Using Tonic’s data de-identification tools, eBay shortens development cycles and fuels its automated testing suite. Starting with eBay’s vast data ecosystem of multi PBs spread over multiple data sources, Tonic enables eBay developers to scale their data down to manageable subsets, rapidly protect those subsets with privacy guarantees, and easily call up the exact variety of data they need to hydrate staging, replicate their most complex buyer journeys, and quickly fix bugs found in production.
Using Tonic’s data de-identification tools, eBay shortens development cycles and fuels its automated testing suite. Starting with eBay’s vast data ecosystem of multi PBs spread over multiple data sources, Tonic enables eBay developers to scale their data down to manageable subsets, rapidly protect those subsets with privacy guarantees, and easily call up the exact variety of data they need to hydrate staging, replicate their most complex buyer journeys, and quickly fix bugs found in production.
“Nothing that we tried in-house is comparable to what we’re doing now with Tonic. It’s a game changer both in terms of the automation we can achieve, as well as on-demand function validation targeting specific use cases with the precise data we need.”
- Srikanth Rentachintala, Director of Buyer Experience Engineering at eBay
The challenge of maintaining staging
In 2019, a fundamental challenge was brought up by the team at eBay: staging was unreliable at best, and a lack of quality data was a large source of the pain. Without quality staging, there was no standard pattern across teams for performing regression testing, and without quality data, their automated testing suite entailed costly overhead to run and frequently failed. “It had become a chicken and egg problem,” states Senthil Padmanabhan, Technical Fellow, VP of UX Engineering and Developer Productivity at eBay. “Since the data wasn’t there, not a lot of effort was dedicated to making sure staging was always available. Productivity was impacted accordingly. It was a constant point of frustration for our developers.”
Enter Tonic. With its advanced subsetting capabilities, Tonic rose to the challenge in meeting eBay’s needs.
Secure, quality data on demand
Given the scale and complexity of the eBay ecosystem, the company’s engineers are deploying Tonic in phases to segment the effort and prioritize their most critical use cases. They’ve hand-selected ten domains for which Tonic is creating clusters of roughly 1 GB subsets. These subsets preserve the source databases’ referential integrity, capturing the breadth and variety of key patterns within the domains, while scaling the data down to a manageable size. The subsets are then de-identified to protect eBay’s users while reducing the risk of reverse engineering.
“This year is all about subsetting the data in key domains, protecting it, and making the de-identified data accessible to our developers,” states Padmanabhan. “We’re in the phase of building the infrastructure, building the fundamentals. As our developers have started to take ownership of the project, we’re really seeing the momentum pick up.”
With the data for these critical use cases now readily available, developers are able to jump directly into the function they need to test and perform their validations without having to work through the time-consuming steps of crafting a trail of data specific to their use case. “You can see the joy in the eyes of our developers using Tonic,” states Srikanth Rentachintala, Director of Buyer Experience Engineering at eBay. “They are finally able to get to the data of their choice without going through a lot of hassle. And some of our use cases are really difficult to run. Now we can get to them very easily—it’s a major change from what we were able to do previously.”
With the time savings, teams are looking to increase their release velocity and expand feature development. “The goal is to ship fast and ship more, with higher confidence,” says Padmanabhan. And they’re already thinking ahead to expand their use of Tonic. Current efforts have focused on subsetting Oracle databases but the next wave will see Tonic pointed at noSQL databases and other more distinctive use cases.
“Often with vendors, there’s push back on our asks for features,” notes Rentachintala. “But with Tonic there’s never been a moment where they said, ‘No, we can’t do this.’ It’s always, ‘Let’s figure out how we can do it.’”
“Tonic is a very engineering-focused organization, and I really admire that about them,” adds Padmanabhan. “They take privacy very seriously. A lot of companies would love to have a solution like this to enable securely moving a subset of production to a development environment. What prevents them is privacy concerns and all the complexities involved. With Tonic, we found the right balance, and privacy is taken care of.”
Tonic.ai’s impact and results
- 8-PB data ecosystem scaled down to 1-GB, referentially-intact subsets targeted to specific domains
- Significant time savings in their automated testing suite, with an increased pass percentage of automation scripts in staging
- Equipping eBay’s developers with quality, privacy-preserving data for their most critical use cases
.svg){kind=icon}
