Traffic pattern
Peak-free and steady while the line runs
- Uninterrupted telemetry stream
- No surge expected
- Capacity planned for continuity
Solution scenario
A reference architecture for production telemetry and industrial data
A production line does not stop, so neither does its data. This page shows how the one-way transfer boundary between plant and cloud and the peak-free continuous stream are planned.
- Primary decision
- Keeping the flow from plant to cloud one-way
- Load pattern
- A peak-free, uninterrupted and predictable stream
- Data layer
- The time series is retained in raw form
Overview
How does industrial telemetry shape the architecture?
What sets this scenario apart is two separate domains. Plant-side systems are attached to the production line; the cloud side processes, stores and reports on the data. The transfer is one-way: data leaves the plant for the cloud, and no control traffic is opened in the reverse direction. While the line runs, data arrives at a steady rate, so capacity is planned for continuity. The raw time series is retained. Record order is preserved.
Typical workload profile
No peak, and no pause either
While the line runs, data arrives at a steady rate, so the capacity plan is built on continuity rather than on a surge scenario. Individual records are small, but because the stream never stops the total keeps growing. The third axis of the profile is not latency but the preservation of record order.
Data volume
Small records, uninterrupted accumulation
- The time series is retained in raw form
- Summary reports produced separately
- Retention tied to the line’s lifetime
Latency and compliance
Ordering integrity matters more than instant latency
- Record order is preserved
- The transfer is one-way
- The plant network stays a separate domain
Measurement comes from the line’s own systems. The cloud side processes, stores and reports on the stream.
Reference architecture
Two domains with a single crossing
The diagram is built as two side-by-side domains: the plant on the left, the cloud on the right. The transfer boundary between them is the single crossing and carries an arrow only from left to right. Plant systems do not act on a command arriving from the cloud.
PortvMind Cloud services used
Bridge, stream, store and report
The router owns the transfer boundary and its one-way rule, VKE the continuously running stream processing, Object Storage the raw time series, and Compute the summary and reporting jobs. VKE is used here not for environment separation but to keep stream components continuously available — a different role for the same service.
VKE (Kubernetes)
Stream processing components run in a cluster so they stay continuously available.
Object Storage
The raw time series is retained as-is for questions asked of the past.
Compute
Summary and reporting jobs run on a resource separate from the stream path.
Router
The transfer boundary between plant and cloud and its one-way rule are defined here.
Each service page links back to this scenario; the link is generated from the same list in both directions. All Services
Evaluation criteria
Direction, continuity, raw data and domain
The table covers transfer direction, the capacity plan, raw-data retention and the separation of the two domains. The transfer is one-way, capacity is planned for continuity, the raw time series is retained, and plant and cloud stay in separate domains.
| Criterion | Question asked | Answer in this scenario |
|---|---|---|
| Transfer direction | Is control traffic opened from cloud to plant? | No — the transfer is one-way; plant systems do not act on commands from the cloud. |
| Capacity plan | Is capacity planned for a surge? | No — the stream has no peak; the plan is built on continuity and ordering integrity. |
| Raw data retention | Is only the summary retained? | No — the raw time series is retained; otherwise no new question can be asked of the past. |
| Domain separation | Are the plant network and the cloud project one domain? | No — two separate domains, with the transfer boundary as their only crossing. |
Operating the plant systems belongs to the organisation. The cloud side covers the one-way transfer, stream processing, the time-series store and reporting.
Related guides and documentation
Reading on clusters and the time-series store
What Kubernetes does and how object storage is structured are covered in the guides; provisioning infrastructure as code and bucket operations are handled step by step in the documentation. Protocol and collector choices on the plant side belong to the organisation’s own systems.
Guides
- What Is Kubernetes? How It Works and When to Use It
Kubernetes is an open-source orchestration platform that automates the deployment, scaling and lifecycle of containerized applications. It does not write application code and is not a container runtime by itself. Instead, it continuously compares your declared desired state with the actual state of a cluster, schedules pods, replaces failed workloads and routes service traffic to available replicas.
- Object Storage Architecture: How It Differs from Block Storage
Object storage keeps data in buckets as independent objects containing the data, a unique key and metadata rather than filesystem blocks. Applications typically access these objects through an HTTP-based, S3-compatible API. Block storage presents a disk volume to a server, where the operating system runs a filesystem or database. The object model is evaluated for media, backups, archives and data lakes, while the block model suits operating-system disks, databases and workloads that need low-latency random reads and writes.
Guides
- Infrastructure Management with Terraform
Go directly to the setup and infrastructure automation guide on Discuss.
- Object Storage and the S3 API
The published Discuss guide for buckets and file usage.
Other solution scenarios
- AI and Data
The training window, checkpoint volume and inference path are planned separately.
- SaaS and Startups
Environment separation, release promotion and a persistent volume per tenant.
- Finance
An intraday transaction band, an end-of-day batch band and a retention period.
Let us build the transfer boundary and the retention plan together.
Share your line count, telemetry volume and the collector setup on the plant side; in the architecture consultation we plan the one-way bridge, the stream processing layer and the time-series store together.