Arsitektur & Komponen Komputasi Tepi

An edge system is organized in tiers that move computation away from a central cloud and toward where data is produced.

• Device tier: sensors and endpoints generating data
• Edge tier: nodes and gateways processing data locally
• Cloud tier: central storage, analytics, and coordination

Edge devices are the endpoints that produce or consume data: cameras, sensors, phones, vehicles, and IoT hardware. They often have limited compute and power but generate large volumes of raw data.

An edge node is a small server located near the devices. It runs application logic, filters and aggregates data, and responds quickly without round-tripping to the cloud.

• Lower latency for time-sensitive decisions
• Reduces bandwidth by sending only summaries upstream

An edge gateway bridges devices and the wider network. It handles protocol translation, security, and buffering when connectivity is intermittent.

Data typically flows: device produces a reading, the edge node processes or filters it, and only meaningful results travel to the cloud.

• Raw data stays local where possible
• Aggregates and alerts go upstream
• The cloud sends config and models downstream

An edge node might average sensor readings locally and only push the result upstream, cutting bandwidth dramatically.

The core advantage of edge architecture is locality. By deciding close to the data source, the system reacts in milliseconds, which is essential for use cases like industrial control or AR.

Managing many distributed edge nodes requires orchestration. Lightweight platforms (such as k3s or container runtimes) deploy and update workloads across nodes from a central control plane.

Edge nodes must keep working when the network drops. They buffer data locally and sync once connectivity returns, providing resilience in remote or mobile environments.

Each tier needs protection: device authentication, encrypted gateway links, and hardened edge nodes. Because edge hardware sits in the field, physical security and secure boot matter more than in a cloud data center.

A complete edge solution combines all pieces:

• Devices generate data
• Gateways collect and translate it
• Edge nodes process and decide locally
• The cloud coordinates, stores, and trains models

Test your grasp of edge architecture components.

You explored the edge computing architecture: devices, gateways, edge nodes, and the cloud tier. You saw how data flows locally first, how orchestration manages distributed nodes, and why locality, resilience, and security define edge systems.