The world needs ever faster data processing. Edge computing will help

Few innovations have transformed our relationship with technology like the cloud. This constellation of huge data centers can tackle computing tasks for digital devices located a mile or a continent away, allowing us to enjoy podcasts, share memes or stream movies nearly anywhere we go.

The cloud, however, has a problem. It’s constantly connecting to new devices — from smartphones to tiny sensors — and offering them more and more services, from e-commerce purchases to turnstile access at a subway. That means more and more data and computing is being shuttled back and forth across the cloud’s connections, and more is coming. Think of how much data self-driving cars will require and how quickly they’ll need to make decisions. There’s just too much happening to rely solely on a central node for the kinds of technology — artificial intelligence, machine learning — that enables these innovations.

Enter edge computing, where more data flow and computation happen at the edges of a network instead of just in the center. Edge computing isn’t new. Networks, by definition, have “edges” that can be used to communicate with machines and users outside the network itself, but now more computing is being done there — when we tap our cards to make a purchase or catch a train, or when we cruise through a tollbooth without tapping the brakes.

Two technological shifts are behind this push to make more use of the network edge. The first is the shift from private to public (and hybrid) clouds. Public clouds are computing and storage for rent that enable businesses to scale up their services without having to build their own data centers. The other shift is the move to 5G wireless data networks, which dramatically speed up the ability to move data through the air to and from devices like smartphones.

For companies that run their own network, such as Mastercard, these shifts are profound — and also exciting, because they open new possibilities for ways to serve people around the world. In the past, someone swiping a debit card in a shop might have been willing to wait several seconds for the network to confirm a bank balance and run security protocols. Merchants, too, might not have expected to glean much information from that purchase, let alone get that intelligence delivered immediately.

Those days are over. As more of our lives have shifted online, we’ve come to expect instantaneous service wherever we interact with a device, app or webpage. Merchants expect to understand their customers’ behavior in real time, and banks expect fraud and credit decisions to get smarter while happening faster. The solution is to constantly evolve the network to make the edge do more work.

Mastercard’s global network, for example, processes tens of billions of transactions a year, linking merchants, banks, government agencies, fintechs and open banking networks. The network started out processing payments for banks and retailers but has expanded into peer-to-peer transactions, digital identity verification, loyalty programs and security and fraud detection. This network has relied on edge computing from the start to cut down on response time, but also to add resiliency by creating backup pathways for data to flow through.

Public cloud and 5G have increased consumer expectations for speed in our digital interactions, but they also allow for a range of innovations.

For example, to enable public transit riders in London, New York and elsewhere to tap with their card from a mobile device or card, Mastercard worked with turnstile manufacturers to embed computing into the turnstiles themselves. That pushed much of the decision-making to the very edge, where riders are, instead of having the card data run all the way home to the network’s center and back again. The result is that riders know if their card is valid in less than a tenth of a second. Any longer and lines would start to stretch.

There are applications, though, with even lower tolerances for lag. Congestion pricing is being used by more cities as a way to discourage drivers from bringing their vehicles into city centers during rush hour. And with 5G and edge computing combined, cars and trucks driving along the tollway can make payments instantaneously.

The increasingly complex business of detecting and preventing fraud also makes a robust edge-computing strategy more vital than ever. Mastercard has stopped tens of billions of dollars in fraud using advanced techniques like machine learning, something no human or rules-based system could replicate. Pushing more of those advanced technologies close to the customer means a faster response based on more relevant data.

Someday soon, we won’t just use our phones or wallets for commerce. Cars and kitchen appliances will play a role, too — recommending when it’s time for new tires, an oil change or a gallon of milk. Managing all that data and creating a unified experience across all of these devices will be a huge challenge. Computing on the edge can make it happen in an instant.