The possibilities are endless. That is what some vendors say about 5G, as they urge operators to partner with manufacturers to transform the technology into a lucrative revenue stream. Their intentions are good. The reality, though, is more complicated. In fact, there’s a widening disconnect between those who understand 5G – primarily network operators – and those who understand industries and manufacturing.
How can this gap be bridged? That is the multi-billion-dollar question facing the 5G ecosystem. At the heart of the issue is a need for the network to support manufacturing, while also providing the ability to make reliable decisions about critical factors such as latency, connectivity and control. In other words, the network needs to perform to the level of industrial-grade precision and reliability.
When multiple parties pursue different business models, however, this becomes a complex conundrum. For instance, there are B2B manufacturers that sell and deliver to other enterprises – while also relying on operators for connectivity. Each of these entities can be distinct and mutually exclusive with no real common meeting point apart from connectivity. The question, then, is how to create a successful business model for 5G.
The solution is to provide open computing architectures that enable all parties to integrate their models within the open infrastructure.
The hypothetical quandary
Take the example of using computer vision in manufacturing, which is expected to radically improve the productivity of manufacturing processes. On the shop floor, computer vision can aid in predictive maintenance, inspect packaging in the supply chain, automate barcode reading, automatically detect errors in the supply chain, improve safety and track logistics. However computer vision implementation can be quite costly if the processing is done at the point of action. Instead, cheaper cameras can stream videos over a reliable, low latency 5G network that aggregates and executes machine vision algorithms in a localized data center via edge compute.
In this scenario, the 5G network could be provided by an operator, and the edge compute could provide powerful computing capabilities with adequate processing powers through specialized computer vision silicon. When developed on open infrastructure with well-defined application programming interfaces (APIs), such algorithms can help application developers and device manufacturers (of smart cameras, for instance) to easily add and modify application content. Having common computer vision and cognitive APIs can make life simpler and the processes more scalable for the manufacturer. This also provides a new revenue stream for specialized vision computing companies.
Another benefit is that of cost savings. Consider the example of an electric grid operator in Asia. If data from the electric grid is sent over a 5G network to a data center in Seattle, it completely nullifies the impact of 5G. While the 5G network has ultra-fast connectivity for access, the reality is that the information is still traveling through an undersea cable from Seattle. It is not making the best use of network resources – even when using a blazing fast 5G network. This is when edge compute resources powered by 5G connectivity can be localized for this specific industry.
Ericsson suggests that operators are attractive partners within the healthcare ecosystem. 5G can enable home care and telemedicine through IoT sensors and 5G-enabled medical devices. Digital therapeutics platforms can enable healthcare providers to analyze patient data and accelerate decision making. In such a scenario, telcos partner with healthcare providers to develop connectivity and data centers to create an ecosystem. Open infrastructure helps healthcare providers introduce new applications into the operator-provided software-as-a-medical-device ecosystem. Reuse and sharing of information between multiple healthcare providers is enabled through the multiuser and API interfaces for open infrastructure.
Additionally, use cases should be centered on the devices for interoperability. Consider another example in which Industrial Internet of Things (IIoT) sensors are monitoring multiple assembly systems in a large factory. The company that created the IIoT sensors provides analytics applications. But how can analytics or intelligence be obtained from those sensors? If the manufacturer decides to use a different set of sensors from another provider, how can they reuse analytics prior to the change? All of this can be achieved if the industry evolves toward open APIs and an application marketplace.
Many of the 5G use cases being developed for the manufacturing sector are hindered by the siloes that exist in skillsets between manufacturers, operators and suppliers. Some operators do not have the skills or knowledge necessary to understand what is happening in the manufacturing environment. This also rings true for many ecosystems in different industries.
Open ecosystems are not without challenges. Security and privacy are of concern, particularly with multiple stakeholders in the ecosystem. Risk will only be mitigated when all members of the ecosystem are required to implement security solutions, processes and risk management in the open infrastructure. As such, creating digital trust across the ecosystem is critical to the success of 5G for industries in an open ecosystem.
A complete solution
The only way that a 5G network can effectively support such use cases is when automation, analytics and intelligence are inherent on the network. This requires the ability to compute locally through edge computing, where analytics are in a feedback loop that doesn’t require every stakeholder’s domain knowledge.
Having open APIs ensures that all stakeholders are able to monitor their parts of the process. In such a scenario, an operator can monitor how the network is working, while the manufacturer can monitor how its sensors are working, and other partners can monitor whether productivity is impacted across the value chain – all in real time, on the same network and independent of one another. Such a model, which supports all stakeholders and enables them to identify and correlate their data and performance, is vital to the success of industrial grade 5G.
5G for industrial ecosystems is still evolving and is to some extent nascent. While connectivity is well understood, the complex, heterogeneous ecosystem of devices, applications, developers, infrastructure, analytics and more slows adoption. This process slows even more if there’s a lack of conviction on the business case, insufficient processes and security concerns. The community needs to come together and create a common builder model with well-defined infrastructure, security and application interfaces to create a successful 5G for industries model.
