Unikernels: The Next Evolution in Cloud Infrastructure
summary
Unikernels fuse application and OS libraries into a single sealed image, claiming drastically smaller attack surface and millisecond boot times; MirageOS is shown as the toolchain that compiles only the needed libraries into images that boot in ~30 ms and target Xen or other runtimes. A live demo pattern is Jitsu, which performs just-in-time summoning of unikernels on request. Tradeoffs include language/tooling limits, per-image memory reservation, and immature ecosystem despite clear gains for microservices and on‑demand scaling.
AI-written summary · about
chapters (27)
- 0:00Intro
- 0:15Unikernels: The Definition!
- 1:03The Picture Today
- 1:48Software is Complex
- 2:53Complexity is Unnecessary
- 4:45Lack of Portability
- 5:45Pre-requisite for Solution
- 6:55Solution: Containers!
- 9:12Introducing Unikernels
- 10:48Unikernel Application Stack
- 12:30Evolution
- 13:33Characteristics of a Unikernel
- 17:17Embedded Systems for the Datacenter
- 19:17Use Case Scenarios
- 20:03The Development Stack
- 20:36The Testing Stack
- 20:55The Deployment Stack
- 22:08Benefit: Cost Savings
- 22:49Benefit: Faster Boot Times
- 23:07Benefit: Better Security
- 23:40Why NOT to use Unikernels?
- 25:12Unikernel Implementations
- 26:38Xen Virtualization Server
- 29:21Example: Mirage OS
- 30:12On Demand Allocation: Jitsu
- 31:15Resources
- 31:38Conclusion
related talks
44mUnikernels and Hyper Elastic Clouds
34mLook ma, no OS! Unikernels and their applications
17mUnikernel based deployments
33mUnikernels and Explorations
38mUnik & Cloud Foundry Integration: Focus on IoT
35mUnikraft: Unikernels Made Easy
18mUnikernels for Serverless Computing
24mUsing Unikernels to deploy real-world apps