Case Study Enterprise UX Project
VMware by Broadcom

"Designing for Redundancy & Reliability"
Dual DPU Support (vSphere 8.0.U3)

Dual DPU Support

📄 Project Overview

👤 Role: UX Designer (collaboration with UX design manager)
🗓️ Timeline: 2022 - 2024
🤝 Cross-functional partners collaboration: Product Managers, UI Engineers, Hardware Partners (NVIDIA, AMD), Architects and Backend Engineers (around 10 separate engineering teams)
🔵 Product: VMware vSphere 8.0.U3

🛠️ Tools: Figma, Jira, Confluence
📚 Design Patterns / Libraries: Clarity Design System, internal VMware enterprise UI patterns for consistency

Summary

I led the UX design for dual data-processing units (DPUs), creating monitoring and configuration flows that improved system reliability, covered 10+ edge cases, and streamlined collaboration between admins and infrastructure teams.

1. Design Problem

With the introduction of DPUs (Data Processing Units) into vSphere, customers gained new performance and security benefits. However, users needed the flexibility to run multiple DPUs (dual-DPU setups) on a single host, which introduced new complexity in terms of setup, lifecycle management, and monitoring.

Business Goals:
- Improve customer satisfaction by enabling reliable, redundant hardware usage.
- Ensure no idle hardware (customers didn’t want standby DPUs sitting unused).
- Support all server and smartNIC vendors in the product.
Technical Requirements:
- Dual DPUs must operate in Active/Active mode (both handling production traffic).
- Support failover if one DPU/NIC fails (e.g. software issue or cable disconnect).
- Limit scope to two DPUs only

UX Goals (Use Cases):

Installer – Install ESXio on both DPUs per host.
Lifecycle Manager – Keep DPU firmware/software up to date.
Health & Sensors – View metrics to confirm DPU health.
DPU Presence – Know if a host has a DPU and view its status.
DPU Monitoring – Performance insight for capacity planning.
Networking – Configure DPUs in Active/Standby mode.

Key challenges / user needs:

Configuration complexity: Admins needed a clear way to configure and assign workloads across two DPUs without breaking workflows.
Lifecycle parity: Ensure that updating and managing dual DPUs followed the same lifecycle as other hardware components (consistency with vLCM).
Monitoring clarity: Users needed clear feedback about which workloads were running on which DPU and whether performance/security policies were applied correctly.
Edge cases: What happens if one DPU fails, is mismatched, or is downgraded in firmware? Users must understand system health quickly. For this project, there were quite a lot of edge cases, especially in the Networking use cases (10+ edge cases)

2. UX Solution + Mockups / Walkthroughs

As UX Design Lead, I collaborated closely with UI engineers, product managers, solution architects, and backend engineers to simplify these complex technical requirements into user-centered workflows and consistent UI patterns.

Installation

Redesigned the installation screens to allow the users to install both DPUs, removed some user overhead by reducing the amount of installer screens and simplifying installation settings. This use case also had to cover a few error state as additional edge use cases.

Installation use case mockups (gallery, 13 screens)

Interactive Figma prototype walkthrough

Monitoring

Added an element to current dashboard screens showing status, metrics, and alerts. The added function made it easy to see if both DPUs were healthy, misconfigured, or failing.

Monitoring use case mockups (gallery, 2 screens)

Interactive Figma prototype walkthrough

Networking

Redesigned the networking setup screens to guide admins through key steps, reducing chances of mistakes in a high-stakes environment.
Designed workflows for creating and managing virtual distributed switches (vDS).
Added options to assign uplink modes and manage advanced settings like Link Aggregation Groups (LAGs) for bandwidth and redundancy.
Simplified the “Manage Physical Adapters” step to reduce admin confusion.

New Switch -> Assign Uplink Modes

Networking Mockups Part 1 (gallery, 14 screens)

Interactive Figma prototype walkthrough

Add / Manage Host -> Assign Uplink to vmnics

Networking Mockups Part 2 (gallery, 13 screens)

Interactive Figma prototype walkthrough

Distributed Port Groups - Advanced Settings

Networking Mockups Part 3 (gallery, 8 screens)

Interactive Figma prototype walkthrough

DPU Presence & Overview

Delivered a new Dual DPU overview card design after 5–6 rapid iterations with the engineering teams
Delivered 10+ edge cases to ensure reliability and clarity in all scenarios, including:
- Single DPU active.
- Dual DPUs (1 active, 1 standby).
- Connectivity issues and failovers.
- Misconfigured or failed DPUs.
- Both DPUs active across different distributed switches.

Dual DPU Presence & Overview Card Mockups (gallery, 24 screens)

Interactive Figma prototype walkthrough

Health and Sensors

Lifecycle flows were kept the same as the current single DPU screens. Only provided Lifecycle screens with the new Dual DPU view.
Redesigned the Dual DPU grid view for better visibility for health indicators, alerts, and metrics.

Health and Sensors + Lifecycle Mockups (gallery, 2 screens)

Interactive Figma prototype walkthrough

Video walkthrough of all the project mockups coming very soon! Stay tuned!

Collaboration & Process

Led weekly cross-functional reviews for a many months in the period 2023 - 2024
Balanced technical constraints with usability for enterprise admins.
Ensured all designs followed WCAG accessibility guidelines, supporting inclusive use across VMware’s enterprise tools.
20+ teams and stakeholders were involved in this project which highlights its collaborative complexity
UX team had to lead many product discussions with the partnering teams to navigate towards the most usable solution for our users

3. Outcome & Results

Shipped in vSphere 8.0.U3 / VCF 9.0 as the first release with full Dual DPU support.
Delivered a more intuitive experience for VI admins:
- Reduced configuration errors by simplifying vDS and DPU setup.
- Cut troubleshooting time by covering 10+ edge cases in the UI.
- Helped users use 100% of their hardware investment (instead of leaving backup units idle).
- Improved hardware utilization by supporting Active/Active DPUs, preventing wasted resources.
Recognized internally for bridging engineering and UX, ensuring complex backend logic was translated into clear, usable workflow for enterprise customers.