Catalog
concept#Platform#Architecture#Integration

Edge Computing

Shifting compute and storage closer to data sources to reduce latency and conserve bandwidth.

Edge computing moves processing and storage closer to data sources to reduce latency, conserve bandwidth, and enable local decision-making.
Emerging
High

Classification

  • High
  • Technical
  • Architectural
  • Intermediate

Technical context

Cloud backends for long-term storage and central analyticsKubernetes-based orchestration (edge Kubernetes distributions)Message brokers and telemetry pipelines (MQTT, Kafka)

Principles & goals

Process as close to the data source as possible.Design for intermittent connectivity and constrained resources.Protect privacy via local data retention and selective synchronization.
Build
Enterprise, Domain, Team

Use cases & scenarios

Compromises

  • Increased attack surface due to distributed nodes.
  • Inconsistent states between edge and central systems.
  • Complexity in update and rollout processes.
  • Filter data locally and synchronize only necessary information centrally.
  • Implement automated rollouts and health checks for edge nodes.
  • Establish secure certificate management and access control.

I/O & resources

  • Sensor and telemetry data
  • Device identity and security information
  • Configuration and policy specifications
  • Local decisions and control commands
  • Aggregated metrics for central analytics
  • Alert and event notifications

Description

Edge computing moves processing and storage closer to data sources to reduce latency, conserve bandwidth, and enable local decision-making. It comprises distributed edge nodes, lightweight platform services and hybrid integration with centralized cloud backends. Common use cases include IoT telemetry, real-time control loops and constrained-network applications.

✔

  • Reduced latency and faster response times.
  • Lower bandwidth usage compared to cloud-only approaches.
  • Improved availability and resilience during network disruptions.

✖

  • Limited compute and storage capacity per edge node.
  • Increased operational effort for distribution and management.
  • Heterogeneous hardware and software landscapes hinder standardization.

  • End-to-end latency

    Measurement of time from event to local/central response.

  • Volume of data to cloud

    Amount of data sent per time unit to central backends.

  • Edge node availability

    Percentage of time edge instances operate correctly.

Autonomous production line

Manufacturing uses local edge controllers for closed control loops and sends only aggregated KPIs to central systems.

On-site surveillance camera analytics

Video analytics runs on-camera or on a local node; only relevant events are archived or uploaded to the cloud.

Branch network with local data retention

Branches keep personal data locally and synchronize anonymized metrics centrally.

1

Define requirements and latency targets and identify suitable edge sites.

2

Evaluate hardware and platform options and deploy pilot nodes.

3

Design network and security architecture for edge-to-cloud connections.

4

Adapt applications for local execution and define data flows.

5

Set up automated monitoring, updates and operational processes.

⚠️ Technical debt & bottlenecks

  • Ad-hoc deployment scripts instead of automated pipelines.
  • Unclear versioning of edge software and modules.
  • Non-standard telemetry schemas across sites.
Network stability and latencyProvisioning of heterogeneous devicesOperational and update workflow
  • Moving sensitive analytics to unsuitable edge hardware without privacy assessment.
  • Disabling full cloud backups, increasing risk of data loss.
  • Running complex consistency logic on severely resource-constrained nodes.
  • Underestimating operational effort for distributed infrastructure.
  • Lack of monitoring and debugging strategy at the edge level.
  • Tight coupling between edge and cloud implementations.
Distributed systems and networkingEmbedded and edge software developmentOperation and automation of distributed infrastructure
Low latency requirementsConstrained WAN bandwidth and cost constraintsData sovereignty and local privacy regulations
  • • Constrained hardware resources at edge sites
  • • Robust security and authentication mechanisms required
  • • Regulatory requirements for data localization