When Campofrio Food Group’s 17-year-old factory in Burgos, Spain, famously burned to the ground, the multinational meat processor turned tragedy into opportunity. With an eye on digital transformation, Campofrio rebuilt the facility as a connected factory, powered by the Internet of Things (IoT). This allowed the state-of-the-art, greenfield meat packing plant to automate processes and provide real-time data on materials, equipment and workers to uncover new business value.
Such “blank canvas” opportunities to bring a factory into the digital age from the ground up don’t happen very often. Most IoT projects are implemented in existing, brownfield environments with traditional legacy systems, requiring an incremental approach. The goal of these gradual integrations has been to optimize or automate processes, gain some efficiencies, and move onto the next low-hanging fruit. But this approach will not work if businesses want to capture IoT’s true transformational value: the creation of new business models, new revenue streams, new products and new markets.
Why an incremental approach to IoT doesn’t work
Initially, the incremental approaches with legacy infrastructure made sense, given the novelty and complexity of IoT and the cost of an infrastructure overhaul. However, this approach no longer applies because of recent advances in technologies powering IoT’s impact, such as artificial intelligence (AI), fog computing (distributed cloud), blockchain, big data analytics and more. Often disparate and rigid, legacy systems simply aren’t capable of anchoring such robust IoT operations – they lack the latency, bandwidth, security and interoperability required for IoT success. In other words, it no longer makes sense to apply IoT solutions on top of outdated infrastructures – you must redo the entire system, workflow and underlying processes to compete in today’s digital age.
Existing infrastructures aren’t the only challenges inhibiting the full potential of IoT. While strides have been made, we need further progress around standards that enable interoperability of technologies with each other and their networks. Such standards help horizontal technologies and protocols to meet legacy Operational Technology (OT) requirements in vertical markets and lines of business. Internal cultural differences (especially between IT and OT), a lack of skill sets and undeveloped supporting ecosystems are also factors hindering the infrastructure transition.
There are numerous challenges to overcome, but they’re not insurmountable. And, you don’t have to do it all at once to take full advantage of IoT. It is possible to take a practical, step-by-step approach based on your resources and appetite for transformation. One thing is clear, though: businesses need to take a revolutionary, rather than evolutionary, approach today.
Elements needed in a 21st century IoT infrastructure
- Open protocols. Organizations that want to experience the full value of IoT need an infrastructure that embraces open standards, ensuring the interoperability and connectivity of the various IoT devices on your network. This involves migrating to IPv6, where IP and ethernet have evolved to meet the new requirements of technology standards such as Time-Sensitive Networking (TSN). Although we are still preparing for a complete, industry-wide move to open standards, you should begin working closely with your ecosystems of technology providers – now – to adopt these standards and evolve those that are not meeting your requirements.
- Distributed workflows. IoT has sparked a shift from centralized batch-focused, “cloud 1.0” networking technologies to decentralized, real-time, “cloud 2.0” technologies – also known as fog computing. Therefore, deploying blockchain-based, decentralized workflows that interconnect across incompatible ERP systems, supply chains, and multiple organizations allows for an instantaneous and trusted exchange of value, without wasting time for transaction reconciliation and troubleshooting.
- Machine learning. Artificial intelligence and machine learning (ML) are crucial to achieving the full value of IoT, as they allow your IoT systems to constantly learn, adapt and improve your project, whether it involves remote operations or predictive analytics. Think of AI/ML as the brains, and IoT as the body of your solution. In your new infrastructure, ensure individual robots or workstations are aligned on “what,” but are making independent decisions on “how.” Here, you must obtain the proper expertise and start implementing selectively.
- Security. Last but not least, you’ll need to build a modern security architecture into your new IoT system. The more intuitive your network, the better positioned you will be to thwart and mitigate cyberattacks. For example, ML-based self-learning systems can instantaneously identify Distributed Denial of Service (DDoS) attacks or unauthorized flows.
With these elements in place, you can take a pragmatic approach and insert key elements of a 21st century architecture into existing workflows. But this should not be a set of disjointed actions; these should be building blocks of the modern, end-goal architecture. Here are a few tips for getting started:
- Chart a vision of the revolutionary end state of both your business and technology, which includes IP, AI/ML, fog computing and other key technologies. Also, ask yourself, in what market do you want to compete? What are your value propositions? What are the technologies’ implications for your workforce and processes?
- Create a flexible “micro-services-like” framework of your future business and start building it in a multi-step process. This involves converting one section or workflow of your enterprise (one “micro-service”) at a time. Here, it is important to have consistent architecture, or you will find yourself re-doing several “one-off” decisions you made in multiple sites or locations.
- Bring in other leading technologies (especially AI/ML, blockchain, fog computing), and start by seeking expertise on how these technologies work with IoT. Together with IoT, AI/ML, blockchain and fog computing can overcome previous barriers to adoption and success, including latency, bandwidth and reliability issues.
- Join organizations that can help you in this journey. Organizations such as Open Fog Consortium, Trusted IoT Alliance and IEEE will not only provide a wealth of resources and expert guidance, but also are key players in the move to open standards – an essential ingredient for IoT success.
Even if you must begin with a brownfield environment and have to make incremental changes atop a legacy infrastructure, you can still realize results. The key, however, is to chart the end-state architecture first, and begin making small changes to your existing infrastructure, consistent with that end state. Then, plan to migrate your legacy systems to open systems.
If you don’t take these crucial first steps, you’ll end up deploying use-case-specific or inconsistent solutions in multiple sites that you’ll eventually have to redo – costing you more in the long run. Further, such patching will prevent you from having to redefine your business as you look to compete against new entrants that embrace new architectures from the ground up.
Once you update your infrastructure – whether in a greenfield or brownfield environment – you’ll find yourself with a flexible and open framework where both physical and digital processes are synchronized. In this end state, workflows, ML data, security criteria and blockchain policies are all managed in a federated and interconnected model, while the inputs and execution are fully decentralized. In simpler terms, your network will be highly intuitive, allotting you the scalability and security not found in a traditional infrastructure to drive your journey toward IoT’s transformational value. As a result, you’ll more readily realize IoT’s potential as a business a disruptive game changer – even when you can’t start with a “blank canvas.”