Learning the Edge Levels

Unlike the cloud or virtualization, edge computing has a multitude of levels each with their own concept of operations (ConOps) and use cases. This added complexity moving back to decentralized computing stratifies the data custody model. With each OEM and vendor pushing their own angle of marketing and presentation, it can be hard to see the bigger picture. Thankfully the following Forbes article and framed summary can help simplify the tech stack and acquisition dynamics. Understanding the ecosystem can help ratify your orgnaization's edge compute initiative as a top-down, or bottom up approach for both technical and human systems.

Classifying The Modern Edge Computing Platforms, Forbes Article

Enablement Platform

 ConOps & Use Cases


The Cloud
  •  Centralized computing era, place all services, apps, and data in a single place
  •  Infinitely scalable, elastic compute resource pool
  •  Software defined everything, delivering IaaS, PaaS, and SaaS capabilities in OPEx acquisition models
  •  Examples: Amazon's AWS, Microsoft's Azure, Google Cloud Platform, and the IBM Cloud


Cloud Edge
  •  Does what CDN did to static content but for distributed dynamic workloads such as containers and microservices
  •  Enables distributing components of an application across multiple endpoints to reduce the latency involved in the roundtrip
  •  Delivered as a managed service, customers won’t have to deal with the hardware and software maintenance
  •  With increased investments in hybrid and multi-cloud environments based on Kubernetes, mainstream cloud providers may eventually offer cloud edge services as an extension of their existing platforms
  •  Examples: Section.ioVolterra.ioMimik, and are some delivering cloud edge capabilities.


  •  MEC moves the computing of traffic and services from a centralized cloud to the edge of the network and closer to the customer
  •  With 5G becoming a reality, MEC is becoming the intermediary layer between the consumers and providers of the public cloud 
  •  The edge infrastructure runs within the facility of a telecom provider co-located within a data center or even a cellular tower
  •  Examples: Telco companies such as Verizon MEC or AT&T MEC, and T-Mobile


Heavy Edge
  •  Typically a hyperconverged infrastructure (HCI)
  •  Comes with an all-in-one hardware and software stack typically managed by a vendor
  •  Demands power and network resources that are available only in an environment like an enterprise data center
  •  Doubles up as an IoT gateway, storage gateway, AI training and inference platform 
  •  Comes with an array of GPUs or FPGAs designed for managing end-to-end machine learning pipelines including the training and deployment of models
  •  Examples: AWS Snowball Edge, Azure Stack EdgeNVIDIA EGX A100 and Nutanix Acropolis

Medium Edge
  •  Represents a cluster of inexpensive machines 
  •  Compute cluster is powered by an internal Graphics Processing Unit (GPU), Field Programmable Gate Arrays (FPGA), Vision Processing Unit (VPU), or an Application Specific Integrated Circuit (ASIC)
  •  A cluster manager like Kubernetes is used for orchestration of the workloads and resources in the clusters
  •  Examples: a Kubernetes cluster of Intel NUC machines or Zotac Mini PC connected to NVIDIA GPUs or Intel Movidius VPUs deployed in a retail store or a restaurant running NVIDIA CUDA/TensorRT or Intel OneDNN/OpenVINO models for AI acceleration

Mini Edge
  •  Runs in disconnected, mobile and remote environments such as trucks, vessels, air crafts, and windmills
  •  Capable of running a full-blown operating system like Linux or Windows
  •  Based on a single board computer with either ARM64/AMD64 architecture
  •  Often uses a software stack and AI accelerator to speed up the inference
  •  Ideal for protocol translation, data aggregation, and AI inference
  •  Examples: NVIDIA Jetson Nano module running NVIDIA JetPack and TensorRT or an X86 board with Intel Movidius Myriad X Vision Processing Unit (VPU) with Intel’s OpenVINO Toolkit

Micro Edge
  •  Sensors connected to the microcontroller generate the telemetry stream that is used by a deep learning model for inference
  •  The most recent incarnation of the edge computing layer running within the context of a microcontroller and a microprocessor
  •  When a microcontroller is capable of running a TinyML AI model, it qualifies to be a micro-edge computing device such as OneTech's MicroAI capable of running on a programable logic controller/microcontroller
  •  Examples: TensorFlow Lite running on a combination of ARM Cortex-M processor co-located with a microcontroller such as nRF52840 or Apollo3 Blue,