Oct 4, 2017
Qualcomm products mentioned within this post are offered by Qualcomm Technologies, Inc. and/or its subsidiaries.
Are you developing in the cloud? Well, your processes may soon be changing. By the end of the decade, edge computing is projected to become the dominant computing model. Peter Levine of Andreessen Horowitz has boldly predicted that cloud computing as we know it is coming to an end because developers and companies see the benefit of moving from centralized computing processes to decentralized ones.
Luckily, developers working with the DragonBoard 410c from Arrow Electronics and other embedded hardware already possess many of the tools needed to make this shift. With the addition of cutting-edge frameworks like AWS Greengrass, developers can start experimenting with edge-computing applications today. Let’s dive into this new process.
What is the edge?
The edge computing model is an updated way of managing cloud computing systems where the data processing is done at the “edge” of the network, closer to the source of the data. By using local processors and storage as the primary computational device for an application, the time spent relaying all information back and forth from a central, remote data center is reduced.
As computing gets more complicated and handles more data, using local computing power can help lower latency. For example, in artificial intelligence (AI)-powered programs, algorithms make hundreds of decisions a second. In industrial manufacturing, multiple processors take input and coordinate responses simultaneously. In augmented reality (AR) systems, a program synthesizes HD graphics and geospatial data in real-time. The latency in a cloud-based computation model is inefficient for these types of applications, which will become even more apparent with the increase in processing power for local devices.
That said, cloud computing is not obsolete. The cloud informs the edge, empowering it with important data and instructions. Whether informing local devices through deep learning models it has previously trained, or supplying application-specific contextual data (such as feature maps for autonomous vehicles), the cloud still serves many critical functions.
Edge your bets
Of course, these predictions could be completely off, but we’re willing to bet the edge is where computing is headed. At Qualcomm Technologies, we’ve been talking about the importance of the edge for years. We’ve seen how the requirements of recent technology trends are driving the modality shift to edge computing, and especially in these key areas:
IoT, which is synonymous with edge computing
Autonomous systems, including cars, drones, robots, medical or industrial machines, will have edge-based systems
The AI that could impact many of our daily experiences will be composed of calculations on edge systems
The vast computations that make VR and AR systems such a natural-feeling experience will also be edge-based
We’re not the only ones who feel this way: venture capital agrees. Peter Levine from Andreessen-Horowitz has provocatively suggested that the age of edge compute is taking over the cloud.
“The current model of cloud computing [is] too slow. A small difference in the time it takes to refresh a machine learning model for a drone or car could be the difference between life and death. Computation will move to the edge. The same drones, cars, and IoT devices that need their models updated quickly will form a peer-to-peer network with which to distribute time-sensitive tasks...cloud servers [will] still be around...responsible for doing offline computation across large data sets.”
Get an edge on the competition
Since edge computing is still in its early stages, now is a great time to get started as a developer. One tool that can help you get rolling is the DragonBoard 410c. Thanks to a few key features, this development board is a good entryway to utilize the established ecosystem of sensors and tools. However, developers working with heavier edge computing loads may want to consider using the Qualcomm Snapdragon mobile platform, since this opens up access to a wider range of tools such as the Qualcomm Snapdragon Neural Processing Engine.
To get started with edge computing, we recommend AWS Greengrass, which is an especially powerful framework because:
It provides the same programming model locally as used in the cloud, simplifying the process for developers
It’s connectivity independent, meaning data and computations run locally and sync with the cloud when the network is available
It’s compatible with many Qualcomm Developer Network tools including: Mesh, Wi-Fi and cellular network connectivity tools, Snapdragon NPE, Snapdragon Profiler and heterogeneous computing libraries and utilities
Ready to get started?
Ready to dive into edge computing? We have a project on QDN for AWS Greengrass on DragonBoard 410c with set up instructions to get you started today. We hope you enjoy developing on the cutting edge. (Sorry, couldn’t resist!)