Jargon buster | The edge vs the cloud

Any tech platform you come across will eventually talk about data either in the cloud or the edge. But what are they, and what makes them different? Chris Carson, CEO of smart traffic enforcement startup Hayden AI, breaks down the uses, benefits and risks of both.

What are the edge and the cloud and what makes them different?

Edge computing is computing in the old-fashioned sense: doing all of the computing for any task on the device itself. Computing in “the cloud” means you outsource computing that task to multiple computers elsewhere.

Calling edge computing “old-fashioned” doesn’t make it obsolete – in fact, it’s the opposite. With demand for more and better real-time data, the value of edge computing technologies has skyrocketed. Edge computing combined with artificial intelligence (AI) can unlock powerful and secure real-time insights for a host of different applications.

In what circumstances would you want to use either?

In most circumstances, you would want to use a combination of both. Cloud computing is useful for tasks that don’t require significant bandwidth to send large, detailed data sets to the cloud, or tasks that need powerful computing power that doesn’t fit on any one device – like Siri. Edge computing is better at processing those large data sets – especially high-resolution videos and photos – in real-time.

For example, at Hayden AI, we install sophisticated, purpose built perception systems on buses. These systems capture and interpret video of urban streets as the bus moves, and if our algorithm detects a potential traffic violation, we process it on the edge and send a small evidence package to the cloud.

Processing on the edge means that only a small fraction of the video we capture is sent to the cloud. A large majority of the video is deleted after a short period of time.  This is part of what makes edge processing often more secure than cloud computing.

Can they work together?

Absolutely – and they regularly do. At Hayden AI, once our patented AI-powered edge processor on the bus determines if a traffic violation occurred, the high-resolution evidence package is sent to a third party entity for processing and issuing tickets to drivers. That processing can happen in the cloud.

What are some of the risks associated with the edge and the cloud, and how do you mitigate them?

Cloud computing carries some unavoidable security risks due to the nature of cloud computing itself, meaning that data could potentially be exposed or stolen if cloud clients do not follow shared cloud security responsibility agreements.

Edge computing helps to minimise the overall risk by reducing how much data flows back to the cloud, as more processing is done at the edge or the location where our perception device and data meet. The edge, however, also opens up and widens out an organisation’s attack surface, which poses challenges for most organisations.

At Hayden AI, we continuously map and monitor our “attack surface” – basically any aspect of our hardware or software that connects to the internet in order to watch out for and prevent cyberattacks. We follow a concept internally known as “jointness,” where each of our security tools and operations is designed to support multiple areas of security, not just one.