Analytics & ML · Study deck

Practical Edge Computing: Workload Placement

A vibration node is a sensor that records machine movement.

Data Dora is your guide for this deck.

edgepatternspractical

Data Dora, the module guide, in a scene from this chapter.

After studying this chapter

Learning objectives

You will be able to:

Use Interactive Latency Tools: Calculate end-to-end latency for different processing locations (device, edge, cloud)
Apply Worked Examples: Follow step-by-step calculations for real-world edge vs cloud decisions
Avoid Common Pitfalls: Recognize and prevent typical edge deployment mistakes
Make Informed Architecture Decisions: Select appropriate processing locations based on latency, bandwidth, and cost requirements

Major section

Start With the Story

The long study needs a saved period, enough detail, and a review trail.

A local computer must warn staff about a warming freezer, while a central service studies months of energy use.
The system owner has to place each job without hiding its cost or its failure path.
Cost is more than the purchase price.

Major section

Start With the Story (continued)

The warning needs a response time, a local safe action, and a maximum age for its input.

Include field visits, power, link use, updates, support time, and the cost of a missed or late decision.
The simple split does not choose a final design.
Under the Hood covers queues, model size, hardware limits, and the standards or service terms that may change the result.

Major section

Phoebe's Field Notes: Why Bit Depth Costs Watts, Not Just Storage

The mathematical gist: The chapter's 1,000 Hz, 16-bit stream is 16 kbps (2,000 B/s) per sensor and has a 98.1 dB ideal converter ceiling.

At a stated 50 nJ/bit, raw transmission costs 0.800 mW per sensor.
Dropping to 8 bits halves radio work but loses 48.2 dB; extracting the chapter's FFT features removes 95% of the stream while retaining 16-bit input evidence.

Major section

Edge IoT Architecture

An edge gateway evaluates and aggregates nearby streams, then forwards only the summaries needed for broader cloud analytics.

Rich industrial systems, constrained devices, and tiny sensors arrive over different local paths, but the gateway gives them one governed crossing into the wider IoT system.
The gateway must preserve the event meaning and source identity needed by the later cloud decision.

Diagram of edge IoT architecture showing sensors connecting to edge gateways for local processing, with filtered and aggregated data flowing to cloud services for advanced analytics and storage

Major section

Data at the Edge Architecture

Four layers of data at the edgeLevel 4: Data accumulation: Persistent storage keeps synchronized history ready for cloud access.

Level 1: Physical devices: Sensors and actuators generate the raw data that starts the pipeline.
Four levels carry device readings through edge reduction to cloud storage and analysis; 1,000 readings become four summaries before moving upward.
The reduction is (1,000 − 4) ÷ 1,000 × 100 = 99.6% fewer values to move onward.
Cloud accumulation then keeps broader history for fleet analysis.

Four-level data-at-the-edge diagram: physical devices send 1,000 temperature readings through connectivity to edge filtering and aggregation. Four summary values remain, a 99.6% reduction; cloud accumulation stores data for fleet analysis and business collaboration.

Major section

IoT System Architecture

Things produce evidence, fog nodes handle nearby decisions, and cloud services provide wider analytics while controlled insight can flow back down.

Their data moves upward to the fog tier, where nearby processing can coordinate devices and make fast local decisions.
The cloud sits above that boundary with internet-scale servers, analytics, and user-facing applications.
Insight or an approved control action can move back toward the physical process, so the path is not a one-way upload pipe.
Each downward step must retain command authority, freshness, and destination identity before an actuator treats the result as current.

Three-tier factory IoT architecture with 500 vibration sensors in Things, 10 edge gateways in Fog, and cloud analytics. The diagram compares a local safety alert near 10–20 ms with a cloud safety path near 150 ms.

Major section

Factory Edge or Cloud Choice

Management needs to decide between edge processing (local analysis with alerts) versus cloud processing (centralized analysis).

The plant has a 100 Mbps internet connection.
Verdict: Cloud path fails the 50 ms safety requirement.
Result: Edge-hybrid architecture saves $4,831 (41%) over 3 years while meeting the critical 50 ms latency requirement that pure cloud cannot achieve.

Deck summary

Key takeaways

The long study needs a saved period, enough detail, and a review trail.

The warning needs a response time, a local safe action, and a maximum age for its input.
The mathematical gist: The chapter's 1,000 Hz, 16-bit stream is 16 kbps (2,000 B/s) per sensor and has a 98.1 dB ideal converter ceiling.
An edge gateway evaluates and aggregates nearby streams, then forwards only the summaries needed for broader cloud analytics.
Four layers of data at the edgeLevel 4: Data accumulation: Persistent storage keeps synchronized history ready for cloud access.

Retrieval practice

Recall check

Data Dora says: answer from memory, then check your reasoning.

Q1A pipeline needs a 20 ms emergency shutdown on vibration sensors; its measured cloud round trip exceeds that deadline. Weekly fleet-wide predictive-maintenance training uses shared history held in the cloud, which individual edge controllers do not have. Where should each function primarily run?

ABoth the shutdown and the model training in the cloud for a single code path

BBoth functions at the edge to avoid any dependence on the network

CShutdown logic in the cloud and model training at the edge for load balance

DThreshold and shutdown logic at the edge; aggregation and model training in the cloud

Show answer

Answer: D Latency-critical safety logic runs at the edge; compute-heavy model training runs in the cloud.

Print reference

Answers

Answer key.

D · Latency-critical safety logic runs at the edge; compute-heavy model training runs in the cloud.