location data is stored | In Noccela’s cloud service | On your server |
| Network set-up | No public IP, static IP or access to your private network | Server and base stations on your local network |
| Hardware on site | Base stations and tags | Base stations, tags and the server |
| Typical fit | Quick starts, pilots, many sites | Strict IT policies, sites without internet, safety-critical use |
Typical server requirements
What a typical on-premise server for one site looks like, so your IT team knows what will run in its network.
| Processing and memory | |
|---|---|
| Processor | Maximum turbo frequency 4.1–5.0 GHz |
| Memory | 16–32 GB (16 GB or 2 × 16 GB DDR4-3200 SODIMM) |
| ECC memory | Optional |
| Storage | |
| Disk | 512 GB–1 TB M.2 SSD |
| Network | |
| Ethernet | 2.5 GbE (2.5 Gb/s Intel Ethernet controller) with Wake-on-LAN |
| Operating system | |
| Supported options | Windows Server 2022, k3OS Linux distribution, or Windows 11 IoT Enterprise LTSC |
| Security and reliability | |
| Trusted platform | TPM 2.0 (Intel Platform Trust Technology) |
| Firmware | Self-healing BIOS |
| Watchdog | Hardware watchdog timer |
| Physical security | Chassis intrusion switch, Kensington lock slot |
| Power | |
| Supply | DC adapter with 100–240 V AC input |
The right size depends on the number of base stations and tags, their update rates and how much history you keep. We confirm the specification for your project.
The security features matter in a production network: a TPM 2.0 chip, a self-healing BIOS and a hardware watchdog that restarts the server if it stops responding.
How an on-premise installation works
-
Size it
Tell us how many base stations and tags you plan and at what update rates. We confirm the server specification.
-
Place it
The server goes in a server room or network cabinet, with power and a connection to the network of the base stations.
-
Connect the base stations
Base stations connect to the server over your local network and calibrate themselves as usual.
-
Connect your systems
Your users open the Location Intelligence Platform in a browser, and your systems take positions and events from the APIs.
Frequently asked questions
Does the on-premise server need an internet connection?
Which operating systems are used?
How do the base stations connect to the server?
Can our own systems use the location data?
Where can I read about security and data protection?
Discuss on-premise with an engineer
Tell us about your IT policy, your network and your site. We will propose the deployment and the server size that fit.