Dell Technologies · Enterprise SaaS · 2022–2023
APEX Navigator for
Multicloud Storage
Building a new product from scratch — and making complex cloud infrastructure simple enough to trust.
My role
Product Designer
Timeline
1 years · 2022
Platform
Dell APEX / Premier
Launched
Nov 30, 2023 · Shown at Dell Technologies World and AWS re:Invent
Overview
This was the first management experience in Dell's APEX Navigator line. There was no existing solution, no prior designs, no internal reference to copy. We built it from scratch. My job was to take all that complexity and give it a clear shape: simple steps, a visual map of the system, and one shared prototype that kept the whole team on the same page from start to finish. I joined in 2022 and worked on it just under a year. Over time I became the go-to designer for APEX patterns — the person other designers and PMs came to when they needed to know how the system should work. The heart of the work was one living prototype that kept a dozen teams aligned — from a blank page to a keynote stage. It worked, though it taught me as much about the limits of that approach as the power of it.
The challenge
Building a new product from scratch — for a system with no map yet.
The challenge was building one from zero, for genuinely complex infrastructure — and giving it a clear shape before anyone knew what "clear" should look like.
No existing product or UX patterns to build from.
The core flows — deploy, manage, monitor — all had to be defined from scratch.
The system could go four or more levels deep.
Different people pictured the system differently, so there was no shared starting point.
The job was to turn all of that into something a user could move through with confidence.
Mid-project pivot
The platform changed near the finish line. The team thought it couldn't be done in time.
In early 2023, with the design almost finished, the business direction changed. The product would now ship through Premier — Dell's B2B online store — instead of the APEX Console we'd designed for.
The team worried it couldn't be done in time. I was confident it could.
Not because I work fast. Because the prototype was built on a solid, reusable component structure — well-organized and used consistently. When the platform changed, the flows and logic still held. Most of what had to change was just the look and layout.
I updated the designs, hit the engineering deadlines, and built a separate version of the prototype for the Dell Tech World keynote — working with marketing and an outside production company to get it stage-ready.
The lesson: good component structure isn't only about saving time day to day. It's what keeps you standing when things suddenly change.
The lesson: good component structure isn't only about saving time day to day. It's what keeps you standing when things suddenly change.
Key Design Decisions
Step-by-Step Deployment Flow
Some decisions don't belong together. Basic setup is a different kind of thinking than performance setup. Putting it all on one page would force people to face choices they weren't ready for — and in infrastructure work, a wrong setting has real consequences. So I grouped each step by decision type: what you're deploying → where it connects → how it's set up. Each step is small enough to finish and move on.
Topology visualization over text-based status tables
The system is truly layered — four or more levels deep. A table would force users to rebuild those connections in their head. For IT users fixing a live system, that's too much. The map shows two things at once: how the parts connect, and whether each one is healthy. A table can't do both. The map isn't decoration — it's the only way to see what you're looking at and what might be wrong at the same time.
Progressive instructions for Cloud Access credential entry
One step needs users to set up AWS credentials by hand — copying values between two systems. It can't be automated; that's an AWS rule. My first instinct was to show all the instructions by default. But early feedback said that felt overwhelming, even for technical users. So I redesigned it: instructions hidden by default, with a small "Where do I find this?" link at each field. The goal was to make the complexity manageable — without pretending it isn't there.
Configuration grouped by decision type
Instead of ordering settings the way the backend processes them, I grouped them the way storage admins actually think: Basics (what it is and where it lives) · Performance (capacity and speed) · Availability (backup) · Networking (connection). This lets people jump straight to the part that matches their job.
Prototype as the team's source of truth
As more teams joined, static mockups weren't enough — people read them differently because they couldn't feel the real flow. Keeping one fully connected prototype always up to date was a lot of work and hard to scale. But it became the "bible": the one reference for leadership, marketing, PMs, researchers, and engineers. It cut confusion in meetings, and it's what made the late pivot survivable.
Final design
An end-to-end experience across four areas: Deployment, Management, Data Mobility, and Cloud Access.
Deployment — Simplifying complex setup
Each step shows only what's needed, right when it's needed, so users can deploy storage on AWS with confidence.
Step 1
Product Selection — Guided Start
Visual cards for Block vs. File storage make the first choice easy. Only the key decisions show (type, provider, region), keeping the screen light.
Step 2
Cloud Access — Secure Connection
Hard technical inputs (Role ARN, trust policies) are made approachable with step-by-step instructions, collapsible help, and inline "Where do I find this?" tips.
Step 3
Deployment Configuration — Resource Setup
Settings grouped into Basics · Performance · Availability · Networking. Inline help and sensible defaults throughout. As users enter values, the interface calculates the minimum required AWS infrastructure — setting expectations before provisioning begins.
Management — one place to see and control everything
Cloud Deployment List — Monitoring across environments
A color-coded health table for a fast read across AWS and Azure, with capacity and group details right there — no drill-down needed for routine checks.
Capacity metrics and group details inline — no drill-downs required for routine checks
System-level and cloud-access views for different user roles
Deployment Details — End-to-end configuration insight
Split into setup (storage, region, speed) and live health. Links out to PowerFlex Manager and CloudIQ for deeper analytics, keeping things connected across Dell tools.
Layout split into system configuration (storage, region, IOPS) and real-time health monitoring
External navigation to PowerFlex Manager and CloudIQ for deeper analytics — maintaining continuity across Dell platforms
Mobility section for cloning and data movement without leaving context
Cloud Access Management — Credential oversight
Tabs for AWS · Azure · GCP, ready for more providers later, with account groups in a light, searchable table.
Tabbed layout segments cloud providers (AWS · Azure · GCP) — scalable for future integrations
Account groups and permission details in a lightweight, searchable table
Data Mobility — Data movement simplified
Mobility — Centralized view of data movement
Source → target arrows make the direction easy to read. A clear "Create mobility group" button shows the next step.
Tabbed navigation for Volume · Application · File mobility groups
Source → target flow arrows make migration direction immediately readable
"Create mobility group" CTA promotes a clear next step without requiring menu navigation
Group Overview — Visualizing data flow across environments
A visual flow — source (on-site) → target (cloud) → copies — with transfer method and progress shown upfront.
Visual flow diagram: source (on-premises) → target (cloud) → clone sets
Transfer method and progress details surfaced upfront for context without drill-down
Clone Sets & Clone Details — Managing replication
A sortable table (name, time, status) with an inline "Create clone" — no switching screens.
Sortable table: clone name · image time · mapping status
Inline "Create clone" — no context switching required
Editable mappings in detail view simplify volume association updates
Impact
The biggest impact wasn't in the product — it was in how the team worked.
Launched November 30, 2023, and available in the US that December. Announced at Dell Technologies World 2023, and shown at AWS re:Invent in Las Vegas that November.
Defined the core UX patterns for a brand-new multicloud product.
Made complex setup understandable by breaking it into clear, guided steps.
Made system structure readable at a glance through the visual map.
Kept product, engineering, and leadership aligned with one shared prototype.
Cut repeated clarification back-and-forth by making the prototype the single reference.
Left behind reusable patterns that shaped later multicloud work.
What teammates said
"Many, many teams have come to rely on the Cirrus prototype as the source of truth for what the Cirrus experience will be. He's the Figma master, and the UX Cirrus team wouldn't have been able to deliver as swiftly or as well without his incredible skills in design and management."
Maggie Harney · Cirrus prototype recognition
"You consistently delivered quality work, on time, and never let ambiguous direction or requirements stop you from designing an excellent user experience for Multicloud customers."
Kathryn Ward (Manager) · Game Changer 2 Award — APEX Navigator for Multicloud Storage
"Min worked meticulously and tirelessly on the Figma prototypes that were absolutely essential for the DTW Day 2 APEX keynote demo. Without Min's help we would not have gotten the work done in the extremely short timespan that was required."
David O'Dell · Dell Tech World keynote recognition
Key Insight
The prototype worked. Every team relied on it. People called it "the bible," and directors asked if I could do the same for their projects.
I said no.
Not because I couldn't — because it wasn't sustainable. Managing a prototype that big was intense. Hundreds of wireframes and components, all connected. I spent most of my days just making sure everything still worked, and problems kept popping up anyway. It held the team together, but it ran on my hours, and that doesn't scale.
That's the real lesson. Making a clickable prototype feel completely real can cost more than it's worth. It aligned everyone, which was valuable — but the same result shouldn't depend on one person maintaining a giant machine by hand. Next time, I'd find where "real enough" ends and "polishing forever" begins, and stop there.