We propose DC-store, a storage framework that offers deterministic I/O performance for a multi-container execution environment. DC-store's hardware-level design implements multiple NVM sets on a shared storage pool, each providing a deterministic SSD access time by removing internal resource conflicts. In parallel, software support of DC-Store is aware of the NVM sets and enlightens Linux kernel to isolate noisy neighbor containers, performing page frame reclaiming, from peers. We prototype both hardware and software counterparts of DC-Store and evaluate them in a real system. The evaluation results demonstrate that containerized data-intensive applications on DC-Store exhibit 31% shorter average execution time, on average, compared to those on a baseline system.
|Title of host publication||Proceedings of the 18th USENIX Conference on File and Storage Technologies, FAST 2020|
|Number of pages||9|
|Publication status||Published - 2020|
|Event||18th USENIX Conference on File and Storage Technologies, FAST 2020 - Santa Clara, United States|
Duration: 2020 Feb 25 → 2020 Feb 27
|Name||Proceedings of the 18th USENIX Conference on File and Storage Technologies, FAST 2020|
|Conference||18th USENIX Conference on File and Storage Technologies, FAST 2020|
|Period||20/2/25 → 20/2/27|
Bibliographical noteFunding Information:
This research is mainly supported by NRF 2016R1C182015312, KAIST Start-up package (G01190015) and Samsung grant (G01190271). The authors thank John Bent of Seagate Technology for shepherding their paper. Myoungsoo Jung is the corresponding author.
All Science Journal Classification (ASJC) codes
- Hardware and Architecture
- Computer Networks and Communications