The adoption of edge computing is accelerating in all types of enterprises. Mission-critical computer components are increasingly found outside of data centres and distributed throughout factories, warehouses, retail stores, hospitals, college campuses, and in pretty much every other sector of the economy. Written by Jonathan LaPorta, Pfannenberg USA.
Explosive growth of Edge computing
In fact, three-quarters of all data generated by Hewlett Packard Enterprises, a multinational information technology company, is now created and processed outside the traditional centralised data centre or cloud.
Edge computing is the placement of application logic as close to the source of data as possible to reduce latency, optimise bandwidth usage, and minimise costs while increasing visibility and control, speeding up processing, and enhancing user experience.
IDC defines edge computing as “the movement of computing resources to the physical location where data is created, dramatically reducing time to value and the instant enablement of business processes, decisions, and intelligence outside of the core IT environment” and predicts worldwide spending on edge computing to grow by 56% to $274 billion from 2022 to 2025. Grand View Research expects the edge computing market to expand at a compound annual growth rate (CAGR) of 38.9% from 2022 to 2030.
The explosive growth of edge computing is fuelled by ubiquitous megatrends like digitization, IoT, and automation. All that data needs to be processed somewhere, and it is usually much faster and more efficient to do it close to the edge where people and things produce and use that information. Keeping and processing data at the edge can also enhance reliability, security, and privacy by minimising the transmission of sensitive information and exposure to network failures and malicious attacks.
Take IoT video cameras and machine vision systems for example. They are literally everywhere. One high-definition camera generates a lot of data, so imagine the bandwidth requirement of dozens or hundreds of devices. Consider a building secured by video cameras. Rather than stream all the video data over a wide area network, it is obviously more efficient to process the content locally with motion-detection algorithms and only transmit to the cloud and store what is useful.
The same principle applies to machine vision systems widely used in manufacturing to perform imaging-based automatic inspection for process control, quality assurance, traceability, and predictive maintenance. It makes sense to collocate the imaging processing and analysis close to the vision system to minimise bandwidth requirements. Even more importantly, the low latency and high reliability of the edge computing architecture enable real-time applications, such as part rejection, automated calibration of equipment, and robot guidance.
Low latency is critical to many edge computing use cases to improve process efficiency and enhance user experience. Large hospitals see the benefit of storing and processing data, such as medical sensors and digital health records closer to the point of use. Warehouses and retail stores find it more efficient to run inventory and point-of-sale systems locally rather than depend on the cloud for each individual transaction.
Need to protect distributed components
Edge computing architecture means that computing equipment is located outside of a climate-controlled data center and is usually far from the watchful eye of an IT professional. For these reasons, it is vitally important to protect these distributed systems that serve mission-critical functions.
The servers and networking gear are often placed in suboptimal locations, in whatever space is available, sometimes in crowded spaces on the factory floor or up in the rafters of a warehouse, or in the closet of a hospital.
Even when there is ample space, the environment may be inhospitable. The ambient temperature of factories can be quite high, and dust and oil in the air can be a problem.
For all these reasons, the distributed computer equipment is typically placed in enclosures for protection from the immediate environment, but paradoxically, these enclosures trap heat that is generated by the electronics, and therefore the enclosures themselves require protection from overheating.
To ensure continuous operation of sensitive electronics, the optimal temperature inside an enclosure is 95°F. A variety of cooling methods have proven effective in the thermal management of electrical enclosures including passive air cooling with fans and active cooling with air conditioning.
Simple, proven cooling solution
For many edge computing applications, a filter fan will provide sufficient thermal protection. The basic concept was originally invented by Otto Pfannenberg in the 1950s. Pfannenberg’s newest design of filter fan, the Datawind Filterfa, has a temperature indicator and is mounted externally to the enclosure to enable full use of all racks within the bays.
Filter fans are an economical solution for thermal management of electrical enclosures where the ambient temperature is always lower than the temperature required by the computer components inside. The fan forces the relatively cool surrounding air into the electrical enclosure, so that a slight overpressure builds up inside the enclosure and relatively warm air is expelled.
Compared to air conditioning, filter fans cost much less to purchase, consume much less energy, install much more easily and are much easier to maintain. Often air conditioning is not even an option due lack of space for a compressor.
More challenging cooling applications
Some applications require enclosures specially designed for challenging environments, from weather and impact-resistant outdoor enclosures to tightly sealed stainless-steel enclosures for food production facilities that must withstand high pressure cleaning. The better the seal, the less heat is naturally dissipated, and the bigger the cooling challenge.
When the interior air of the electrical enclosure must be protected from the exterior environment, a closed loop cooling system is required. Similar to a filter fan, the closed loop air to air system requires fairly cool ambient air to be effective, but is more protective from a dirty outside environment. One of the more energy efficient technologies to use when there is a temperature difference between the internal target temperature and the surrounding temperature (∆T≥10°C) is air to air technology. An air to air heat exchanger removes heat from inside of the enclosure to the cooler environment using the least amount of energy while still providing a closed loop ingress protection.
Applications where the electronics generate a high heat load and/or the ambient air is too warm to provide adequate cooling, an active cooling system may be required. These air conditioning systems utilize a compressor for cooling in a closed loop design that isolates the interior of the enclosure from the exterior environment.
For extremely demanding industrial and outdoor applications, specially designed cooling units such as the Pfannenberg DTS Series of side mounted cooling units provide NEMA 4 level of protection from washdown conditions and caustic air and can operate in ambient temperatures up to 140° F (60° C).
By implementing the optimal cooling solution for your distributed components, you can take advantage of the many advantages of edge computing with the peace of mind of knowing that all your mission critical IT infrastructure is protected from overheating.