The 5G campus network is set to revolutionize the manufacturing industry. After the first operational campuses came online in 2019 in China, the German airline carrier Lufthansa was among the pioneers of the 5G campus network outside of Asia. Their 5G test bed was set up shortly before the Covid-19 pandemic hit, and faster than anticipated became an essential and productive interface for technical inspections. In the meantime, automotive OEMs like BMW, Mercedes-Benz, VW, and Ford have announced their plans to build their own 5G campus networks, at least for testing purposes. However, many of these plans are yet to bear fruit, so it’s still early days for the factory of the future.
What makes 5G so interesting for manufacturing? 5G is a game-changer in terms of the amount of data that can be transmitted over a mobile network, as well as the number of devices that can be connected: 5G can enable peak data rates of up to 20 Gbit/s, with up to 1 million devices connected per km2, and data transmission is highly reliable, with a latency of as low as single millisecond. This means that the newest generation of mobile communication technology can cater well for the special requirements of a whole range of modern smart factory solutions, which depend on low latency and high bandwidth connectivity. Applications ranging from Digital Twins for design, testing, and predictive maintenance, to Robotics as a Service (RaaS) and 3D printing, to virtual and augmented reality for the enablement of customization and remote maintenance, and on to intelligent and autonomous logistics, are hungry for high-performance connectivity. The smart factory of tomorrow demands seamless, utterly reliable, high-bandwidth and low-latency (wireless) communication between hundreds, if not thousands, of sensors, machines, AI capabilities, and multiple external networks – just for a single campus to function.
Connectivity within and out of the campus
The design of connectivity for a 5G campus network must encapsulate two primary goals: First, connecting devices within the campus with each other at low latency, and secondly connecting the campus to the outside world. The devices on the campus need to be networked with each other in order to ensure that the latency (response time) between the many terminals in the factory is kept to an absolute minimum, ensuring that critical information regarding the quality of products, the state of the machines, and the safety of the humans in this environment can be exchanged at times down to as low as 1ms round trip time. 5G cells can cover several hundred meters in every direction, so theoretically, a single antenna can encompass an entire campus.
There is also a need for the campus to be connected to external networks in order to access cloud services and partner networks. Very often, the company data center itself will be outside of the perimeter of the campus network. So, in the first place, the campus network needs to connect to the company internal network to access and store company data. The IoT devices in the smart factory often depend on communication with cloud services (e.g. to store and aggregate data) and cloud-based applications, such as the ERP (for inventory management and supply chain management) and AI services (for analytics, etc.). Beyond this, connecting to a variety of partner networks is necessary for the basic functioning of a smart factory: at the very least, in order to manage just-in-time deliveries with multiple suppliers, but even more so, for the operations of robotics in the plant, often leased in the form of RaaS models. Therefore, the robotics partner needs secure access to the machines, to monitor production quality, make adjustments, and perform analytics, predictive maintenance, and reporting. Finally, some use cases – such as immersive VR environments for the customization of high-end luxury vehicles – may require an interface to enable certain data to be transferred from the company data center or the campus network to end user access networks. All of these links to external networks must be implemented in a way that ensure not only the lowest latency on the connection, but also the highest security and resilience.
Designing resilient connectivity for the campus network of the future
When designing a 5G campus network, one initial choice that needs to be made is whether to create a standalone and non-standalone 5G network – this is a question of the extent to which the campus network should be able to guarantee data flows independently of a single mobile operator. In Germany, for example, it is possible to apply for a private 5G frequency spectrum license, enabling the operation of completely independent networks. The advantage here is one of resilience: by running their own 5G network, companies can overcome any disruption in the general mobile network coverage of a specific provider. Clearly, further resilience can be built into the design by using multiple redundant 5G antennas, even though each one alone would provide sufficient coverage. Securing the internal IoT network through a zero-trust policy that authenticates devices within the network adds a further layer of trust. Here, 5G technology enables a greater level of security than alternative wireless technologies, through SIM-based identification and network slicing potential.
The really interesting challenge is connecting the 5G antenna to external networks. Having created a secure and resilient high-performance network within the campus boundaries, it is sensible to ensure the same kind of performance and security for the connections that leave the network. Again, for the purposes of resilience, the manufacturer should avoid being dependent on a single carrier to ensure this essential connectivity. Here, working with multiple providers to ensure redundant pathways for critical data flows offers more robust connectivity, while also avoiding vendor lock-in. The easiest way to achieve this is by working with a data center and carrier neutral interconnection service provider, operating a distributed platform that connects a diversity of infrastructure partners and demonstrates network density.
Once connectivity to a secure and resilient distributed interconnection platform has been established, there are a number of options for ensuring SLA-backed, high-performance, low-latency, and highly robust connectivity to clouds and partner networks. One option is a point-to-point private line, which can be used for secure connectivity to a single business-critical network – such as from the campus network to the company’s own data center. But an interconnection platform with an integrated Cloud Exchange can do more: enabling direct and dedicated connections to cloud services, bypassing the public Internet. This ensures much greater security for the network’s traffic, but it also has the advantage of ensuring the shortest data pathway and therefore the lowest possible latency to the clouds in question. This can be further enhanced with cloud-to-cloud communication to optimize the manufacturer’s multi-cloud strategy. Equally, connecting to a set of trusted partners for a given use case, such as operating the production plant at the campus, can be achieved through building a Closed User Group (CUG) – this functions like a highly secure “mini Internet” for invited participants, with authentication and policy compliance built in. In this case, the manufacturer, as the owner of the CUG, sets the rules for its invited participants, and the interconnection provider can support with compliance auditing and connectivity requirements for these partners. Finally, peering with end-user access networks (in the markets where the manufacturer is active) provides the most efficient, lowest-latency, scalable high-bandwidth connectivity for allowing end-user access to resources that the manufacturer makes public – such as VR and AR simulations as part of the customization and ordering process.
Shielding data flows against real-world events
All connectivity infrastructure is physical infrastructure embedded into real-world environments. As such, it is at the mercy of real-world events – be they mistakes on building sites, power outages, or extreme weather. As such, redundancy in connectivity – not only at the physical hardware and geographical level, but also at the contractual provider level – is necessary to protect data flows. The first step here would ideally be to create redundant connections from the campus to the interconnection platform via different connectivity providers and different pathways, and connecting to the platform from geographically separated data centers where it is accessible. Such redundancy, built on multiple levels and using of a diversity of infrastructure providers, reduces significantly the chances of the connectivity suffering any downtime – especially if the connections are active-active, so that there is no need to re-route traffic in the case of an incident with one of the providers. Because incidents do happen in the real world, and critical data flows need to be shielded from collateral damage.
When it comes robust connectivity, a diverse and healthy interconnection ecosystem based on provider neutrality enables a heightened security against outages. There is strength in numbers that vastly surpasses what any single provider can achieve in isolation. Take what we do at DE-CIX as an example: We place enormous value on physical redundancy to support the SLAs we guarantee our customers. Globally, we share our capacities across multiple sub-sea cable routes and we check the exact pathways, including GPS coordinates, to ensure that these paths do not overlap at any point. We purchase connections as diversely as possible along multiple different routes, so that even in the case of outages, connectivity can be maintained. On the local level, the same applies: multiple redundant and non-overlapping cable connections between every data center where our platform is accessible creates a highly robust and resilient, failure-safe interconnection environment. We ensure the highest levels of diversity on multiple levels: different operators, different cable stretches, different upstream products, geographical distribution and redundancy for our platform and routers. This model can be replicated by enterprises to also ensure resilient connectivity: What we at DE-CIX do on the global and metro levels is what every company must also do for their own critical data flows and value chains.
5G enables low-latency, high bandwidth, flexible, scalable, and highly resilient connectivity within the confines of a campus network. However, outside of these boundaries the same connectivity characteristics need to be built up through diversity. A distributed data center and carrier neutral interconnection can simplify this process enormously.