David Jensen - Beginning Azure IoT Edge Computing: Extending the Cloud to the Intelligent Edge

Any source code or other supplementary material referenced by the author in this book is available to readers on GitHub via the books product page, located at www.apress.com/9781484245354 . For more detailed information, please visit http://www.apress.com/source-code .

If you have been paying attention to advances in technology over the past 78 years, you have no doubt heard the term cloud. Cloud has come to mean any processing or data collection that happens somewhere other than on my machine, which gives way to terms like private cloud, public cloud, self-service cloud, managed cloud, blah, blah, blah. With such a broad definition comes a plethora of misunderstandings, misrepresentations, and misguided advice.

To complicate matters more, even if you have navigated through the pitfalls of cloud overuse, there are architectural patterns that are easier to design and implement in the cloud-aware world in which we live, but those, too, have been overrun with vague definitions and meaning. Youve no doubt heard of the Internet of Things (IoT) and Big Data.

And, lastly, one of the latest victims of vague-term disorder is Intelligent Edge. You have most likely heard that term or you wouldnt be reading this book. If this term has lost its significance for you, then read on.

To level set, let me clarify the meaning of cloud for the purposes of this book. In this book, cloud refers to a public, managed platform that uses a consumption-based cost model (pay only for what you use) and supports public-facing endpoints as well as endpoints that are provisioned to only be available within a private network, even though they are hosted on a public platform. This description of cloud also includes a self-service management portal for developers to create, remove, and otherwise manage the resources and services running on the platform.

Shortly after this type of cloud emerged on the horizon, one of the architectural patterns mentioned above, Internet of Things (IoT), also began to gain popularity and a significant amount of attention. Over the past few years, it has been one of the most intriguing technologies and patterns that has come about.

The canonical example of the IoT pattern involves using an electronic device to instrument and monitor a physical, real-world device. Once instrumented, the electronic device transmits measurements from the physical device for collection and analysis in some back-end data store. The building blocks of this pattern have existed for a few decades, but they have not been as affordable and, consequently, widespread as they are now, leading to the massive increase in interest it has generated.

As more and more companies bought into the idea that they could quickly and cheaply implement their own instance of an IoT architecture and the cloud was gaining momentum in other areas, the idea of push everything to the cloud became the mantra of the most edgy or disruptive companies. If you wanted to be known for pushing the envelope and really innovating, then you had better be pushing all your data to the cloud along with the other thousands of companies that were doing the same thing, sometimes without considering whether it was the right approach.

As more and more companies architect solutions that involve the cloud in some way, there are more and more examples of when that cloud or bust mentality breaks down. Not every problem can tolerate the bandwidth and latency required to push everything to the cloud. Not every company can access a public-facing endpoint from their devices.

So, here we are, a decade after the cloud came on the scene, and we are learning that maybe we dont need to push everything to the cloud. Maybe we need to consider when to push data to the cloud and when to process the data where it originates. Enter The Intelligent Edge .

The edge in the intelligent edge refers to the proximity to where the data originates . Intelligent edge computing then indicates that some intelligence has been pushed out from the cloud to a point as close to the data origination as possible. There are examples of this all around us.

Think about cars that automatically apply the brakes for you if they sense an object in front of the car. Or the alerts that sound if you attempt to change lanes while another vehicle is in your blind spot. Those decisions must be kept as close to the edge (data origination) as possible. When deciding whether to stop the car or not, milliseconds of latency could potentially be a life-or-death situation.

Having the option to apply intelligence to data as it originates gives options to developers and architects as they architect cloud-enabled solutions. No longer must all data be sent to the cloud to perform analysis on it. Analysis can now happen on the edge, and based on that analysis, a subset of the raw data streams can be sent to the cloud. This saves bandwidth, time, money, and storage.