Swisscom Global Public Cloud AWS Tiger Team did a deep dive into Industrial Internet of Things. We built an IIoT PoC which shall prove the feasibility of some crucial technical challenges for a future platform which has tens of million machines in mind and multiple geographic locations as a requirement. Additionally, we need to be able to scale with a hundred thousand users and ten thousand things producers. We explored the open standard OPC UA, Edge Gateways and analytics streams. We learned a few lessons for the new manufacture with IIoT on the global public clouds that we want to share.
An Overview of IoT
Internet of Things (IoT) has been around for a while, the term is invented already in the year 1999, with first promotion of RFID. Maybe you are thinking about Amazon Echo, Google Nest or Philips Hue. A fit bit or wearable is also a type of IoT. We had a chance to have a close look at a different kind of IoT. Industrial IoT – a subset of IoT with focuses specifically on industrial applications such as manufacturing or agriculture. It deals with mainly two areas, increases efficiency and improving health and safety. An example of the potential of IIoT is predictive maintenance. A broken machine in a manufacturing process can mean millions of dollars in lost productivity while production halts to fix the issue.
Every IoT solution has three parts. The Edge Tier with things that sense and act. The Platform Tier to ingest, store and compute data, as well as to manage the IoT devices. Last part is the Enterprise Tier with Intelligence based on the data analytics. Such as insights and logic to action. There is no exception in IIoT. Additionally, industry adds extra demand for security and interoperability.
In the latest Gartner report, they predict 25.1 billion IoT units by end 2021. 80% of PoCs do not convert into full implementation with the consequent payback. The convergence of IT and OT (operation technology), along with increased adoption of IoT by the line of businesses, has raised interoperability, integration and security issues that enterprises cannot avoid.
When the different vendors manufacture machines, they agreed to one open standard – OPC Unified Architecture. OPC UA can be used independently of platform and manufacturer. Increase machine to machine and machine to cloud communication. It is one of the building blocks of Industrial 4.0.
Proof of Concept (PoC) Connecting customer machines
Before we start getting into the architectural details I just wanted to recall the idea of this PoC: It should deliver us important insights about some technical challenges to be expected for a future (SaaS-)platform which shall be able to support millions of devices, thousands of users with a global reach. The PoC shall deliver input as one of the decision criteria if and how the mentioned platform shall be built.
In our PoC, we are connected to the OPC UA server in two different ways. We had two customer machines connected to Swisscom IoT Cloud by using Telit solution with our certified gateway device. We know this solution is comprehensive. It can do many things, just to name a few advantages, it includes mobile connectivity with a machine to machine sim card. A broad M2M services interface including to AWS and Azure. Triggers for “on the edge actions”. High transaction reliability and so on.
The second one is through AWS IoT. We run the device gateway with AWS Greengrass Core on the raspberry pi. An OPC UA client has been developed and deployed to the Greengrass Core as a long-running lambda. The data collection flow is same as Telit solution, but then the similarities end as AWS IoT has an entirely different philosophy. It is a framework or tool sets. It does not give us anything plug and play. We have to put all puzzle pieces together, sometimes even build a bit of a puzzle. With AWS IoT, security & scalability by design and large industrial adoptions and last not least, highly active feature development.
The lesson learned
My personal lessons learned are divided in more organisational/processual topics on the one hand and technical topics on the other hand
- An organization must reach a certain maturity level across all stakeholders for successful IoT project implementation. E.g. IT and OT with different focuses must converge into a shared responsibility model.
- Managing different viewpoints is tough. IIoT touches multiple various viewpoints including data architecture, device management, edge aggregation. Multiply with software life cycle, connectivity, manageability, and security. Therefore enough time should be taken in consideration for these alignment tasks – geographically working closely together could also help a lot.
- Take the data ownership issue in the early stage. More data collected, harder it is to redistribute the permissions.
- Build the architecture for IIoT solution early on. Use reference architectures from international bodies, such as IIC or Industrie 4.0, rather than single vendor architectures.
- Create and implement an end-to-end IoT viewpoint, and a testing plan is the key to success!
- Side note: If you want to build a complete new platform as it was the case in this project: Choose a partner to assist in building the business case/business story first before partnering for products and technologies to deliver it. The business case comes shall guide the discussions and not technology.
Do you have an IoT or IIoT project which you want somehow to combine with the global public clouds and need some support for it?
Swisscom Global Public Cloud Team can help you dive into both AWS IoT and Azure IoT and more. Get in touch with our experts!