The Google File System
This paper presents GFS, the scalable distributed file system implemented at Google to support data-intensive applications with fault tolerance mechanisms and high aggregate performance to a large number of clients.
Crash Consistency: FSCK and Journaling
Paper presenting the Journaling implementation in Linux file systems, along with the techniques used for allowing robust crash recovery from disasters during disk operations.
The Design and Implementation of a Log-Structured File System
Paper presenting Sprite LFS, a sytem implementing a new technique for disk storage management that speeds up both file writing and crash recovery.
RadixVM: Scalable address spaces for multithreaded applications
Paper from MIT CSAIL presenting RadixVM: a new virtual memory system design that enables fully concurrent operations on shared address spaces for multithreaded processes.
Practical, transparent operating system support for superpages
Paper presented on the 5th Symposium on Operating Systems Design and Implementation, describing a superpage management system that improves the performance of applications by increasing the TLB coverage and reducing TLB misses.
Non-scalable locks are dangerous
This paper shows that non-scalable locks can cause dramatic collapse in the performance of real workloads, even for very short critical sections, and demonstrates a replacement of the offending nonscalable spin locks with MCS scalable locks.
Eraser: A Dynamic Data Race Detector for Multithreaded Programs
This paper presents the design and implementation of Eraser, a tool for dynamically detecting racing conditions (data race) in lock-based multithreaded programs.
Serverless Computation with OpenLambda
This paper presents the advantages and main challenges of implementing OpenLambda, a platform for building applications using the serverless computation model.
The Singularity system
This paper shows how safe, modern programming languages let Microsoft rethink the architectural trade-offs in a experimental operating system.
Xen and the Art of Virtualization
This paper presents Xen, an x86 virtual machine monitor which allows multiple operating systems to share conventional hardware in a safe and resource managed fashion.