Intel DAOS is an open-source software ecosystem revolutionizing high-performance storage for data-centric computing.
DAOS is a complete I/O architecture that aggregates SCM and NVMe storage into globally-accessible object address spaces, providing consistency, availability and resiliency guarantees without compromising performance. DAOS was created to meet the needs of enterprises producing an abundance of data that can be stored and made instantly available when needed.
Designed for Non Volatile Memory (NVM), DAOS offers companies elastic storage options to optimize business performance and cost.
The Distributed Asynchronous Object Storage (DAOS), built by Intel, came to Launchable through our Discord community looking to use AI to optimize their Jenkins tests and find bugs or tests that run for a long time.
Starting in early 2023, Launchable’s AI-driven Predictive Test Selection (a form of Test Impact Analysis) was able to see a significant decrease in monthly test run times.
They used zero-input subsetting, which assumes they ran all their tests every time. However, in reality, they never run all their tests — they run different groups of tests at different times. Their pre-Launchable runs ranged from a couple of seconds up to ten hours.
With Launchable’s AI Test Intelligence Platform, DAOS accelerated its development cycle by only running the tests relevant to the changed source code, reducing its test run times to an average of 116 minutes and saving over 2,000 hours a month.
Now, DAOS can use AI for their DevOps pipeline to find failures fast, iterate faster and release software with high confidence so they can continue to help other companies continue to store valuable AI-produced data.