From the course: Understanding Edge Computing in a Cloud Computing World

Evolution of edge computing

From the course: Understanding Edge Computing in a Cloud Computing World

Start my 1-month free trial

Evolution of edge computing

- [Instructor] What is edge computing? Ultimately, it's a way that we're able to leverage a distributed architecture that is decentralized but is able to also leverage the resources of some sort of a centralized service such as a public cloud. Keep in mind that during an edge computing process that we're processing the data as close as we can to the source of the information. This removes a lot of the latency and this actually makes it more simplistic. And at some time the data is transmitted to a data center or a public cloud provider where additional processing occurs. So in other words, it's a way to push data and push processing closer to the entity that's leveraging it but still engage a backend system such as a public cloud. So keep in mind that there's really three basic components. There's a cloud or some other centralized system, it doesn't necessarily have to be a cloud but it typically is, some sort of communication networks such as a 5G cellular network, and the ability to communicate to some sort of device or entity or compute system. And as we'll see later in this video, that can be many different things. So what is the evolution or the start of edge computing? We'll keep in mind that we first started to leverage remote data storage as a way to put information as close to the entity that's going to leverage the data. This allowed us to remove some of the latency or the performance problems that occurred at times when the information was stored far away from those who are accessing it. As it progressed, we also learned that processing was a good idea in some instances to be stored on the edge. And ultimately this is where information would be processed and the compute system would in essence deal with the information that was being stored on the remote device. So instead of transmitting all pieces of information back to some sort of a centralized storage system such as a public cloud, we do the data processing at the device or at the edge as well. Then finally, the ability to deal with coupling. And this is the concept that we're able to take information and processing that's occurring at the edge and couple it with information and processing that's occurring within a public cloud provider and some other sort of centralized system. And the idea being is that they're able to interact one to another and coordinate processing to an essence solve the needs of the business or basically solve the needs of the application. So what are edge devices, anyway? It could be a mobile phone that we all carry. If you think about it, lots of processing and lots of information can be stored on these devices. And they're connected via network, either a cellular network or some sort of a broadband network that you have in your house or at your business. They can be a traditional computer that sits in a data center. So in other words, it doesn't necessarily have to be a small device even though many of the edge devices are small devices. It could be a traditional computer that is in essence taking part in an edge based architecture by providing edge based services that just happens to sit in a data center. It could be a small device such as a raspberry PI or something else where processing is stored locally, has local data storage, has networking, basically most of the capabilities of a low powered computer. Can be a drone such as those that are leveraged in agriculture. In other words, the processing and the data storage may occur within the drone. And because there is latency over the network that's connecting the drone, since it's typically a low bandwidth cellular network, then edge base architecture works much better. And then finally medical devices. The fitness watch that you may be wearing on your wrist, the hearing aid, other things that we deal with in the medical world each and every day have the potential if not to be edge devices now, which many of them are, but to be edge devices in the future. And if you think about it, we live in a world where we can't constantly depend on network connectivity. And so this allows these devices to operate independently, which is a good thing because we have to depend on them constantly. So the basic edge architecture is some sort of a centralized processing system, typically cloud but it doesn't have to be a cloud, can be a centralized data center. And sub components of that processing system and things that exist on the edge would be data storage, the ability to do knowledge processing such as artificial intelligence systems, the ability to do traditional computing and the ability to in essence manage these devices and also secure these devices. So the centralized processing system in the edge base architecture would be called centralized processing. And again, it's anything where there's a one to many relationship between the centralized processing system and the edge based devices. Edge processing would be anything that exists down at the device-based level or something that's designated as the edge computing system which basically would take place and processing all of the capabilities that we list here. Data storage, knowledge processing, compute, also management monitoring, and security. So all this is really combined to create the edge base architecture. So when you think of edge computing, it's not just defining something that exist on one system. It truly defines an architecture that exists between many systems that are configured in such a way that edge computing is the way in which they're approaching the processing system.

Contents