Edge computing (EC) allows data generated by the Internet of Things (IoT) to be processed near its source, rather than sending the data great distances, to data centers or a cloud. More specifically, edge computing uses a network of micro-data stations to process or store the data locally, within a range of 100 square feet. Prior to edge computing, it was assumed all data would be sent to the cloud using a large and stable pipeline between the edge/IoT device and the cloud.
Typically, IoT devices transfer data, sometimes massive amounts, sending it all to a data center, or cloud, for processing. With edge computing, processing starts near the source. Once the initial processing has occurred, only the data needing further analysis is sent. EC screens the data locally, reducing the volume of data traffic sent to the central repository.
This tactic allows organizations to process data in “almost” real time. It also reduces the network’s data stream volume and eliminates the potential for bottlenecks. Additionally, nearby edge devices can “potentially” record the same information, providing backup data for the system.
A variety of factors are promoting the expansion of edge computing. The cost of sensors has been decreasing, while simultaneously, the pace of business continues to increase, with real-time responses providing a competitive advantage to its users. Businesses using edge computing can analyze and store portions of data quickly and inexpensively. Some are theorizing edge computing means an end to the cloud. Others believe it will complement and support cloud computing.
The Uses of Edge Computing
Edge computing can be used to help resolve a variety of situations. When IoT devices have a poor connectivity, or when the connection is intermittent, edge computing provides a convenient solution because it doesn’t need a connection to process the data, or make a decision.
It also has the effect of reducing time loss, because the data doesn’t have to travel across a network to reach a data center or cloud. In situations where a loss of milliseconds is unacceptable, such as in manufacturing or financial services, edge computing can be quite useful.
Smart cities, smart buildings, and building management systems are ideal for the use of edge computing. Sensors can make decisions on the spot, without waiting for a decision from another location. Edge computing can be used for energy and power management, controlling lighting, HVAC, and energy efficiency.
A few years ago, PointGrab announced an investment in CogniPointTM, and its Edge Analytics sensor solution for smart buildings, by Philips Lighting and Mitsubishi UFJ Capital. PointGrab is a company which provides smart sensor solutions to automated buildings.
The company uses a deep learning technology in developing its sensors, which detects the occupant’s locations, maintains a head count, monitors their movements, and adjusts its internal environment using real-time analytics. PointGrab’s Chief Business Officer, Itamar Rothat stated:
“CogniPoint’s ultra-intelligent edge-analytics sensor technology will be a key facilitator for capturing critical data for building operations optimization, energy savings improvement, and business intelligence.”
Another example of edge computing is the telecommunication companies’ expansion of 5G cellular networks. Kelly Quinn, an IDC research manager, predicts telecom providers will add micro-data stations that are integrated into 5G towers, or located near the towers. Business customers can own or rent the micro-data stations for edge computing. (If rented, negotiate direct access to the provider’s broader network, which can then connect to an in-house data center, or cloud.)
Edge Computing vs. Fog Computing
Edge computing and fog computing both deal with processing and screening data prior to its arrival at a data center or cloud. Technically, edge computing is a subdivision of fog computing. The primary difference is where the processing takes place.
With fog computing, the processing typically happens near the local area network (but technically, can happen anywhere between the edge and a data center/cloud), using a fog node or an IoT gateway to screen and process data. Edge computing processes data within the same device, or a nearby one, and uses the communication capabilities of edge gateways or appliances to send the data. (A gateway is a device/node that opens and closes to send and receive data. A gateway node can be part of a network’s “edge.”)
Edge Computing Security
There are two arguments regarding the security of edge computing. Some suggest security is better with edge computing because the data stays closer to its source and does not move through a network. They argue the less data stored in a corporate data center, or cloud, the less data that is vulnerable to hackers.
Others suggest edge computing is significantly less secure because “edge devices” can be extremely vulnerable, and the more entrances to a system, the more points of attack available to a hacker. This makes security an important aspect in the design of any “edge” deployment. Access control, data encryption, and the use of virtual private network tunneling are important parts of defending an edge computing system.
The Need for Edge Computing
There is an ever-increasing number of sensors providing a base of information for the Internet of Things. It has traditionally been a source of big data. Edge computing, however, attempts to screen the incoming information, processing useful data on the spot, and sending it directly to the user. Consider the sheer volume of data being supplied to the Internet of Things by airports, cities, the oil drilling industry, and the smart phone industry. The huge amounts of data being communicated creates problems with network latency, bandwidth, and the most significant problem, speed. Many IoT applications are mission-critical, and the need for speed is crucial.
EC can lower costs and provide a smooth flow of service. Mission critical data can be analyzed, allowing a business to choose the services running at the edge, and to screen data sent to the cloud, lowering IoT costs and getting the most value from IoT data transfers. Additionally, edge computing provides “Screened” big data.
Transmitting immense amounts of data is expensive and can strain a network’s resources. Edge computing processes data from, or near, the source, and sends only relevant data through network to a data processor or cloud. For instance, a smart refrigerator doesn’t need to continuously send temperature data to a cloud for analysis. Instead, the refrigerator can be designed to send data only when the temperature changes beyond a certain range, minimizing unnecessary data. Similarly, a security camera would only send data after detecting motion.
Depending on how the system is designed, edge computing can direct manufacturing equipment (or other smart devices) to continue operating without interruption, should internet connectivity become intermittent, or drop off, completely, providing an ideal backup system.
It is an excellent solution for businesses needing to analyze data quickly in unusual circumstances, such as airplanes, ships, and some rural areas. For example, edge devices could detect equipment failures, while “not” being connected to a cloud or control system. Examples of edge computing include:
Internet of Things
- Smart streetlights
- Home appliances
- Motor vehicles (Cars and trucks)
- Traffic lights
- Mobile devices
Industrial Internet of Things (IIoT)
- Smart power grid technology
- Magnetic resonance (MR) scanner
- Automated industrial machines
- Undersea blowout preventers
- Wind turbines
Edge Computing Compliments the Cloud
The majority of businesses using EC continue to use the cloud for data analysis. They use a combination of the systems, depending on the problem. In some situations, the data is processed locally, and in others, data is sent to the cloud for further analysis. The cloud can manage and configure IoT devices, and analyze the “Screened” big data provided by Edge Devices. Combining the power of edge computing and the cloud maximizes the value of Internet of Things. Businesses will have the ability to analyze Screened big data, and act on it with greater speed and precision, offering an advantage against competitors.
Data Relationship Management
Device Relationship Management (DRM) is about monitoring and maintaining equipment using the Internet, and includes controlling these “sensors on the edge.” DRM is designed specifically to communicate with the software and microprocessors of IoT devices and lets organizations supervise and schedule the maintenance of its devices, ranging from printers to industrial machines to data storage systems. DRM provides preventative maintenance support by giving organizations detailed diagnostic reports, etc. If an edge device is lacking the necessary hardware or software, these can be installed. Outsourcing maintenance on edge devices can be more cost effective at this time than hiring an in-house maintenance staff, particularly if the maintenance company can access the system by way of the internet.