On this page (9)
- Platform: .NET 10, MongoDB and Blazor Server
- Your own server or the cloud: two deployment models
- Backups, retention and two-server operation
- User permissions and data visibility scope
- Network requirements
- What the architecture does not cover
- Leaving the service and handing over the database
- IT department questions about the architecture
- Related
Platform: .NET 10, MongoDB and Blazor Server
The application server runs on .NET 10, on Windows or Linux, so it fits into the server room you already have rather than forcing a second platform on you.
The database is MongoDB: one database with around ninety collections. That gives you one place to back up and one place to maintain, with no data spread across separate instances.
The interface is built with Blazor Server and runs as an application in the browser, so nothing is installed or updated on the workstations.
TraxSystem is a modular monolith: 75 modules, one deployment, one database, with the available features switched on by the licence. There are no microservices to orchestrate, no queues to look after and no synchronisation between databases.
Gate states, alerts and events arrive live over SignalR, without refreshing the page.
Number plate recognition runs locally: the recognition engine works on the server or on an edge device at the gate. Object detection also runs locally, on the ONNX runtime. The camera feed is converted and streamed to the browser, and it is not written to disk.
Your own server or the cloud: two deployment models
Both TraxSystem deployment models run the same application, and the licence covers the same list of capabilities. What differs is where the server stands, who maintains it and who carries legal responsibility for the data.
| On your server | In our cloud | |
|---|---|---|
| Where the server stands | in your server room | in Poland |
| Whose infrastructure | yours | ours or a third-party provider’s, your choice |
| Who maintains it | your IT department | we do |
| Updates, backups, monitoring | your responsibility | included in the subscription |
| Stated availability | not applicable | 99% |
| Backups | your responsibility | every 8 hours, continuous with two-server operation |
| Role regarding personal data | you are the sole data controller and no data processor is involved | we are the data processor and a data processing agreement (DPA) is required |
| Operation if the internet link fails | the site keeps working | depends on the configuration, agreed during the analysis |
| Access to devices | directly on the local network | requires a link and configured access |
| Purchase route | one-off or subscription | subscription only |
The two options carry equal weight, and each is chosen for a specific reason. Your own server makes sense when company policy does not allow data to leave the premises, or when the site has to keep working if the internet link goes down. The cloud makes sense when you have no IT department, no server room, or no wish to look after yet another machine.
Backups, retention and two-server operation
A TraxSystem database backup is an archive of dump files, with a configurable scope and frequency. For cloud deployments we take one every eight hours.
Retention and anonymisation run as a separate mechanism: rules with deadlines, executed once a day. Anonymisation cannot be reversed; there is more about it on the data security page.
Device monitoring and diagnostics report when a camera, reader, weighbridge or controller stops responding, before your staff notice.
Two-server operation works in an active–passive arrangement: a database replica set, a floating IP address, a watchdog service on Windows and on Linux, a configuration wizard in the application interface, planned switchover and notifications. The whole set-up is configured from within the application.
User permissions and data visibility scope
TraxSystem keeps permissions, meaning what a user can do, apart from visibility scope, meaning which data they see. Each is set separately, ideally as early as the proposal stage.
There are 217 permissions in 24 groups, and they are granted only through roles. They are deliberately fine-grained: registering an entry, opening the barrier and access to all gates are three separate rights.
Visibility scope has two dimensions: target sites and entry types. The rule is fail-closed: no role means no access to data, not access to everything. The scope also applies to reports and exports, so nobody can get round it by downloading a file.
Network requirements
The TraxSystem server communicates with the devices on site: access controllers, cameras, barriers, weighbridge indicators, kiosks and terminals. IP addressing and firewall rules are agreed during the pre-implementation analysis.
What the architecture does not cover
We do not configure your plant network. TraxSystem monitors devices and links, but it does not manage switches or firewalls.
Leaving the service and handing over the database
With a deployment on your own server, the database is already with you, so there is nothing to hand over.
With a cloud deployment, if you cancel we hand over a copy of your instance’s database. It is the complete database in a format that can be restored, not an export of selected data to a spreadsheet.
IT department questions about the architecture
Do we need a server with a graphics card?
Not for the application itself, because the plate recognition engine runs on the CPU. With more cameras and plate recognition running on the server, a graphics card speeds up image processing, so it is taken into account in the pre-implementation analysis. The deciding factors are the number of cameras and whether recognition runs on the server or on an edge device.
What resources does the server need?
That depends on the number of gates, cameras and workstations. We give the requirements for your specific configuration with the proposal.
Can it run on a virtual machine?
Yes, and most deployments on customers’ own servers run that way.
How do updates work?
An update is a short maintenance window on the server. Nothing needs doing on the workstations, because the application runs in the browser.
Do you provide an API?
Yes. The integrations described on the integrations page are built on it.
Related
Data securityIntegrations and APISupported hardwareLicence and support
Further reading: Your own server or the cloud · Edge vs cloud ANPR · ERP integration: what to require