Architecting Edge Computing Infrastructure Is Simplified With New Vertiv Report and Digital Configuration Tool

Edge computing is a simple concept. Bring processing and storage closer to devices and users to better manage the tsunami of data being generated and consumed in your business and enable new digital applications.

But its execution is not so simple. Edge use cases have different requirements that must be considered in your Edge strategy. And edge computing sites are deployed in vastly different physical environments. While some edge use cases may support IT racks in a regional collocation facility, others require placing it at the back of a store, on the factory floor, or on a city street corner.

This complexity is compounded by the decentralized nature of edge computing. Organizations embarking on an edge computing strategy, and our research indicates that roughly half are currently – typically, they will need to deploy multiple edge sites to achieve their goals, increasing the importance of standardized, intelligent edge infrastructure.

Using archetypes and models to configure the edge computing infrastructure

As the network perimeter expands, the need to simplify the process of configuring and deploying edge computing sites increases. The first step in that process was to categorize edge use cases based on their latency, bandwidth, availability, and security requirements. Vertiv tackled that challenge define four archetypes of edge computing.

With archetypes, your organization can quickly identify the key attributes of your perimeter network based on whether your use case is primarily data-intensive, human latency sensitive, machine-to-machine latency sensitive, or critical to performance. life.

The next step, which we recently completed, was to define edge infrastructure models that allow for practical decisions around the physical infrastructure.

graphic Vertiv

Each model plays a role in supporting various edge use cases and has different infrastructure requirements. Regional edge data centers, for example, are being used to support data-intensive and human latency-sensitive use cases such as high-definition content delivery and cloud gaming.

For use cases that require lower latency than a regional edge data center can provide, distributed edge data centers and micro edge sites will be required. The Distributed Edge Data Center model offers higher availability than can normally be achieved in Micro Edge, but sacrifices some of the latency reduction provided by that model. Micro Edge sites allow deployment closer to data sources to enable the ultra-low latency that many use cases require.

Simplify the configuration of the edge computing infrastructure

Through the lens of archetypes and models, we can now more efficiently focus on the key factors that need to be addressed to effectively configure the physical infrastructure required to support edge computing. These include:

  • Case of useUse case availability and latency requirements dictate which data center models should be deployed.
  • Location and surroundings: The data center model dictates the location and environment in which edge computing will operate.
  • Number of racks: The closer the Edge Computing facility is to the point where data is consumed or generated, the less processing and storage is typically required.
  • Power requirementsBased on the number of racks and the physical environment of the site, the power requirements can be determined and the workload and physical characteristics of the infrastructure required to support a specific use case can be clarified.

If you are thinking that it would be nice to put all that information into a digital tool that can determine the specific requirements of your use case, we are one step ahead of you. Along with our new report, Archetypes 2.0: Deployment Ready Edge Infrastructure Models, we have released a configuration tool designed to simplify the Edge Infrastructure configuration process based on common use cases. Try it today.

Copyright © 2021 IDG Communications, Inc.

Leave a Comment