Is 2018 the Year of Edge-Computing for Apps?

The cloud is going through a bit of a transformation right now. Electronic manufacturers are pushing for more edge computing. But it’s not just physical electronics that are facing the challenge of moving away from the cloud for split-second computing; almost all mobile app developers use the cloud to process data from a user’s interactions with their app.

This shift towards in-product computing isn’t new, but it’s becoming more necessary to create a faster future.

Speed Is the Name of the Game

Even though deep learning is advancing every day and GPUs are more powerful than ever, it still takes too long to utilize incoming data. The issue lies in the time (and therefore delay) it takes to send up to hundreds of megabytes (MB) of data to the cloud per second.

A couple of seconds can make all the difference. Self-driving cars regularly generate roughly 100 MB of data per second. This needs to be analyzed and acted upon instantaneously. Enter edge computing.

Edge computing refers to data processing at the edge of networks, at the sensors that are on the electronic device itself. In the case of apps, edge computing usually occurs on the mobile device.

Keeping the Cloud Around

It is important to still have the cloud there to store data in case of emergencies or problems; it should remain an integral part of new hardware so that it can be utilized by data scientists and AI experts to continue improving their technology, a type of “flight recorder” for software.

As AI and hardware continue to improve and become more accessible to the general public, it’s more imperative to set up industry-wide best practices that handle the cloud and edge computing as they should be: distinctly separate but intertwined.

The Promise of Edge Computing

While data scientists work quietly in the background operations of most companies these days, data storage, processing, and analysis are becoming more indispensable than ever before. Since response time correlates with the method and means of data storage and processing, data science is evolving towards optimization over automation for edge computing.

Even though it should be instant for your Tesla to detect a possible collision when you’re in New York and the data center is in San Diego, the technology isn’t quite up to speed yet. But it’s getting there.