JM

Network Throughput Test

A web app test that lets network engineers emulate realistic application traffic mixes to validate device performance and policy enforcement under load.


Brief

Modern, app-aware network devices need to recognize traffic accurately and maintain performance under load. At the time, most test tools relied on scripts or dense configuration panels, making it difficult to assemble realistic traffic mixes or run repeatable tests without deep protocol expertise.

We set out to streamline that workflow—making it easier to build representative mixes, ramp load in a controlled way, and observe how devices respond. The goal was to reduce setup friction while revealing bottlenecks and classification issues more clearly.

Early reviewers noted how unusual it was to see a throughput test presented through a clean, approachable UI—one even joked that it “made network testing look almost… sexy.” It confirmed that thoughtful design could meaningfully elevate even the most technical tools.

Jobs to Be Done

Job Performer

Network hardware engineer (primary); validation specialist or lab manager (secondary).

  • Assemble representative application-traffic mixes for a target device under test
  • Ramp and control load to observe throughput limits and policy behavior
  • Capture and export results in a repeatable format for comparison

User Group

Vendor and enterprise lab teams characterizing modern, app-aware devices. Prior tools were script-heavy, OS-specific, and rigid—making it slow to build realistic mixes and hard to reuse scenarios. This project focused on streamlining setup, controlling load predictably, and making results easy to interpret and reuse across devices and labs.

Main Job

Validate device throughput and traffic-classification behavior under representative, controllable load conditions.

Top 3 Desired Outcomes

  • Minimize the time it takes to assemble a representative application-traffic mix for a new test.
  • Minimize the likelihood that generated traffic or classification deviates from the intended mix under load.
  • Minimize the effort it takes to recreate and reuse prior test conditions across devices or labs.

Impact

  • 🧩 Mix builders and presets reduce setup friction for realistic scenarios.
  • 📈 Controlled load ramps reveal performance thresholds and policy behavior clearly.
  • 📤 Consistent, exportable results support comparison and downstream analysis.

Design Concepts

Early exploration focused on simplifying a traditionally script-heavy workflow into a clear, guided interface. Wireframes defined how engineers would build traffic mixes, ramp rates, monitor key KPIs (throughput, latency, classification accuracy), and export reproducible results.

We iterated through multiple layouts to balance configurability with clarity, ensuring the test flow felt predictable and repeatable—even under complex device scenarios.

Wireframes and prototype
Early layout, wireframes/prototype explorations.

Design Concepts

Final UI: traffic mix editor, ramp controls, and real-time performance panels was created by collaborating with the UX team's visual designer. My role was to art direct as the wireframes were interpreted using our design system. New components were added based on this new faeture.

Design sketches
Final throughput test views.