We have (unfortunately) entered the era of the idle cloud, characterized by rampant over-provisioning being used as a shoddy and expensive workaround for the severe shortcomings arising from having built the cloud on legacy technology.

Such technology was designed for seconds or minutes semantics in a world of millisecond cloud traffic. We believe cloud infra is ripe for disruption, and have created a revolutionary, from-first-principles, millisecond cloud platform.

With Unikraft Cloud, millisecond cold boots, scale-to-zero and autoscale, as well as high server density (e.g., thousands of instances on a single server) are the norm – all while ensuring strong, hardware-level isolation across instances and a familiar Docker-based user interface. A millisecond platform for a millisecond cloud traffic world.

To get here, seven years ago Unikraft’s founding members embarked on a research journey to fundamentally revisit how the cloud is built, taking a first principles approach: given that hardware-level isolation is key, what’s the thinnest, most efficient and secure software layer that an application requires to run?

The research lead to the creation, 5 years ago, of the Linux Foundation Unikraft project, with the explicit aim to create a truly cloud-native OS and cloud stack that would remove all of the inefficiencies of current deployments (cold boots, slow autoscale, bloated images, large memory consumption, exploding cloud bills, etc.).

It is time to take a leap forward and enter a new era of true serverless, where users simply deploy without headaches, efficiently and without over-provisioning.