“Only when the information has been processed and refined is the data sent to the cloud, if at all,” the company continues. “Edge computing is becoming more and more relevant with the growing popularity of the Internet of Things (IoT).”
Testifying to this is a 2018 CB Insights report – ‘What Is Edge Computing?’ – which notes edge computing’s particular relevance in the realm of autonomous vehicles. “An autonomous vehicle is essentially a large, high-powered computer on wheels that collects data through a multitude of sensors,” says the report. “For these vehicles to operate safely and reliably, they need to respond to their surroundings right away. Any lag in processing speed can be deadly.”
Pros and cons
What makes edge computing an effective data processing option is that it creates an environment that acts as a hybrid between cloud and local processing – combining the key attributes of both.
“The cloud brings flexibility, scalability and services decoupled from hardware to improve end-user experience,” says Dalia Adib, edge computing practice lead at STL Partners. “But processing locally reduces the amount of data traversing through the network, decreases bandwidth costs and ensures latency is kept to a minimum to support mission critical services. As data privacy concerns grow, edge computing makes it easier for businesses to manage their data and avoid it being stored in remote public clouds.”
While businesses that adopt edge computing are looking for enhanced application performance and reduced costs, they must also be aware of its limitations.
“The challenge is that there are different edges, including device edge, on-premises and network edge, and each has its pros and cons,” explains Ms Adib. “The network edge can dramatically decrease latency to 5 to 10 milliseconds, but could potentially be more costly for a developer to use than the cloud. It is more difficult to benefit from economies of scale and the number of edge locations is limited, at least in the short term. In reality, latency is affected by many factors – hardware (processing power), location, application and architecture, among others – and determining an IT architecture is not a straightforward decision.”
While the adoption of edge computing is generally viewed as being at an early stage, its burgeoning status as a transformative business operation suggests it is on the cusp of mainstream adoption. Indeed, according to the International Data Corporation (IDC), by 2022, 40 percent of companies’ cloud deployments will include edge computing.
“The edge computing market is not new – we already have applications running on various edges, such as device edge, on-premise edge and CDN edge,” says Ms Adib. “For example, Android phones run machine learning (ML) models on the device to adapt the keyboard based on user behaviour and only aggregated data is sent back to the data centre.
“Manufacturers are using edge computing at their production facilities to run IoT applications, for applications like predictive maintenance,” she continues. “What we will see in the next five years is growth in network edge – edge servers running at locations on the telecoms network. This is a nascent market and very few operators have announced any live, commercial deployments yet.”
A potentially powerful force for the future of IT and business, edge computing, despite its history, is still viewed as a new paradigm by many – a mechanism to reduce latency and streamline data traffic which may, in time, replace cloud computing as the favoured data storage solution.
© Financier Worldwide