Edge Computing

Edge computing is a distributed information technology (IT) architecture in which client data is processed at the periphery of the network, as close to the originating source as possible.

Data is the lifeblood of modern business, providing valuable business insight and supporting real-time control over critical business processes and operations. Today's businesses are awash in an ocean of data, and huge amounts of data can be routinely collected from sensors and IoT devices operating in real-time from remote locations and inhospitable operating environments almost anywhere in the world.

However, this virtual flood of data is also changing the way businesses handle computing. The traditional computing paradigm built on a centralized data center and the everyday Internet isn't well suited to moving endlessly growing rivers of real-world data. Bandwidth limitations, latency issues, and unpredictable network disruptions can all conspire to impair such efforts. Businesses are responding to these data challenges through the use of edge computing architecture.

In simplest terms, edge computing moves some portion of the storage and compute resources out of the central data center and closer to the source of the data itself. Rather than transmitting raw data to a central data center for processing and analysis, that work is instead performed where the data is generated -- whether that's a retail store, a factory floor, a sprawling utility, or across a smart city. Only the result of that computing work at the edge, such as real-time business insights, equipment maintenance predictions, or other actionable answers, is sent back to the main data center for review and other human interactions.

Thus, edge computing is reshaping IT and business computing. Take a comprehensive look at what edge computing is, how it works, the influence of the cloud, edge use cases, tradeoffs, and implementation considerations.

Edge computing brings data processing closer to the data source.

Edge computing is all a matter of location. In traditional enterprise computing, data is produced at a client endpoint, such as a user's computer. That data is moved across a WAN such as the internet, through the corporate LAN, where the data is stored and worked upon by an enterprise application. The results of that work are then conveyed back to the client endpoint. This remains a proven and time-tested approach to client-server computing for most typical business applications.

But the number of devices connected to the internet, and the volume of data being produced by those devices and used by businesses, is growing far too quickly for traditional data center infrastructures to accommodate. Gartner predicted that by 2025, 75% of enterprise-generated data will be created outside of centralized data centers. The prospect of moving so much data in situations that can often be time- or disruption-sensitive puts incredible strain on the global internet, which itself is often subject to congestion and disruption.

So IT architects have shifted focus from the central data center to the logical edge of the infrastructure -- taking storage and computing resources from the data center and moving those resources to the point where the data is generated. The principle is straightforward: If you can't get the data closer to the data center, get the data center closer to the data. The concept of edge computing isn't new, and it is rooted in decades-old ideas of remote computing -- such as remote offices and branch offices -- where it was more reliable and efficient to place computing resources at the desired location rather than rely on a single central location.

Edge computing is closely associated with the concepts of cloud computing and fog computing. Although there is some overlap between these concepts, they aren't the same thing, and generally shouldn't be used interchangeably. It's helpful to compare the concepts and understand their differences.

One of the easiest ways to understand the differences between edge, cloud, and fog computing is to highlight their common theme: All three concepts relate to distributed computing and focus on the physical deployment of compute and storage resources in relation to the data that is being produced. The difference is a matter of where those resources are located.

Edge Computing is the deployment of computing and storage resources at the location where data is produced. This ideally puts compute and storage at the same point as the data source at the network edge. For example, a small enclosure with several servers and some storage might be installed atop a wind turbine to collect and process data produced by sensors within the turbine itself. As another example, a railway station might place a modest amount of computing and storage within the station to collect and process myriad track and rail traffic sensor data. The results of any such processing can then be sent back to another data center for human review, and archiving and to be merged with other data results for broader analytics.

Cloud Computing is a huge, highly scalable deployment of computing and storage resources at one of several distributed global locations (regions). Cloud providers also incorporate an assortment of pre-packaged services for IoT operations, making the cloud a preferred centralized platform for IoT deployments. But even though cloud computing offers far more than enough resources and services to tackle complex analytics, the closest regional cloud facility can still be hundreds of miles from the point where data is collected, and connections rely on the same temperamental internet connectivity that supports traditional data centers. In practice, cloud computing is an alternative -- or sometimes a complement -- to traditional data centers. The cloud can get centralized computing much closer to a data source, but not at the network edge.

Fog Computing environments can produce bewildering amounts of sensor or IoT data generated across expansive physical areas that are just too large to define an edge. Examples include smart buildings, smart cities, or even smart utility grids. Consider a smart city where data can be used to track, analyze, and optimize the public transit system, municipal utilities, and city services and guide long-term urban planning. A single edge deployment simply isn't enough to handle such a load, so fog computing can operate a series of fog node deployments within the scope of the environment to collect, process, and analyze data.

It's important to repeat that fog computing and edge computing share an almost identical definition and architecture, and the terms are sometimes used interchangeably even among technology experts.