Edge Computing
Edge Computing
Edge Computing & MEC
1. Overview of Edge Computing, Moving Data Processing to the Front Line
flowchart LR
A["Centralized cloud"] -- "Shift to real-time distributed processing" --> B["Edge computing"]
Definition: A distributed computing technology that processes data in real time at or near the physical location where it is generated (the edge), rather than sending it to a central cloud server.
Characteristics: (Low-latency processing) Processes data near where it is generated, delivering real-time response without a round trip to the cloud. (Reduced bandwidth) Local processing reduces the volume of data sent to the cloud, cutting communication cost and network load. (MEC evolution) Combined with 5G/6G as MEC (Multi-access Edge Computing), evolving into core infrastructure for autonomous driving and smart factories.
2. Edge Computing Architecture and Layered Structure
A. The Cloud-Edge-Device Three-Layer Model
flowchart TD
subgraph CLOUD["Central Cloud (Core)"]
C1["Big-data analytics<br/>long-term data storage<br/>training complex AI models"]
end
subgraph EDGE["Edge Layer"]
E1["Edge gateway<br/>MEC server<br/>real-time data processing"]
end
subgraph DEVICE["Device Layer (IoT)"]
D1["Sensors, cameras, terminals<br/>data collection and transmission"]
end
DEVICE <-->|"Real-time control"| EDGE
EDGE <-->|"Summary data transmission"| CLOUD
CLOUD <-->|"Model/policy deployment"| EDGE
style EDGE fill:#E3F2FD,stroke:#1976D2,stroke-width:2px
| Layer | Role and Key Function | Response Speed |
|---|---|---|
| Central cloud | Global optimization, large-scale computation, backup | Hundreds of ms to seconds |
| Edge layer | Local analysis, filtering, immediate control | A few ms to tens of ms |
| Device layer | Data sensing, simple threshold judgment | Instant (real-time) |
B. Core Technology Elements of Edge Computing
flowchart TD
EdgeTech["Edge computing technology"] --> Virt["Virtualization and containers"]
Virt --> LightweightK8sK3s["Lightweight K8s (K3s)"]
Virt --> Docker["Docker-based microservices"]
EdgeTech --> Network["Networking"]
Network --> 5GMECMultiaccessEdgeComputing["5G MEC (Multi-access Edge Computing)"]
Network --> SDNNFV["SDN / NFV"]
EdgeTech --> Data["Data management"]
Data --> FlinkSpark["Stream data processing (Flink, Spark)"]
Data --> DB["Distributed databases"]
EdgeTech --> Sec["Security"]
Sec --> ZT["Zero Trust edge security"]
Sec --> FederatedLearning["Federated Learning"]
| Technology Element | Details | Notes |
|---|---|---|
| MEC | Provides compute capability at the edge of the mobile network | A core 5G technology |
| Micro-DC | Small-scale regional data centers | Physical infrastructure |
| Edge AI | Performs inference on edge devices | Linked to on-device AI |
3. Expected Benefits and Key Use Cases of Edge Computing
| Category | Key Expected Benefits | Practical Application and Use Cases |
|---|---|---|
| Performance optimization | Dramatically reduces response latency | Autonomous vehicles (V2X), real-time process control in smart factories |
| Cost reduction | Minimizes the volume of data sent to the cloud | Analyzing and filtering high-resolution CCTV footage in smart cities |
| Privacy | Local processing and storage of sensitive data | On-site processing of medical data, smart-home privacy protection |