SemperOS: A Distributed Capability System
summary
SemperOS implements a multi‑kernel, HW/SW co‑designed distributed capability system that handles cross‑partition delegation and revoke races by adding a two‑way handshake for delegation and a distributed mark‑and‑sweep for revocation. It builds on M3‑style DTUs and microkernels, runs services on dedicated cores, and hits up to 348k capability ops/s in simulation and 72–78% parallel efficiency when replaying 512 application instances while using ~11% of cores for OS services.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:23Motivation
- 2:27Characteristics of Capability Systems
- 4:11Base Capability System - MS
- 5:05SEMPEROS- A Distributed Capability System
- 5:35The Distributed Data Lookup - DDL
- 6:37Distributed Capabilities - Revocation
- 10:29Distributed Capabilities - Interferences
- 10:52Evaluation
- 11:24Using the Capability System - Services
- 14:11Application Benchmarks - NGINX webserver
related talks
55mProtected Execution Facility On Power
01h 03mMINIX 3
21mEmbedded solution for multiprocess application
01h 06mIt's Time for Operating Systems to Rediscover Hardware
21mHarmonizing Performance and Isolation in Microkernels with Efficient Intra-kernel
15mNrOS: Effective Replication and Sharing in an Operating System
56mMeet All Scenarios OS: A Distributed O.S. with Feet on the Ground
48mDemystifying the Secure Enclave Processor