CyberKshetra Admin Scenario Builder

CyberKshetra Admin Scenario Builder

Designing a set up phase where creators can create topology by drag and drop Nodes and Configure this.

INTRODUCTION

01

Project Overview

Project Overview

THE PROJECT

Created a clear workspace for configuring live cybersecurity on OpenStack.

My Role

Product Designer (Phases 1 & 3), collaborating with 2 BAs, 4-5 network engineers, and 4 developers.

STATUS AND TIMELINE

~9 months from discovery to shipping.

THE IMPACT

Launched a config system that removed ‘error hunt’ for Scenario Builders, using one UI for six node types.

PROBLEM

02

The Problem Space & Architectural Tension

The Problem Space & Architectural Tension

A scenario isn’t just content; it is live infrastructure standing up on OpenStack.

A scenario isn’t just content; it is live infrastructure standing up on OpenStack.

The Complexity

The Complexity

Node types need specific inputs. A Router needs a name, IP, MAC, and interfaces; a Network needs a gateway and DHCP.

The Stakes

The Stakes

Get one input wrong, and the entire cyber-range environment fails to stand up.

The Tension

The Tension

Should we create six separate forms, causing user relearning, or one rigid template that doesn’t fit well?

These constraints meant the design couldn’t simplify the inputs — only the experience of navigating them.

These constraints meant the design couldn’t simplify the inputs — only the experience of navigating them.

CONSTRAIN

03

Constraints & Discovery

Constraints & Discovery

The Fixed Complexity

The Fixed Complexity

Cybersecurity analysts set fields and validation rules based on cyber-range platforms. Complexity could not be changed.

The Design Lever

The Design Lever

Since I couldn’t remove the complexity, the only lever left was the experience—how inputs are grouped, how nodes maintain familiarity, and how mistakes surface.

Ownership Boundary

Ownership Boundary

Analysts and BAs defined what the system must do; I owned how a human actually does it.

TRADE OFFS

04

Trade-offs & Explorations

Trade-offs & Explorations

The ‘Where vs. What’ Trap

The ‘Where vs. What’ Trap

The validation model turned bad nodes red. Builders saw red and knew the error, but didn’t know what was wrong. They had to open the node and hunt for the mistake.

Discarding Inline Annotations

Discarding Inline Annotations

I actively chose not to rely on standard inline error text. On a dense topology canvas, scattered errors become visual noise. A builder doesn’t need to hunt through a map; they need a checklist.

THE SOLUTION

04

The Final Solution

The Final Solution

SCREEN 01

The Workspace

The Workspace

A clean canvas where builders can drag nodes, connect topologies, and trigger system-wide validation when ready.

SCREEN 02

One Pattern, Many Nodes

One Pattern, Many Nodes

I designed a single structural skeleton applied across Networks, Routers, Servers, and End devices. Only the specific input fields swap out. The builder learns the layout logic once and applies it everywhere.

SCREEN 03

Aggregation over Annotation

Aggregation over Annotation

I replaced the ‘red node’ hunt with a single, grouped sidebar list of every error and its specific reason (e.g., Router 1 — IP address error; Network — missing gateway). Clicking the error navigates the user directly to the broken field.

SCREEN 04

Constraints at Entry

Constraints at Entry

Implemented required-field cues, character limits, and format hints to catch edge cases before the user hits validate.

SCREEN 05

The Final Polish (Phase 3)

The Final Polish (Phase 3)

A dedicated step to review the topology, upload documentation, attach questionnaires, and finalize the scenario.

THE SOLUTION

04

Business Impact & Retrospective

Business Impact & Retrospective

Powered a Live, Top-Tier Event

Powered a Live, Top-Tier Event

This architecture was tested in a real deployment. It was reused in the CTF module for a live competition at IIT Delhi, supporting 150 students.

User-Driven Architecture

User-Driven Architecture

The system was shaped by testing. The error panel emerged from observing a Scenario Builder struggle with the validation model.

-64%

-64%

reduction in scenario configuration time

-90%

-90%

decrease in failed OpenStack environment spin-ups

What I’d Do Differently

What I’d Do Differently

Validation was done with expert Scenario Builders. I need to test this flow with newcomers to see if the ‘learn one node, know them all’ design intent holds for first-time users.