top of page

Amazon Relational Database Service (RDS) Custom SQL Server CEV support

Custom SQL Server CEV support cover phot
Role: UX Designer and UX Researcher
Company: Amazon Web Services
Duration: April 2022-January 2023


Amazon Relational Database Service (Amazon RDS) is a collection of fully managed services that makes it simple to set up, operate, and scale databases in the cloud. RDS Custom gives the flexibility having both an automatic managed environment same as RDS and also gives flexibility getting responsibilities by customer. Customers can manage their media with RDS Custom. RDS custom is available for Oracle and SQL Server. In RDS Custom for Oracle, customers can create Custom Engine Version (CEV) which allows them to bring their own media and third party applications. With this project, I was asked to redesign the RDS Custom experience to add CEV support for customers of Microsoft SQL Server as well. I tasked with owning the user experience as a whole. 

My individual contribution

I was the only UX designer on an Agile team comprised of 2 product managers, 5 back end teams, 2 front end engineers, 1 doc writer, and one doc editor.
I was responsible for determining the overall design direction of the project, while collaborating with the rest of the team on problem discovery and ideation. I worked on redesigning the whole experience of creating CEV and using this CEV to create an RDS Custom database instance for SQL Server. Worked closely with front end engineers to add key metrics to the console during implementation to be able to get data driven results after launch. 
I also worked closely with the AWS design system team to make sure our product followed the UI patterns and guidelines.


  1. Leading the whole project by myself while I did not have any knowledge about the technical part of the feature.
  2. How might we redesign to add the availability of CEV for SQL Server to the current state of the feature in a consistent way?
  3. RDS Custom SQL Server’s current feature does not match customers’ need. They are blocked by the restrictions of RDS Custom SQL Server with pre-built engine versions


  1. Research with customers and stakeholders to better understand the feature
  2. Redesign the CEV experience of RDS Console​
  3. Follow AWS design system UI patterns and guidelines
Situation & Task

Approach to take Action

Double diamond design thinking approach:

Of course It was not a linear path for me!
I went back and forth between these stages in order to fully understand what the problem is and how I can solve it, or improve it on an existing solution. 


The actions that I took in the discovery and research stage of the approach was to better understand the technical side of this feature and also to analyze the problem better in order to come up with the best solution. The steps were as follow:
  • I explored the research that were conducted previously on this topic and also had a deep dive reading into the PRFAQs (Press Release (PR) and Frequently Asked Questions (FAQ)) that published for RDS Custom for Oracle. 
  • I explored AWS documentation (which is a good resource to understand and learn about released features).
  • I interviewed 10 stakeholders including solution architects, product managers, internal and external customers.
    1. I started my interview with asking questions: I wanted to understand who are my customers and what do they need from technical side of the issue. 
    2. Next, I conducted UI audit:  I wanted to understand how they use the console for RDS Custom SQL Server pre-redesign and to analyze the problems.
Following is the key findings of the research:

Key findings

What is RDS Custom?
Current use cases of RDS Custom:
Insights gained from UI audit of pre-redesign of RDS Console pages with stakeholders: 
Amazon RDS console, CEV page pre-redesign: 
Frame 204.png
Amazon RDS console, CEV creation page:
Frame 205.png
Amazon RDS console, RDS Custom for SQL Server database creation page:
Frame 207.png
Amazon RDS console, RDS Custom for Oracle database creation page:
Frame 206.png


Who is our customer? 
After discovering the problem from both technical side and the UI that need to be change, I had a clear understanding of who is my customer and what they need. So I created a persona and shared it with the team to all be in the same page about what business and customers challenges we are going to solve: 
Customer journey:
As the next step, I used the user research findings to map user journey and understand the path a customer may take to reach their goal in the current state of the feature pre-redesign and without CEV. It helped me to identify customers needs and opportunities to improve their experience. 
Based on information I got from research and after defining my customer and their journey map. I wanted to have a clear understanding of the problems that I am going to solve, so I used 'How Might We' questions to be able to generate creative solutions while keeping the team focused on the right problems to solve. 
How Might We: 
  • Make the experience of creating CEV for RDS Custom for Oracle for SQL Server in a consistent way to RDS Custom for Oracle.
  • Inform and learn customers about CEV creation of RDS Custom for SQL Server.
  • Make availability of CEV to the current console’s state in a sustainable way.


Ideation for redesign of CEV page:
  • I started the ideation with CEV onboarding page, as it is the first page that customers navigate through to get information about CEV creation steps and then as the next step they will navigate to CEV creation page.
  • The redesign of this page must be in a sustainable and scalable way in order to make it possible to add more engines in the future.
  • As we already had the onboarding panel for RDS Custom for Oracle, It was challenging for me to redesign the panel in a way to consist of SQL Server engine as well and at the same time gives informational steps for each engine version.
  • I spent a fair amount of time to ideate and brainstorming on CEV onboarding page. This activity would have me to engage with engineering and product in order to solve problem while gaining context of constraints and possible barriers for development. I scheduled weekly team workshops to design each design idea with the team and get their ideas.
  • During ideation I followed AWS design system UI pattern and guidelines.
Following is the 4 design ideas that I finalized with the team and next I conducted Dot voting method to prioritize on one of the design ideas:
Design idea 1: 
Design idea 2: 
Design idea 3: 
Design idea 4: 
Dot voting on 4 design ideas for CEV page:
​Next, I conducted Dot voting method to prioritize on one of the design ideas:


Usability Testing

Testing the early design work with internal customers:
Based on the workshops and brainstorming session that I had with the team, I created the flow which consists of redesigning the pages that I audit at the beginning of the project and I tested it with internal customers.  
  • Number of participants: 3 customers
  • Interview time: 60 min
  • Role: Database specialist solution architects
  • RDS Custom usage: Customer who are familiar with RDS Custom and have shared actionable feedbacks during the research phase of the project.
Following is the flow of early design work that I tested with customers:

High fidelity design and Implementation

Based on insights that I got during usability testing, I iterated on the design and created the high fidelity wireframes. 
CEV onboarding page:
Create database with CEV:
Create database with CEV:

Implemented prototype



I collaborated with the console engineer to add tracking metrics during the implementation, I worked backward and created a funnel with showing metrics that we need to implement. It helped to better communicate with the engineer and for both of us to have a clear understanding of what and why we need these metrics. I synced with the product manager and we decided to monitor these metrics 1 month after releasing the feature. Following is the funnel tracking metrics illustration:
I also decided to track the same metrics that were already implemented for RDS Custom Oracle and compare the data with the same metrics implemented for RDS Custom for SQL Server. Following are the insights that I got:
  • 87% of customers successfully created CEV for SQL Server
  • Only 12% of customers got failed status of creating CEV and it was due to MBS validation process
  • RDS Custom SQL Server DB instance creation is increased 2.1% by internal customers and 1.4% by external customer. Currently 93% of internal customers successfully create RDS Custom SQL Server with CEV which it is 85% of internal customers for RDS Custom for Oracle. 
bottom of page