Project Background

The client, a company specializing in the optimization and control of distributed energy systems, was operating a system based on off-the-shelf industrial hardware, combining an industrial PC with GSM communication modules.

While this approach provided the required functionality, it resulted in a high cost per installation. With a planned rollout of approximately 3,000 units, the existing architecture, costing around £1,200 per installation, was not economically sustainable at scale.

The client’s objective was to evaluate whether a more cost-effective solution could be achieved, either through custom hardware development or by selecting a more suitable industrial controller platform, without compromising system functionality or reliability.

The Challenge

The challenge was to determine the most cost-effective system architecture for large-scale deployment.

At the same time, transitioning to a new platform required careful evaluation of:

• development and unit production costs
• implementation timeline
• risks associated with custom hardware development
• feasibility of migrating the existing software

The client needed a clear, data-driven comparison between a custom-developed solution and available industrial controller platforms in order to make a decision aligned with both cost targets and project timelines.

Our Contribution

Feasibility study and system evaluation

Before writing a single line of code, we ran a structured feasibility study to answer one question the client actually needed answered: build or buy?

We designed a conceptual custom hardware solution and costed it properly – development, validation, regulatory certification, and production at 3,000 units.
The result was roughly £140 per unit, but with a development timeline that carried real risk and an upfront investment that needed to be justified.

In parallel, we evaluated available industrial controller platforms.
The Siemens SIMATIC IOT2020 came out as the most viable alternative: £260 per unit, available off-the-shelf, and with a known software environment based on a Yocto Linux distribution we could work with directly.

The £120 per-unit premium over custom hardware was straightforward to justify: no certification overhead, no hardware re-spins, no timeline uncertainty.
At 3,000 units the difference is significant in absolute terms, but the alternative carries higher risk.

Software platform adaptation and extension

With the platform selected, the work shifted to migrating the existing application from the original industrial PC environment to the Siemens IOT2020.

The base Yocto image required extension before it could support the existing system. Our work focused on adapting and extending this platform to meet both system and client-specific requirements.

Platform adaptation

• migration of the existing application from industrial PC to the new controller
• alignment of system services, dependencies, and runtime behavior with the new hardware

Yocto-based system customization

To support the required functionality, we extended the Yocto-based build system by:

• integrating Modbus communication for interaction with external heating and industrial devices
• adding Perl runtime support, enabling compatibility with existing client-side scripts
• configuring and maintaining the required layers and packages within the Yocto build environment

System delivery and iteration

• development of a custom Yocto image tailored to the target hardware
• validation and testing of the full software stack
• iterative updates of the build system based on integration feedback
• delivery of a reproducible and maintainable software baseline

This approach ensured that the system remained stable, extensible, and aligned with the client’s internal development workflows.

Results

The total investment in feasibility analysis and software migration was approximately €20,000. (around €10k for the feasibility assessment and €10k for porting).

At the same time, the selected solution reduced the per-unit cost from ~£1,200€ to ~£260, resulting in a total cost reduction at scale (3,000 units) from ~£3.6M to ~£780K.
In parallel, the existing application was successfully migrated from the industrial PC platform to the selected industrial controller, without major changes to system functionality.

This created a clear ROI of 14,000%, where the initial engineering effort was negligible compared to the achieved cost savings.

By selecting a proven industrial platform and adapting the software stack accordingly, the client achieved:

• predictable implementation timeline
• reduced implementation risk
• faster deployment