As cloud services move from being startups to enterprise applications, they inevitably must support more and more bespoke use-cases unique to their increasingly upmarket clientele. Given how mission-critical enterprise applications are for large customers, reliability is important. Losing ground on either enterprise functionality or reliability inevitably risks disruption by new startups.
In modern software development, various testing approaches are used to ensure applications remain reliable as new features are added. Many small, cheap “unit tests” are used for sanity checks, but don’t offer complete assurance on reliability.
By contrast, “end-to-end” (E2E) testing offers rigorous checks on whether the software continues to function as advertised, but these tests are often brittle and slow. As cloud software teams deploy software ever more frequently, the time taken for tests to run can become a bottleneck in the software development lifecycle.
Given the ever more important role E2E testing plays in enterprise applications, some teams have been working on making these tests more maintainable and faster for the cloud era.
Here are two examples. In 2020, the startup Walrus.ai raised seed funding for a possible solution to maintainability by adding a human-in-the-loop. When a brittle test failed, Walrus.ai would refer the test to a human for manual execution. This concept was ultimately unpalatable for the developer market and, at the start of 2022, Airtable tweeted that it had acquired the team behind the Walrus.ai tool.
One of its competitors, LambdaTest, has instead focused on improving the technology that runs these E2E tests and has seen more success, having raised $16m in series B funding in 2021. LambdaTest not only offers dashboards to identify unreliable flaky tests and monitor performance, but is also making progress in improving the underlying technology behind these testing frameworks.
LambdaTest’s focus on cutting latency in testing was recently shown in its new HyperTest offering, which is claimed to be “up to 70% faster than other traditional cloud test execution platforms”.
This experience shows us just how important performance is becoming for developer tools. As technology plays a leading role in commercial success, technology teams need developer tools that allow for short, quick deployment cycles. For those involved in building developer tools, especially those focused on an enterprise audience, speed is becoming ever more critical.
Junade Ali is an experienced technologist with an interest in software engineering management, computer security research and distributed systems