ux case study

Demystify Prisma SD-WAN Site Issues

Help Network Operation Engineers monitor and troubleshoot site problems efficiently with a simple workflow and metrics that matter.

A screenshot of web page
My Role

Lead and solo designer:
User research
Sketch and wireframe
Interaction and prototype
Visual design
Design QA

The Team

1 product manager
6 engs, 1 QA

Status

Research and design: 1 month

Created in 2021

Released in Sep 2021

IMPACT

The new design was proved to be a huge milestone and enhancement of the product


Would like to skip the design process? Check out the final design and recap directly.

View final design and recap
Context

What is Prisma SD-WAN and site?

Prisma SD WAN, formally CloudGenix, was acquired by Palo Alto Networks in mid 2020.
SD-WAN is a software-defined approach to managing the wide-area network. It takes user and application centric approach to connect users to the applications. It is an autonomous network system, leveraging ML and data science, reducing manual configuration and troubleshooting.
There are 2 types of sites: Branch and Data Center

  • A branch is where the users and services are located
  • A data center is where the enterprise applications and services are hosted.

Each site contains:

  • The ION devices for network connection: a branch typically has 2 devices, and a data center may have 10+ devices
  • Circuits (typically 1-4) to connect with carriers: each circuit contains secure fabric links (VPN), typically 2- 100+, to connect between sites or data centers

Target user and problems

Understand the target user and their pain points of site troubleshooting

From a few quick internal domain exports and user interviews, I learned about the users and their current workflow of site troubleshooting, and summarized their pain points.

Target user- NOC Engineer

Today, when a site has a reported problem, from a specific end user or system, users do investigation on the “activity” page.

Old UI- Activity View
Old UI - Activity View

Firstly filter the “activity” page by the site. If the site name is not available, users need to manually figure out which site to investigate.

Old UI- filter and select a site on the Activity View
Old UI - filter and select a site on the Activity View

Then apply complex filters one by one to view site, device, circuit, and link level metrics

Old UI - The filters on the Activity vew
Old UI - The filters on the Activity view

After the filter is updated each time , click around the tabs to view the metrics

Old UI - The tabs and sub pages within the Activity view
Old UI - The tabs and sub pages within the Activity view

Require a lot of manual analysis and navigating to other pages in the product to take actions (eg: update site configurations and policies)

Old UI - The graphs under Activity view
Old UI - The graphs under Activity view

User pain points of site troubleshooting
Design process

How might we help users with site troubleshooting?

Brainstormed ideas around user goals

Design process of presenting site and circuit health information

Since site and circuit health metrics are the most important site level metrics user cares about, I will be focus on my design process around them in this case study.

Initial Design

Based on the user stories generated from user needs and PM requirement, I created the initial design with some hypostases.

Initial design solution based on the user stories

Positive user feedback

Then I took the design to 3 user interviews to validate the design direction and the metrics presented, I received some positive feedback that proves that the design aligns well with user needs and their expectations.

Positive feedback from user interviews

Critical feedback

At the same time, I also got some critical feedback and insights that help me to improve the design.

Critical feedback from user interviews
Outcome

Final design walk through

How users troubleshoot a site with the new design

Discover

Receive a ticket states that an end user is experiencing slow response on multiple applications.

Discover problem

Identify

Navigate to the Prisma SD-WAN UI and look up the site. The sites monitoring list gives user a comprehensive view of all sites performance with key metrics. Then user can easily search the site by user IP address from the ticket, and land of the site summary view to investigate problems.

Identify the site

Investigate

The site summary contains key metrics that users need when investigate site issues, such as: circuits, secure fabric links, and applications. Without applying any filters and clicking around, user should be able to quickly find out the problems.

Investigate site problems by viewing the key metrics

Troubleshoot

After the problem is identified, user need to take action on the problem area. The site summary view also provide links to the other frequent used pages within the product, for example: "Flows" for users to investigate on individual traffic flows of the site, or "Device configuration" for user to modify the device setup, etc.

Troubleshoot after problem is identified

Measure Success

Since the new design was released recently, there is little data to prove the success of this feature, but leveraging Pendo and user research, I would like to track and compare more quantitative data.

Time on tasks

  • Resolution time on alarms/ tickets/ calls about a site problem
  • Analysis time on the site metrics

Engagement

  • # of active users using the new features
  • # of actions initiated from the Site Summary Dashboard

Adoption

  • Frequency of using the new dashboards vs. the old “Activity” view

Recap

Challenges and learnings

Advocate design thinking and process

This is the first project for Prisma SD-WAN product, before I joined, there was no designer and engineers did the design. It was a tremendous challenge to advocate the user-centric design process, when working with the Product Manager, I always challenge and ask a lot of "why" for the features they are asking.

Deal with complexity and tight deadline

The domain was new to me and the deadline was only a month while I had a couple of other projects to work on. I scheduled sync-ups and reviews with PM and domain experts regularly to learn about the context and domain quickly and come up with ideas and solutions efficiently.

Push for user research to validate the design

I also tried to ask for the opportunity to do some user research to support the ideas and improve the design, which was never done previously. After the research sessions, I present the user insights and iterations to the PMs and Engineers, to showcase the importance of involving users early in the product development cycle to get better results and reduce engineering efforts.