What is edge computing?

Edge computing has grown in importance as data volumes continue to grow at a phenomenal rate. This increase in data has been driven in a large part by sensor data gathered from Internet of Things (IoT) devices. Edge computing has the potential to deliver near real-time capture, process and analysis of data. It can filter out any unnecessary data too, so only vital information is sent to a data centre.

The performance of applications such as augmented reality (AR), virtual reality (VR) as well as machine learning (ML) and artificial intelligence (AI) are enhanced by handling data processing at the edge. Edge computing also allows data to be used in remote locations and can support compliance by not moving data out of its source of origin. Also, using the local area network (LAN) for data processing and storage means less data has to be sent to the cloud, helping to lower operational costs.

While keeping data at the edge is seen as more efficient than streaming to the cloud or a data centre, whatever data remains at the edge still needs to be secure. Data encryption policies need to be in effect to protect all data being streamed and stored to protect from theft or a cyberattack. Maintaining a strong edge security posture is key to the viability of edge data processing.

In answering what is edge computing, research analyst Gartner describes it as being “part of a distributed computing topology in which information processing is located close to the edge — where things and people produce or consume that information”. Distributed computing is nothing new and has been in use for decades, for example by businesses with remote operations, where the use of a centralised IT does not make sense.

Working with the traditional model of moving data to a centralised data centre via the internet is now recognised as inadvisable when modern businesses need faster insights to inform decision makers. The traditional internet was simply not designed to handle the data traffic that businesses will require in the future. The internet can also be subject to outages that lead to downtime, causing slowdowns, unavailable connections and ultimately lost profits.

Any solution at the edge has to process two distinct types of data. The first kind is system data that supports business operations and that is most often stored in smaller relational databases than those that reside in an on-premises data centre or in the cloud. The second kind of data is user data. The vast majority of the data handled at the edge is generated by sensors, making localised processing more efficient. As well as this sensor data, organisations have an increasing number of connected devices to manage — such as smartphones and laptops — that can be better handled using an edge computing model.

Edge computing is really a complement to the centralised model of computing, which is more suitable for compute-intensive workload processing. Edge is better suited to real-time processing and to generate faster insights. And edge computing can also help in reducing the load on the internet simply by keeping data processing close to the data source, minimising congestion. IDC predicts that by 2023, more than half of new enterprise IT infrastructures will happen at the edge and not in the data centre.

Lowers latency

Processing data close to where it is generated takes far less time than sending that data to a remote data centre. Outages and network congestion can increase latency and slow down real-time analytics. Wireless networks will work most effectively with an edge computing infrastructure already in place.

Increases business resiliency

Edge services can help organisations overcome a series of issues, such as unplanned downtime caused by a lack of internet availability. Keeping data within a single site means companies can better guarantee more secure connections and improve business performance simply by keeping data within the walls of a business’s operations. This can improve the user experience by delivering rich content, such as video, close to where it’s needed.

Improving data sovereignty

A big issue for many businesses is when data travels to regions and across national borders. Data sovereignty issues can be solved by processing and keeping data close to where it is created. In healthcare, this could be sensitive patient records or video data and any identifying information about an individual. It also helps ensure compliance, such as for the European General Data Protection Regulation (GDPR), by processing data at the edge and anonymising it before it is sent to the cloud or central data centre.

Limits congestion

IoT sensors will only add to the need to move and process more data. In the age of cloud computing and streaming services, the internet performs sufficiently for now. Adding in data from billions and billions of devices will put an enormous amount of strain on the global internet infrastructure. IDC predicts that the data generated by 2025 will reach 175 trillion gigabytes, an almost unimaginable volume. Edge computing is set to represent 90 per cent of that data. Severe internet congestion can be easily avoided with edge computing.

There are many and varied uses for edge computing where real-time insights can provide significant benefits for multiple business-critical operations. This could be for predictive analytics in a manufacturing setting, where maintenance of plant equipment supports continuous output. In logistics, edge computer vision or video analysis can be used in distribution centres on the packing line. Parcels can be checked on the fly to ensure they contain the right products, eliminating order errors.

Edge computing has multiple possibilities and is generating excitement in industry, local government, farming and more. Here are some edge computing examples.

Transport
Self-driving cars must prove themselves in terms of keeping both passengers and pedestrians safe. This requires real-time data processing for the up to 20 terabytes of data that each autonomous vehicle may need to process in a single day. This includes road, traffic and weather conditions, as well as information about speed and, of course, location. Sending this data to a remote data centre would slow down analysis and have serious safety implications. Edge computing will help self-driving vehicles process data more quickly and bring the vision closer to reality.

Manufacturing
Less productivity equals less profit in manufacturing. Analysing data at the edge would have important implications for perfecting production processes. Gaining insight into levels of operational effectiveness, by using data gathered by sensors across the entire production line, can help identify areas for improvement. Sensors can perform predictive analytics by checking the health of machinery and plant functionality. Plus, items such as stock control and worker safety can be monitored to keep operations efficient and risk-free.

Agriculture
In a world of finite resources and unpredictable weather, sensors can be used to monitor soil moisture levels and nutrient content. Using sensor data, farmers can accurately predict optimum harvesting times to ensure maximum yields. Drones can monitor crops to check for diseases or pests. Sensors can help to predict patterns in weather, helping farmers better protect their crops from flooding or wildfires. Overall, this type of smart agriculture can help create efficiencies, reduce operational costs and provide farmers with much greater levels of control.

Healthcare
Data is constantly being generated in a healthcare environment. Scans from medical devices such as ultrasound, heart monitors and wearable devices can generate large amounts of data. Instead of sending this to a central data store, it can be processed and analysed at the edge. Medical experts can get real-time results and make diagnoses faster, cutting down on the need for additional appointments and reducing waiting lists. Robot-assisted surgery is made safer with data being processed on site rather than sending it to a distant data centre.

Smart cities
Imagine a world where parking is easy to obtain, there are no overflowing rubbish bins and crime is under control. The smart city could bring this closer to reality with edge analytics. IoT devices can be used to support key services. Street lighting can be monitored so people feel safe on streets at night. Overseeing traffic and road conditions could reduce congestion or keep citizens out of harm’s way. Environmental sensors can be used to reduce energy use.

And more

The future of edge computing is tightly linked to the advent of 5G wireless and WiFi 6 networking technologies. This is bringing fresh impetus to edge computing with the promise of much improved data performance and near real-time analytics. Many industrial operations already use 4G and WiFi, meaning existing networks can be easily upgraded rather than starting from scratch. Coupling data processing at the edge is critical to ML and AI and for supporting the use of AR/VR-driven processes, too.

It is important to note that edge computing complements the cloud. Edge computing will not deliver the benefits for businesses in their cloud use. Cloud services enable organisations to extend their infrastructure on a global level as well as spin up compute resources for heavy workloads only when needed. Cloud reduces the need to invest in expensive IT hardware and lowers management costs. Edge computing is set to grow and complement cloud networks where processing is required far away from the central network.

OVHcloud can bring its 20 years of data centre experience to support the future of edge computing. Our Data Centre as a Service delivers an on-premises, pre-integrated cloud system featuring a wide array of DCaaS, IaaS and PaaS services.

Organisations can use OVHcloud products within their own data centre environment to enable complete sovereignty over their data, reduced operational and capital costs and the ability to benefit from the latest hardware technologies.