What "edge" and "cloud" actually mean
Edge computing means doing work on a device physically at your site — the edge gateway that reads your PLCs and sensors. It can store data, run logic, raise alerts and buffer locally, all without an internet connection. Cloud computing means doing work on central servers reached over the internet — storing years of history, running heavy analytics, serving dashboards to anyone anywhere, and integrating with business systems. They're not alternatives; they're the two ends of an IoT system, and good architecture deliberately splits the work between them.
The South African clincher: any architecture that can only function when the cloud is reachable will go dark during load shedding and on every flaky link. That's why "where does it run when the internet is down?" is the question that settles most edge-vs-cloud debates here — and the answer has to involve real edge capability.
What the edge is good at
- Resilience — keeps logging and alerting locally through power and network outages, then syncs when they return (the core of our load-shedding resilience).
- Real-time response — local logic acts in milliseconds without a round-trip to the cloud, for time-critical alerts and interlocks.
- Bandwidth & cost — process and summarise locally, sending only what matters; vital on metered cellular or low-power links.
- Privacy & control — sensitive data can be processed or kept on-site.
What the cloud is good at
- Scale & storage — years of history across many sites, without buying servers.
- Heavy analytics — trending, modelling and cross-asset analysis that a small edge device can't do.
- Access anywhere — one dashboard for management, across all sites, on any device.
- Integration — feeding ERP/MES and other business systems from one place.
Side by side
| Edge | Cloud | |
|---|---|---|
| Location | On-site device | Central servers, over internet |
| Works offline? | Yes — local & buffered | No — needs connectivity |
| Latency | Real-time, local | Round-trip to cloud |
| Best at | Resilience, real-time, buffering | Storage, analytics, multi-site access |
| Load shedding | Keeps recording | Goes blind unless edge buffers |
The pattern that wins: both
The healthy architecture is a partnership. The edge collects from your equipment, handles anything time-critical, and buffers everything locally so nothing is lost when the power or network drops. It then syncs to the cloud (or your on-prem server) for long-term storage, heavy analytics, multi-site dashboards and business-system integration. You get real-time resilience and scalable analytics — neither alone delivers both. And "cloud" doesn't have to mean someone else's data centre overseas: in-region cloud or fully on-premise are both valid, depending on your data-ownership and security needs (see industrial IoT security).
This is exactly how the in-house addaNet platform is built: edge devices that read your equipment, buffer through load shedding and act in real time, syncing to in-region cloud or on-prem for storage, analytics and cross-site visibility. It's the architecture that lets a South African operation have both resilience and reach. For the bigger picture, see our guide to Industrial IoT in South Africa.
Frequently asked questions
Is edge or cloud better for industrial IoT?
Neither — they do different jobs and the best architecture uses both. Edge handles local, real-time and resilient processing (and keeps working offline); cloud handles scalable storage, heavy analytics and multi-site access. Edge for resilience and real-time, cloud for scale and analytics.
Why is edge computing especially important in South Africa?
Because of load shedding and patchy connectivity. An architecture that depends on the cloud to function goes blind every time the power or network drops, losing data and alerts. Meaningful edge capability — local logging, buffering and real-time logic — keeps the system working through outages and syncs when connectivity returns.
Do I still need the cloud if I have edge devices?
Usually yes, for what the edge can't do well: long-term storage across many sites, heavy analytics and modelling, dashboards accessible from anywhere, and integration with ERP/MES. The edge keeps you resilient and real-time; the cloud gives you scale and reach. They complement each other.
Does cloud mean my data leaves the country?
Not necessarily. "Cloud" can be in-region (hosted in South Africa) or you can run fully on-premise on your own servers. Data ownership and hosting location should be your decision based on security and compliance needs — a good platform supports both rather than forcing data offshore.
What runs on the edge versus the cloud?
On the edge: reading equipment, local buffering, time-critical alerts and interlocks, and summarising data to save bandwidth. In the cloud: long-term history, cross-site dashboards, heavy analytics and business-system integration. The edge decides and survives; the cloud stores and analyses.