The Different Flavors of Edge Computing

Edge computing is slippery and undefined. The only statement that can be applied to all things Edge is it’s an environment designed to move data processing nearer business operations and the consumer.

The challenge with Edge is it has many variations, with professionals focused on Enterprise computing regarding it as an evolution of the distributed ‘lights out’ data center concept.

No matter how intelligent the end-point all Edge approaches share the same architecture. Core data center(s) with satellite locations that store and process data and interact with end-points. The number of Edge layers varies however, there is no standard Edge definition. There is no standard Edge environment. Edge is defined by each business and enabled by application architecture. 

Edge consists of network gateways, data centers and all things IoT. The purpose of the Edge is to deliver distributed application services, provide intelligence to the end-point, accelerate performance from the ‘core’ or collect and forward data from the Edge end-point sensors and controllers. 

Edge footprint can be the size of sensors and controllers, a small number of network/server racks, a container full of equipment or a large air-conditioned data center.  

Edge common attributes include equipment diversity, distributed remote locations, and ‘lights-out’ with no local support.

Edge Delivers

  • Data-stream acceleration, including real-time data processing without latency. 
  • Smart applications and devices to respond to data almost instantaneously, as its being created, eliminating lag time. 
  • Efficient data processing of large amounts of data processed near the source.
  • Reduced internet bandwidth usage eliminating costs, ensuring applications can be used effectively in remote locations.
  • Data processing without placing it into a public cloud adding a layer of data security
  • Better customer experience

So, What is Edge Computing? 

  • The Edge is a place, it’s where things are, and it’s not the data center or the cloud
  • The Edge will house the IT products that are in data centers and clouds
  • The Edge is not a hardware stack, it is defined by how software is architected and used to deliver business value 
  • Edge computing implementation varies wildly, defined by a broad range of use cases spanning industries and companies within industries
  • Companies will find Edge a major advantage and a major challenge as it changes how IT is delivered and managed
  • Edge didn’t just happen – for many large enterprises distributed IT was already being delivered however, it’s growing and becoming more business critical, demanding increased attention, investment and focus

Edge Computing Market Definition

The absence of an agreed and accepted Edge computing definition demanded we create our own. This has resulted in the Edge computing market being split into three different types of use-case:

  1. Enterprise Edge Computing. Remote ‘Lights Out’ Edge Data Centers.  Industrial Edge (control systems, self-contained environments). This use-case is sometimes referred to as Operational Technology (OT) but this narrative is narrow and just focuses on the unification of Data Center Information Management (DCIM) and systems monitoring.  The Remote ‘Lights Out’ Edge Data Centers use-case is a broad market where most enterprises reside. It is the closest relation to current data center and cloud environments. It can be a small equipment rack in multiple remote locations or multiple large data centers. It is the most diverse, non-standard Edge environment, requiring new organizational models, sophisticated software application architectures and a high level of abstraction to visualize, deliver low touch control and the ability to scale and manage a heterogenous, mix of equipment. 
  2. Container IT Edges. (Self-contain – cell tower, reselling something they failed at a decade ago ‘data center in a box’ – micro-data centers). This is left to the customer’s imagination, but it is where converged systems live. This Edge ‘in a box’ approach consists of a solution stack comprising of one or more of the following; servers, OS, storage, network and optimized power and cooling to support all the equipment in the contained environment. The containers are highly standardized however, customization is available to suit specific Edge requirements with options for additional components. This Edge option will continue to grow as it provides cost effective, scalability and the ability to place an Edge near the consumer no matter where they are.
  3. Internet of Things (IoT), where highly available processors enable real-time analytics for applications that can’t wait too many milliseconds to render decisions. IoT end-points continue to get smarter with greater ability to work independently and make decisions without regular communication with a core platform. They are becoming self-aware, discovering other IoT systems and working together to provide greater value. IoT end-points exist everywhere and can scale to millions of devices. 

The Edge content in this blog site focuses on the first definition. This is Edge Computing not driven directly by IoT or 5G wireless networks. It’s an Edge that for many enterprises has existed for years providing data and communications at remote locations and managed either using local staff or by the IT NOC team. However, the business importance of the remote locations continues to grow as more data and data processing moves to the remote locations that were once managed ‘good enough’. For these enterprises, especially the ones who had spent years optimizing the IT support organization and moved to cloud, the Edge challenge is going to get serious.

Leave a Reply

Your email address will not be published. Required fields are marked *