According to a report by IoT Analytics, up to 11.3 billion devices of various types are currently connected to the internet worldwide. Analysts expect this figure to more than double to 27.1 billion by 2025, despite the pandemic and shortage of chips. Looking more closely, it is clear that the increase in industrial devices connecting to the internet is driving this boom, says Nathan Howe, VP of emerging technology at Zscaler.
The concept of the internet of things (IoT) is deceptively simple, devices gather information from their surroundings and send it to a central point for processing. This could be a sensor on your wrist, for example, whether it is a smart watch, fitness wristband, or heart rate monitor. The information is sent to a server to be processed and then returned to the sender and possibly shared with an additional set of services for further analysis. A process such as this involves a veritable tsunami of data streams which require increasing amounts of computer capacity, causing the associated costs to rise. This volume of data being collected and sent will create new opportunities in the industrial environment, as well as in the private sector. In industry, it will be essential to break new ground in terms of information processing at the edge.
The IIoT, or Industrial Internet of Things, makes it necessary for every factory or production line to have a central location for data processing. Sensors, which occur at a density of every five to ten centimetres of equipment in a production environment, will generate information that involves many data streams. If both financial and latency costs of transporting these data streams is to be kept low, then this service needs to be located on site and close to the data source, which is where edge computing comes into play. This enables information to be recorded, collected, and processed locally.
For example, a forklift truck drives autonomously along the aisles of a warehouse and uses its sensors to measure the distance to objects in its environment. If the sensors detect an object or person in the way, the forklift stops. However, this triggers a domino effect of data traffic, as all associated processes have to be stopped and reassessed when the machine stops.
Edge computing solves the latency problem
The IoT concept is reliant on information flowing quickly. The more sensors that are involved, the more decisions there are that need to be made simultaneously, which then impacts the speed of data processing. Some decisions need to be made in the blink of an eye, especially when issues of security are involved. For instance, the processing speed in critical environments is set at less than 15 milliseconds which is the how long it takes a person to smell something and locate that smell in their brain. The further away the computer capacity needed for the processing is from the sensor, the harder it is to maintain low latency. Here is where edge computing solves the problem of long data paths.
By having the necessary technology in a smart box at the edge, millions of IoT devices can be connected via the cloud, and decentralised data processing can then be achieved at the ‘edge’ of a network, close to the data source. This smart box at the edge of a network performs various functions in edge computing. It holds software for processing inside a container, where it maps virtual functions. This makes it possible to monitor the status and maintenance of the devices, as the box can be controlled externally. Edge computing not only ensures low latency in processing, shorter routes for data streams, and lower costs, it also goes hand-in-hand with another revolution, virtual access to IoT devices, which replaces the need to have service technicians on site.
This vision also means that huge racks of hardware in the date centre will be replaced by a handful of small, highly intelligent boxes at the edge of the network, which can then be used to orchestrate a range of different processes. Access to the box can be controlled from anywhere. Hyperscale providers ensure this is the case by using AWS Wavelength, regardless of whether it is for controlling a smart home or a factory system. Due to the almost unlimited possibilities offered by this technology, Gartner has predicted that edge computing will have a market potential of around [$11 billion(€10.45 billion)] by 2026.
Augmented reality (AR) applications are already demonstrating the potential of edge computing. One example of an edge workload that requires communication with low latency is the use of AR for searching for goods in a warehouse. Warehouse workers use their data glasses to navigate around displayed routes to locate the ordered goods and pack them. Barcodes are scanned via these glasses, and the goods are recorded as a message sent back to the SAP system.
However, there are also risks. Provisioning of any workload, let alone a new AR workload, as an open service via a network such as a telecommunications provider, exposes the service to other non-trustworthy users and potential attacks, misuses, and disruptions. It is clear that, just like for all other interconnected-based data streams, security also needs to be embedded in edge computing as a top priority.
Zero trust also secures IoT workloads
Zero trust has established itself as a paradigm for securing user-based data streams to applications, regardless of where they are hosted. The granular nature of zero trust controls also makes sense for protecting workloads that are carried out in edge computing environments. Access to the isolated workload is only granted from the authorised source, meaning that the edge workload is protected from unwanted access or general external visibility. As an access concept like this is no longer provided at the traditional network level, it offers the granular protection that is needed for shared edge services.
Cloud providers, such as AWS, are already working with Vodafone in Europe to make services for secure access to edge computing applications possible. The demand for edge computing services is going to increase when companies recognise the competitive advantages to be gained for Industry 4.0 or applications such as augmented reality. Edge computing services provide access for anyone who shares the network with the edge, and therefore anyone who is connected to the operator network. Going forward, even a smartphone that is connected to the edge service via a carrier network will be enough to access it. None of the telecommunication provider’s other users will be able to see or access the company-specific applications.
The 5G service provided by the telecommunications provider will ensure that the required speed is available for access. This means that, in the future, an unimaginable number of applications will be possible in a production environment, with smart factories, networked vehicles, and innovative retail experiences.
The merging of IoT applications with edge computing will make the required data transfer speed a reality. It is important not to ignore security in the overall structure. The combination of edge, IoT and zero trust via 5G will enable secure, high-performance application scenarios to become a reality, protected from unwanted prying eyes.
The author is Nathan Howe, VP of Emerging Technology at Zscaler.