How Long Can IoT Gateways Survive?

In the early IoT architecture, IoT gateways were a crucial part, but with the development of edge computing and other technologies, many new technologies have emerged that are likely to phase out gateways.

Are IoT gateways still important?

For many years, IoT gateways have been fundamental components of both consumer and enterprise IoT architectures, serving as a bridge between different networks, allowing data generated by IoT devices to be transmitted to the cloud.

However, with the development of IoT applications across various industries, especially in the past three to four years, the architectural requirements have changed. With enterprises adopting IoT on a large scale, the definition of IoT gateways and the role devices play in a broader IoT architecture have become blurred.

The most basic function of an IoT gateway is to interact with sensors or other devices within a local area network and aggregate data for transmission. While gateways still play this bridging role in some architectures, the demand for edge computing capabilities, growing security concerns, evolving IoT customer needs, and the overall recognition of the merging of IT and OT worlds have triggered an evolution beyond gateway-centric IoT architectures.

Overall, the IoT gateways of 2015 or 2016 are the same as today’s IoT gateways. They still provide a means to get from one place to another. The difference is that in 2015, you only had IoT gateways, while in 2019, you have many other devices to drive the IoT architecture, not just IoT gateways.

Traditionally, IoT gateways have power, memory storage, and functions similar to PCs, but in rugged enclosures, they can serve as bridging elements, connecting OT devices in factories or industrial environments, gathering data from machinery, controls, or other elements in OT settings, and delivering it to the IT world or data centers or the cloud where technical standards exist for IoT reporting and analysis that cannot operate in OT settings.

Some IoT gateways still perform repetitive tasks today, most likely appearing in IoT architectures where enterprises use managed cloud or data center services to analyze and store data generated by their IoT devices. For these service providers, traditional IoT gateways remain important.

However, some enterprises are using IoT devices to address issues closer to the network edge. This can be seen as an evolution of devices, where gateways are becoming routers, and routers are becoming edge computing resources.

Initial systems focused on running simple scripts and binaries through their own hardware processors, now have evolved to include container-based network functionalities, allowing them to interact with managed clouds and essentially push machine learning algorithms to the edge to support local decision-making.

This evolution is largely related to the development of IoT architectures in recent years, from isolated proof of concepts to more data-generating endpoints across local and wide area networks. These applications can benefit from more edge computing processing power without having to send every point (or byte) of data to a distant data center or cloud.

More and more enterprise customers have realized that the once independent OT and IT environments can now communicate, which is central to application development. For example, traditionally, industrial process automation and control applications used PLC controllers. As more IoT sensors are applied in industries, IoT gateways have been used to collect and forward data from these sites, but they still remain independent of traditional OT architectures. There was a clear stipulation that the IT world and the OT world should not interact. Nowadays, with new analytics and software capabilities validated across many different industry domains, there is a clear understanding that only by somewhat merging the OT world with the IT world can synergy be achieved and efficiency improved.

At the same time, merging the OT world with the IT world means less data traveling back and forth between customer endpoints and the cloud, and faster decision-making, which significantly reduces bandwidth costs. Given that the distance to cloud or other organizational resources is no longer an issue, the decision-making around data can even be automated. Previously, multinational companies had to be careful not to violate international regulations regarding data transfer between different countries, but now they can more easily leverage more edge processing and analytics capabilities to manage data transfers between different countries/regions.

The result of this shift is that IoT gateways have lost their position in IoT architectures, replaced by new platforms: edge routers or converged edge systems that combine networking, IoT management, processing power, and security, even software-defined networking.

A major aspect of moving beyond traditional IoT gateways is routers, which not only include connectivity but also use SD-WAN to allow dynamic use of connection paths. For example, a police car may have a dual-modem router, one pointing to FirstNet and the other to Verizon, and can set specific application policies for the network based on specific performance characteristics or other factors. Adding such advanced capabilities to edge routers is one of the reasons traditional gateway demand will ultimately decrease.

In any type of IT environment, we realize that once we have more computing power and storage, we can start taking different approaches. For example, when a business begins working with customers of traditional IoT gateways, they may encounter some issues, but if they can reduce the distance between where data is collected and where it is processed, these issues can be resolved. Unlike before, solving this problem has a value-added effect.

The building affairs in most enterprises and industrial IoT environments can be described as a “mixed environment,” where the prevalence of traditional IoT gateways or other types of platforms depends on use cases, the data generated, the data that needs to be processed and stored, and how much intelligence and security the edge requires. More advanced networks can go further down this path, where IT not only merges with OT, but can also reshape OT, realizing the vision of highly automated enterprises and industrial environments.

Meanwhile, traditional gateways are still needed to bridge the gap from OT to IT. Just in manufacturing, how many factories worldwide (through IoT deployments) have achieved connectivity in the past three years, and how many factories remain offline? This specific number is unclear, but it is believed that factories connected through IoT deployments represent only a small portion of the total.

As more companies deploy IoT, some organizations may find that the transformation function of IoT gateways is sufficient to meet their use cases. This means that some form of IoT gateway will be with us for a long time to come.