Edge computing is an IT deployment designed to put applications and data as close as possible to the users or “things” that need them. A distributed computing paradigm called "edge computing" brings computation and data storage closer to the data sources. This should reduce bandwidth usage and speed up response times. It is not a particular technology, but an architecture. It is a type of distributed computing that is topology- and location-sensitive. Edge computing, or simply "edge," relocates computer processing and storage (today frequently referred to as "compute") to the network's edge. Most importantly, it is here that it is as close as it can be to data sources, users, and devices. Frictionless store checkout in retail, which enables customers to pick up things off the shelves and walk out the door after getting checked out without waiting in line, is one of the most cutting-edge applications of edge.
What is Edge Computing?
Why is Edge Computing Necessary?
Edge computing is necessary to address shortcomings in cloud-based applications and services with respect to performance and regulatory requirements. In short, cloud computing can’t always meet the required demands in terms of response time that critical applications require. Companies that face government regulations regarding where data is stored may also find cloud computing can’t deliver the sort of local storage they need.
It’s an issue because the trend toward digitization to improve efficiency and business performance is fueling demand for applications that require peak performance, particularly Internet of Things (IoT) applications. IoT applications often require lots of bandwidth, low latency, and reliable performance while meeting regulatory and compliance mandates, making them classic candidates for the edge.
Deploying Edge Data Centres
While edge computing deployments can take many forms, they generally fall into one of three categories:
1. Local devices that serve a specific purpose, such as an appliance that runs a building’s security system or a cloud storage gateway that integrates an online storage service with premise-based systems, facilitating data transfers between them.
2. Small, localised data centres (1 to 10 racks) that offer significant processing and storage capabilities.
3. Regional data centres with more than 10 racks that serve relatively large local user populations.
Regardless of size, each of these edge examples is important to the business, so maximising availability is essential.
It’s critical then, that companies build edge data centres with the same attention to reliability and security as they would for a large, centralised data centre. This site is intended to provide the information you need to build secure, reliable, and manageable high-performance edge data centres that can help fuel your organisation’s digital transformation.
How IoT is Driving the Need for Computing at the Edge
The IoT involves collecting data from various sensors and devices and applying algorithms to the data to glean insights that deliver business benefits. Industries ranging from manufacturing, utility distribution, traffic management to retail, medical and even education are making use of the technology to improve customer satisfaction, reduce costs, improve security and operations, and enrich the end user experience, to name a few benefits.
A retailer, for example, may use data from IoT applications to better serve customers, by anticipating what they may want based on past purchases, offering on-the-spot discounts, and improving their own customer service groups. For industrial environments, IoT applications can be used to support preventive maintenance programmes by providing the ability to detect when the performance of a machine varies from an established baseline, indicating it’s in need of maintenance.
The list of potential use cases is virtually endless, but they all have one thing in common: collecting lots of data from many sensors and smart devices and using it to drive business improvements.
Many IoT applications rely on cloud-based resources for compute power, data storage and application intelligence that yields business insights. However, it’s often not optimal to send all the data generated by sensors and devices directly to the cloud, for reasons that generally come down to bandwidth, latency and regulatory requirements.
Why Edge Computing is Needed in IoT Applications?
Bandwidth
The volume of data some IoT applications create can be staggering, similar to the costs associated with sending it all to the cloud, making local processing more practical and beneficial. It’s also a gating factor for any application that requires streaming large amounts of content, including high-definition video that may be used in oil and gas exploration applications.
Latency
Some applications require extremely low latency, which is the time it takes a data packet to travel to its destination and back. Any application having to do with safety, for example – such as driverless cars, healthcare or industrial plant floor applications – require near instantaneous response time. Cloud services are not optimal in such cases due to the delay inherent in the round-trip to a centralized service.
Regulatory requirements
In highly regulated industries and regions (such as in Europe with the General Data Protection Regulation, GDPR), the way in which personal information is handled is tightly controlled, including where it is stored and how it is transmitted, driving the need for localized data centres.
In all these instances and more, edge deployments are critical in addressing these issues.
What are the benefits of edge computing?
Of course, all IT is about addressing business requirements, and edge is no different. Edge computing is helping businesses as they embark on digital transformations and use IoT applications to improve the customer experience and operational efficiency as well as develop new revenue streams.
Improving the customer experienceCustomers see examples of IoT applications all around them. Digital signage improves their retail shopping and transportation experiences. Industrial field service personnel use augmented reality applications to help them more easily service complicated machines and devices. You can now do most of your banking from your phone and have your healthcare devices monitored from afar. IoT applications are making life easier for customers in just about every walk of life.
Improving operational efficiencyIoT applications help improve operational efficiency in areas such as predictive maintenance for all sorts of machinery and equipment, be it in industrial environments or data centers, to rectify issues before they cause downtime. Radio Frequency Identification (RFID) tracking helps retailers with inventory management and loss prevention, and enables healthcare providers to track expensive equipment, such as computers on wheels carts. Cities use IoT applications to monitor busy intersections and control traffic lights to reduce traffic jams. Indeed, improving operational efficiency is probably the biggest single reason companies deploy IoT applications.
Develop new revenue streamsEntirely new industries are cropping up based on IoT technology. Uber and Lyft wouldn’t be possible without it, nor would short-term bicycle and scooter rental services. Logistics companies can offer new services based on their ability to provide the real-time status of where containers are and whether climate controls are working properly. Predictive maintenance services that are valuable to customers also mean new revenue for manufacturers and service providers. A slew of home monitoring services now exist that rely on a series of sensors and Internet connectivity. Healthcare providers can now offer “digital hospital” services including remote device monitoring and analysis.
Future of edge computing in your industry
Any company, in any industry, can apply IoT technology and edge computing to develop new revenue streams as well as improve customer experiences and operational efficiency. The principle behind the applications is the same, regardless of the exact implementation: devices or sensors at one end sending data to an edge data center for processing and perhaps some analytics, then to a more centralized application (often in the cloud) that delivers the promised benefit to the company.
It is clear that some vertical industries are emerging as early adopters of IoT technology and implementing successful applications. The lessons they’re learning apply to other verticals as well, so examining where they’ve had success can help fuel ideas for leaders in other industries.
Addressing Challenges with Edge Computing
To realize the benefits that IoT applications promise, however, requires that edge data centres have the performance and reliability that the applications demand. That presents some challenges, because edge data centres can be located literally anywhere: in a wiring closet or server room, in an office populated with employees, in a retail establishment full of employees and customers, or in a harsh outdoor environment.
No matter where it is located, ensuring reliability and performance of edge data centres involves addressing three key requirements: remote management, rapid and standardized deployment, and physical security.
Most edge data centres have few to no IT staff on site to manage them, whether it’s a remote, outdoor facility driving utility IoT applications or a retailer with hundreds of stores. In such instances, the ability to remotely manage and service the edge components is critical. Maintenance needs to be predictive and proactive, to ensure the site has no downtime and to reduce the cost of service calls. A cloud-based management platform that takes advantage of intelligent analytics applications can be an effective solution.
Find out more about remote management of edge computingStandardised and Rapid Deployments
Given the large number of edge data centres that many organisations are going to have, it’s important they be delivered in a standardised, repeatable and rapid manner. The alternative – a series of ad hoc IT deployments – creates a nightmare scenario for both speed of deployment and ongoing management.
The solution here involves using a reference architecture that ensures consistency in each edge deployment. Such architectures define a baseline level of devices and services, while allowing for some variation depending on the requirements of each location. Even better is to have a finite number of reference designs from which to choose for each site, to ensure consistency.
Prefabricated, modular micro data centres are often a good solution for edge data centres. They include all the required power and cooling infrastructure as well as management software. It’s all pre-integrated and installed in a rack or enclosure, ready to accept IT equipment – which is typically installed by an IT solution provider or systems integrator. Some micro data centres are also certified by leading converged and hyperconverged IT equipment manufacturers, which helps ensure good performance and reliability.
Read more about standardised and rapid deployments of edge computingPhysical Security
Edge data centres may be located in server rooms and IT closets, under cash registers or desks. Even if they are in a dedicated room, it may not be secured. This leaves the edge infrastructure open to accidental damage, attack from nefarious actors who intend to do harm, as well as employees with good intentions who simply don’t know any better.
Providing proper physical security requires three components:
Monitoring the physical space, using sensors that can report on temperature and humidity levels, and detect environmental changes caused by fire, smoke, flooding or the like.
Control over the space, to ensure only authorised personnel have access to edge infrastructure.
Supervision of the environment using audio and video, with recording, so you can visually see who is accessing edge spaces.
Perhaps not surprisingly, those three elements figured prominently among respondents to an IDC survey* about the top concerns over edge deployments. Issues around security, monitoring and controlling access to the physical space accounted for five of the top six concerns the 200+ respondents had about edge computing.
Find out more about physical security at the edge* IDC, Edge Computing: The Next Stage of Datacenter Evolution, April 2018.
Author: Jamie Bourassa
Vice President of Edge Computing & Channel Strategy for the Secure Power Division of Schneider Electric
Jamie is responsible for enabling the Secure Power Division commercial strategy and ensuring that Schneider Electric aligns to the market evolutions related to Edge Compute, IoT, and other disruptions that increase the criticality of local computing for customers across all commercial and industrial segments. With a global career in IT Channels Strategy, Sales Operations and Offer Management, Jamie brings a unique set of competencies needed in evaluating and delivering on the current disruptions in the market.
Read more of Jamie Bourassa’s Edge content