[Research] Professor Eom Young-ik's laboratory (Distributed Computing Laboratory, DCLab.) is approved of publication of the SOSP 20
- 소프트웨어융합대학
- Hit833
- 2023-08-28
The 29th ACM Symposium on Operating Systems Principles (SOSP 2023) has been approved for publication in the paper "MEMTIS: Efficient Memory Tiering with Dynamic Page Classification and Page Size Determination" by Professor Eom Young-ik of the Distributed Computing Laboratory and Dr. Lee Tae-hyung. The SOSP Society is the world's leading society for researchers, developers, and programmers in computer systems (BK21+ Recognized International Congress for Computer ScienceIF=4).
This paper proposes how to effectively build large memory systems required by modern data center and cloud computing environments. Professor Eom Young-ik's research team proposed MEMTIS, a new hierarchical memory system utilizing DRAM, non-volatile memory (NVM), and CXL memory devices, which are next-generation hardware. Based on its own high-performance memory page management techniques, MEMTIS delivers up to 169% higher performance than modern tiered memory systems. This study was conducted as an international joint study between Professor Eom Young-ik's research team and Professor Min Chang-woo's research team from Virginia Tech in the United States.
In addition, through the publication of this SOSP paper, the distributed computing laboratory is the first domestic laboratory to publish two or more papers (SOSP 2021 FragPicker, SOSP2023 MEMTIS). In addition, Professor Eom Young-ik's research team will publish its third top-tier conference paper this year alone, following ASPLOS and MobiCom.
[SOSP 2023]
The 29th ACM Symposium on Operating Systems Principles, October 23-26, 2023
[About the thesis]
MEMTIS: Efficient Memory Tiering with Dynamic Page Classification and Page Size Determination
Taehyung Lee, Sumit Kumar Monga, Changwoo Min, Young Ik Eom
29th Symposium on Operating Systems Principles (SOSP 2023)
Abstract:
The evergrowing memory demand fueled by datacenter workloads is the driving force behind new memory technology innovations (e.g., NVM, CXL). Tiered memory system is a promising solution which harnesses such multiple memory types with varying capacity, latency, and cost characteristics in an effort to reduce server hardware costs while fulfilling memory demand. Prior works on memory tiering make suboptimal (often pathological) page placement decisions because they rely on various heuristics and static thresholds without considering overall memory access distribution. Also, deciding the appropriate page size for an application is difficult as huge pages are not always beneficial as a result of skewed accesses within them. We present Memtis, a tiered memory system that adopts an informed decision-making for page placement and page size determination. Memtis leverages access distribution of allocated pages to optimally approximate the hot data set to the fast tier capacity. Moreover, Memtis dynamically determines the page size that allows applications to use huge pages while avoiding their drawbacks by detecting inefficient use of fast tier memory and splintering them if necessary. Our evaluation shows that Memtis outperforms state-of-the-art tiering systems by up to 169.0% and their best by up to 33.6%.
Distributed Computing Lab: http://dclab.skku.ac.kr/xe/