Edge computing moves compute and storage closer to where data is generated — factories, retail locations, hospitals, cell towers — rather than centralizing everything in distant cloud regions. Latency, bandwidth, and data sovereignty requirements drive this shift.

Why Edge Matters

When milliseconds matter (industrial control, real-time video analytics, autonomous vehicles) or when bandwidth to the cloud is constrained (remote locations, offshore platforms), processing at the edge avoids the round-trip to a cloud data center.

Cloud Provider Edge Offerings

AWS Outposts

AWS Outposts delivers AWS infrastructure — servers, storage, and networking — to your on-premises location. It runs native AWS services (EC2, EKS, RDS, S3 on Outposts) using the same AWS APIs and console.

  • Outposts Rack — full 42U rack of AWS hardware
  • Outposts Server — 1U or 2U for space-constrained locations
  • Managed and monitored by AWS; connectivity back to parent AWS region required

Azure Stack

Microsoft offers multiple edge products:

  • Azure Stack Hub — full Azure services in a disconnected or partially connected environment
  • Azure Stack Edge — hardware appliance with ML inference capabilities at the edge
  • Azure Arc — manage on-premises and multi-cloud resources from Azure control plane without Azure hardware

Google Distributed Cloud

  • Hosted — Google-managed hardware at co-location facilities
  • Edge — customer-managed hardware in air-gapped environments
  • Runs GKE and other Google services

OpenNebula at the Edge

OpenNebula is well-positioned for edge deployments. Its lightweight architecture — unlike OpenStack — runs efficiently on small clusters. The OpenNebula Edge Cloud model provisions cloud resources at remote locations and federates them into a unified management plane, a direct evolution of the distributed computing research done at UCM.

Edge vs. Fog vs. CDN

  • CDN edge — static content caching close to users (CloudFront, Cloudflare)
  • Fog computing — intermediate nodes between devices and cloud
  • Edge computing — compute at or near the data source

Use Cases

  • Manufacturing — real-time quality control, predictive maintenance
  • Retail — local inventory management, in-store analytics
  • Healthcare — medical imaging analysis without sending patient data off-site
  • Telecommunications — 5G network function virtualization at base stations

Back to Blog | Multi-Cloud Strategy