Are you ready to be living on the edge? Edge computing, that is.
The state of technology is so advanced as compared to what it was only half a century back that it seems that engineers have found answers to the deepest challenges posed by nature itself. The speed of communication is beyond the expression and imagination of humans. However, the people who actually work in IoT, networking, and cloud industry know it well that there’s still a way to go.
Traditional datacenter architecture is all about central computing powerhouses, from where information is sent and received across globally spread networks. Here, the larger the distance between the endpoint and the datacenter, the higher the response time. In many applications, this incrementally larger time gap is inconsequential. However, in many others, it’s critical.
Examples? Sure, here are some:
- Virtual reality (VR) and augmented reality (AR) content experiences are more fulfilling when the computation required for rendering the content is carried out close enough to AR and VR devices.
- Autonomous vehicles require near real-time feedback from external networks to make course corrections and avoid collisions
- In IoT, many analytical actions need to be carried out closed to the devices that generate the source data.
- HD video content, if cached closer to large concentrations of people who’re likely to access it, means that providers can avoid large costs of transmission over networks provisioned by third-party carriers.
How is edge computing the answer?
Well, edge computing is all about achieving geographical distribution so that computing power can be taken closer to the endpoints that need it most. So, instead of only relying on a dozen giant datacenters, edge computing provides for the cloud to come closer to places/people/devices where there’s a business case for reducing response times even by a few hundred microseconds.
Why does edge computing matter so much?
Before we answer, some stats:
- By 2020, it’s expected that there will be more than 5,600 million smart sensors and connected IoT devices across the globe.
- The data generated by these devices will be to the tune of 5,000+ zettabytes.
- The IoT market size is expected to reach $724 billion by the end of 2023.
Most of this data will be generated at enterprise endpoints located on the “edge” — such as sensors, machines, smartphones, wearable devices, etc. We consider them located on the “edge” because they’re far away from the central datacenter of the organization.
This massive data can’t simply be relayed to the central server because it could easily overwhelm the entire network. This requires enterprises to implement edge computing so that massive data doesn’t have to be transported to corporate datacenters. Instead, it’s used for advanced operational analytics at the remote facilities, enabling site managers and individuals to act in real time on the available information.
How is global IT adjusting to the edge philosophy?
Any technology that can help solve latency problems can also help with bandwidth problems. Companies understand that they can’t stress their bandwidth, particularly in use cases where it’s a win-win for everyone to perform computing close to endpoints instead of in the central server.
Tech giants — Apple, Google, Amazon, and more — seem to have a lot of focus on running AI on end-user devices instead of in the cloud. There are rumors that Amazon is working on building AI chips that will be integrated into Echo devices, reducing the smart speaker’s dependence on the cloud and delivering quicker voice search results. Google is trying hard to make websites better using the same principles as edge computing. Progressive Web Apps are a good example, with offline-first functionalities. Google Clips is another example where data is kept local and AI comes to your device rather than the data having to go to a server where the AI magic then takes place.
Edge computing — Making futuristic more realistic
“Real-time,” though often used, is still a distant dream for technology. Theoretically, real-time might be impossible to achieve, but edge computing brings tech as close to real-time as realistically possible. And because of that, edge computing has massive applications. Examples:
Oil and gas monitoring
Critical infrastructures such as oil and gas facilities require the highest levels of precautions to avoid system failures that could escalate into catastrophes. Edge computing allows for data from temperature and humidity sensors, IP camera, pressure and moisture sensors, and handheld devices. The data is analyzed, processed, and then sent back to users in near real-time, helping them prevent malfunctions.
Edge video orchestration
Mobile edge computing helps media and entertainment players deliver video feeds instantly to users rather than waiting for video content to be transmitted back and forth on a central network. In large stadiums and music arenas, where enhanced real-time video content is important for the user experience, edge computing is a defining moment.
Consider an unmanned aerial vehicle where even a single degree of a digression from the desired path could cause the vehicle to reach a different destination than intended. In such vehicles, several sensors produce massive data every instant, which needs to be analyzed quickly, so that corrective feedback can be sent back to the devices. Edge computing is the solution here — bringing the most important computing elements close to the device.
A word of advice for enterprises
While companies invest in edge computing, they must also plan to draw out strategic and operational advantages from it. Auto alerts and machine automation can help companies completely eliminate assembly line problems by foretelling network issues, machine breakdown, and infrastructure issues. When edge computing begins to deliver business benefits, companies can roll out the strategy to all endpoints that can benefit from the edge architecture.
Closer to the endpoints
A parting thought — edge computing invariably means that companies get closer to endpoints. This could trigger debates on how much control users would be willing to give up to tech giants. As long as the use cases deliver ample benefits, let’s hope the debate will not blow up into propaganda.
Featured image: Flickr / Jane Boyko