Technical Support for Antennas & Related Gear

Edge Computing: Data Processing at Cloud's Edge for Bandwidth Savings and Security

George Hardesty
14 minute read

Listen to article
Audio is generated by DropInBlog's Blog Voice AI and may have slight pronunciation nuances. Learn more

Edge Computing:

Edge Computing - Internet of Things:

One of the key mid-decade trends for the Internet of Things is its increasingly integrated relationship with Edge Computing. This phenomenon continues the unprecedented shift towards a more distributed and decentralized implementation of technology.

Exponential growth in the number of IoT connections globally necessitates computational solutions that possess the processing power to manage previously unimaginable volumes of data. The demand is likely to already outstrip cloud computing capacity meaning that novel technologies will be needed that work in new ways.

Edge computing is coming to the fore as a solution that can match the scale of IoT and meet its requirements for speed, security, and data integrity. Deployments are already underway with data processing outside of the cloud or at its edge becoming a favored means for providing responsive support for IoT applications, saving time and bandwidth.

This article explores the relationship between edge computing and IoT and evaluates if this technology will be able to shoulder the burden of the ever-expanding Internet of Things.

Edge computing: Key IoT Trend for 2022 and Beyond

Edge computing is an on-demand computing resource that is made available spatially, near to the source of the data to be stored or processed as opposed to Cloud computing where the computation takes place after uploading to the Cloud.

It is a form of distributed computing, where the different components of a computing system are spread across several networked devices that communicate with each other to fulfill computational tasks. Edge computer systems are distinctive in that they are asynchronous, constituent devices operate concurrently and the failure of individual network components does not compromise the whole network (independent failure).

Although it is closely associated with the Internet of Things, Edge computing and Io are not synonymous and the data processing for IoT devices is not undertaken on the devices themselves.

Edge computing brings the processing to the data

Edge computing provides a less centralized provision of computing power as it is performed near to the data source, usually an IoT device - literally, on the edge of the network. This form of computing has already been demonstrated to be effective in multiplayer online gaming and a variety of peer-to-peer applications.

Edge computing is reliant on two streams of communication between IoT devices and the edge of the Cloud:

  • Upstream: data flowing from the IoT device to the Cloud. This data transfer is provided as part of the Io software application that is used with the device.
  • Downstream: this data stream is delivered by the Cloud Service Provider from the cloud to a particular IoT device.

In a significant departure from the Cloud computing model, data is stored locally rather than centrally within the Cloud. Within an IoT network that is supported by edge computing, IoT end user devices both generate and consume data. These devices, which include smartwatches, home sensors, cars, and utility meters, have their upstream and downstream data transferred via the edge, which includes routers, mini-servers, and gateways. The edge interacts with the Cloud, fulfilling processing tasks and service provision and requesting data and services where needed from the Cloud.

Edge network resources are provided at any point between the IoT device and the Cloud. The required computational power can come from any devices that can be integrated in the upstream or downstream including smartphones, routers, hubs, and laptops. The data is outsourced to these devices where it can be processed without draining the computation resources of the Cloud.

Why is edge computing important?

The priority in implementing edge computing architecture for the support of IoT networking is the urgency of tackling the sheer volume of data that these networked devices are generating. According to Network World, the global volume of data is expected to jump to over 175 zettabytes by 2025, signaling the end of the zettabyte era of the 2010s and a transition toward Yottabyte (one million trillion megabytes) levels of data.

This means mid-2020s computing faces significant upheavals because of the growth not only of the IoT but of all forms of public and non-public digital data. The massive rise in IP traffic alone is causing a crisis among data centers that cannot keep up with the energy, hardware, bandwidth, and processing power required to support the creation, storage, and consumption of data that is taking place.

Data centers will soon be unable to provide a suitably guaranteed service which may be important for the competent functioning of certain critical applications. High data consumption by the networked devices also strains centralized data center service providers.

The decentralized approach of edge computing increases physical (geographical) proximity to end-user devices which already are consuming high volumes of data. This is to be achieved by using smart IoT devices including phones and gateways that sit at the edge of the Cloud to execute the tasks and provide services that would be provided by the Cloud. This offload or outsourcing of data management to peripheral to edge devices is expected to improve the performance of networks and with improvement in speed and responsiveness.

How IoT with edge computing works

Edge computing can fulfill the computational tasks that are offloaded by the Cloud including:

  • Application service delivery
  • Data storage
  • Data caching
  • Data processing
  • IoT device management

These devolved functions are executed by edge devices, leaving the Cloud focused on big data management and processing.

Within an IoT network that is supported by edge computing, IoT end user devices both generate and consume data. These devices, which include smartwatches, home sensors, cars, and utility meters, have their upstream and downstream data transferred via the edge, which includes routers, mini-servers, and gateways. The edge interacts with the Cloud, fulfilling processing tasks and service provision and requesting data and services where needed from the Cloud.

Before edge computing, IoT sensors and other devices would generate data that would be directly uploaded to the Cloud. The Cloud would then be responsible for storing the data on a centralized database, processing it, and responding with any action required for the IoT device to execute. Though this process takes seconds, an interruption or downtime in Cloud service delivery or poor connectivity will impact the functioning of the IoT device, which could be catastrophic if it is responsible for home security or health.

With edge computing, data from IoT devices flows to an intermediate node or even a processing module within the device itself that can process the data and provide any action that needs to be executed by the device. In almost all cases, the Cloud is completely bypassed, with a reduction in the distance and time of data transfer. Edge devices can therefore function as independent network nodes that are capable of controlling client IoT devices whether they are connected to the internet or not.

Features of edge computing

Edge computing has several key features that provide benefits but also challenges to its large-scale implementation with IoT.

1] Edge networks need to be able to be scaled

To support a large proportion of IoT connections, edge computing needs to be able to operate efficiently and effectively at scale. This appears possible in theory, but industry stakeholders have yet to standardize how edge computing works, or release protocols that could be commonly adopted. The scalability of edge computing is currently being questioned due to the wide range of devices, platforms, and networks that make up the Internet of Things. This heterogeneity places a unique demand on edge computing at scale because of the responsiveness required without the robustness, security, and uniform data management processes of a central Cloud data center. With large-scale edge computing, the power, throughput, and security requirements for a range of devices simultaneously are likely to come at the expense of low latency.

2] The reliability of edge computing is vital

Edge computing is advantageous because of the independence of its constituent nodes, which can provide a local service even if there is an internet outage. Edge devices like personal assistants and smart home hubs can control a home network offline.

Also, if one node fails, the whole network does not go down. This means other areas of the network can operate uninterrupted. Like other types of decentralized networks, edge networks need a way of alerting the network owner that a node is down which may need to be monitored by other nodes at a local level.

3] Edge computing has low bandwidth consumption

As edge computing diverts data from being uploaded to the Cloud. It reduces the requirement for bandwidth as it is no longer competitive with the large volume of cloud computing applications. There simply is not enough bandwidth to optimally service the increasing number of applications requiring cloud computing. An autonomous vehicle system is expected to generate over a terabyte of data per second. An edge solution will keep these large volumes of upload data from overloading bandwidth availability.

4] Edge computing can support low energy IoT devices

Data transfer for cloud computing is energy-intensive and will affect the power consumption of typical battery power IoT devices. Transmission of data to the edge is faster and keeps transmission time short and battery consumption low.

5] For optimum IoT functionality, speed is non-negotiable

Edge computing is expected to improve the speed and responsiveness of IoT networks because the key processing functions are executed closer to IoT devices. Edge platforms have been demonstrated to exceed the performance of Cloud systems with extremely short response times achieved by well-designed networks. This makes edge computing a candidate for supporting autonomous vehicles.

Edge computing systems have enhanced efficiency due to the proximity of processing power to the end device. Edge networking interacts well with AI technologies that may be used in data processing at the edge of the system.

As edge computing exists between IoT devices and the internet, it has the potential of providing wider efficiency savings in taking traffic offline. Local networking solutions for data processing bypass the internet and can provide bandwidth savings that will provide an uplift in performance.

7] The privacy and security of edge computing is a mixed picture

Edge computing potentially is the answer to the privacy concerns associated with centralized “Big” data. As data is processed locally, end-users have more ownership and control. Because edge computing is distributed, data will be transferred via different types of nodes with differing levels of connection to the internet and encryption which may affect vulnerability to an attack. Cloud security is far more mature and sophisticated, with a centralized infrastructure and clear trust model. Novel, comparative encryption and trust models need to be developed for edge nodes that operate at a local level.

Examples of real-world edge computing IoT solutions

Distributed computing has the potential to make IoT applications extremely responsive, with reaction times that are comparable to those of humans. If low latency can be achieved by utilizing the edge, key IoT applications like autonomous vehicles could be realized. Here are some examples of Edge/IoT applications that are being developed or already deployed.

[A] Edge computing + IoT for haulage vehicles

Edge computing is being explored as a solution for delivering vehicles that have automated responses to road conditions and hazards. A key example of this technology is the development of systems to control the platooning of truck convoys, using a combination of IoT sensors and edge computing carried onboard the vehicles.

Automated platooning enables the synchronized acceleration or braking of a convoy of HGVs. This means that they can travel closer together which enhances fuel efficiency and safety, as well as reducing pollution and congestion.

Vehicle telemetry and other sensor data would be transmitted wirelessly to an edge computing node that would be able to evaluate and adjust the positioning and speed of the vehicles to maintain the platoon.

[B] Edge computing + IoT for the Oil and Gas Industry

Telecommunications and networking are critical areas of investment in the oil and gas sector. Oil and gas sector assets are often in remote, inaccessible, or frankly hazardous locations. They may also be moving by road, rail, or sea. The use of remote monitoring via IoT applications is advantageous as personnel do not need to directly access components of pipelines or oil platforms that are failing. Asset monitoring for oil and gas sector equipment can use edge computing which is advantageous due to its proximity to the assets being monitored using sensors or SCADA in surface or subsurface facilities. Low-latency edge computing also reduces this sector's reliance on expensive satellite internet connectivity to function. Real-time monitoring can take place with timely alerts or action implemented if there is a failure in the equipment or components being monitored. Data can also be related rapidly to personnel who can take direct action where necessary.

[C] Edge computing + IoT for 5G virtualized radio access networks (vRAN)

Radio Access Networks (RAN) is a key component within cellular networks. Like a router, the RAN will provide and distribute access to the cellular network. They provide connectivity to end-user devices by using a transceiver to make a connection to the mobile core network, now using 5G New Radio connections. The 5G network provision is more complex than previous generations due to its inbuilt massive machine-type communications provision for IoT networking. This adds switching complexities that are reliant on algorithms and significant processing power to provide competent service delivery.

For this reason, many parts of the 5G cellular networks are virtualizing their RAN. Virtualization relies on edge computing to provide the processing required near each cell tower for a low latency service.

[D] Edge computing + IoT for traffic management

As the Smart Cities agenda gains momentum, developers are exploring how edge computing can be integrated into intelligent transportation systems and road networks. For this type of IoT networking to be safe and effective, low latency is key. Edge computing processes the data near to the vehicles, traffic lights, signage, drivers, and pedestrians that generate it. It has the capability of matching human response times which is advantageous for rapid alerts and notifications of hazards. Using local computing resources, rather than the Cloud, edge computing can also harness data from sensors monitoring the flow of traffic and other parameters in the transportation system to control lane openings, traffic lights, and effect diversions.

[E] Edge computing + IoT for smart homes

Edge computing is already used as a solution for effectively managing the sheer volume of data generated by smart home sensors. Sensor data like room temperature, lighting, or the position of locks are increasingly diverted from central Cloud-based processing to local hubs based within the home. An example of this is the Amazon Alexa personal assistant that can be used to control ZigBee certified devices. The Alexa assistant is now able to operate as a hub, receiving data inputs from ZigBee devices and sending instructions to the networked devices in the home. Operating as an edge device, Alexa can fulfill its monitoring and control function even if there is a loss of internet connectivity.

Robust and reliable wireless connectivity is essential for the optimum performance of every edge computing + IoT solution

For the majority of IoT devices, high-speed data transfer will be wireless. This means one of the most critical components in an IoT product or solution is the antennas. As one of the Southwest's leading wireless networking equipment suppliers and distributors, we are on hand with the quality and expertise required to supply the performance antennas, cables, and other IoT networking equipment that make your solution work optimally. With a near-comprehensive inventory providing broadband radio spectrum coverage, we can supply antennas for IoT solutions that use the leading wireless technologies, including:

« Back to Blog