Computation can occur at several locations
A modern service distributes work according to latency, bandwidth, privacy, reliability, energy and cost requirements.
Device-to-cloud pathway
Device & Sensor→Mobile or Local Network→IoT Gateway→Edge Computing→Cloud Platform→Data Centre
Compare the layers
| Layer or model | Purpose | Typical consideration |
|---|---|---|
| Mobile computing | Provides portable access to applications and data through wireless networks. | Battery, coverage, mobility and interface constraints. |
| Internet of Things | Connects sensors, devices and actuators that exchange operational data. | Device identity, reliability, security and data volume. |
| Edge computing | Processes data near its source before or instead of sending everything to the cloud. | Low latency, resilience, bandwidth and local privacy. |
| Cloud computing | Delivers configurable compute, storage, networking and applications on demand. | Scalability, shared responsibility, cost and availability. |
| Data centre | Provides the physical computing, network, storage, power and cooling infrastructure. | Capacity, redundancy, energy, cooling and maintenance. |
Virtualisation as an enabling layer
HostThe physical computer providing processor, memory, storage and network resources.
HypervisorThe control layer that creates and manages isolated virtual machines.
Virtual machineA software-defined computer with allocated resources and its own guest operating system.
ContainerA lighter isolated environment that shares the host operating-system kernel.
Virtualisation improves utilisation, isolation, portability and recovery, but physical capacity and security boundaries still matter.
System-thinking scenario
A smart traffic camera can analyse urgent events at the edge, send summaries to a cloud platform, archive selected footage in a data centre and provide dashboards to mobile users. Moving every task to one location would increase latency, network traffic or operational risk.