Adaptive Page Migration Policy With Huge Pages in Tiered Memory Systems

Heo, Taekyung; Wang, Yan; Cui, Wei; Huh, Jaehyuk; Zhang, Lintao

doi:10.1109/tc.2020.3036686

Cited by 10 publications

(7 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While in-memory-analytics presents 70% of its performance with the 40% local memory ratio, kmeans shows 6.3% of its performance. We qualitatively analyze the root cause with the tools presented in prior studies [38,19], which allow us to analyze the memory access frequency. We found that the reason behind the different sensitivity on the local memory ratio is the various memory access patterns and intensity.…”

Section: Macrobenchmark Resultsmentioning

confidence: 99%

“…Moreover, the page replacement algorithm has been designed in the context of swapping pages to a slow disk, which is quite different from the emerging memory systems. A few recent studies tried to optimize the page replacement policy in the context of tiered memory systems [56,19], and Google shared the improvement for page replacement, the multi-generational LRU [33,34] to the Linux kernel community.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Hardware-assisted Trusted Memory Disaggregation for Secure Far Memory

Heo,

Kang,

Lee

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Memory disaggregation provides efficient memory utilization across network-connected systems. It allows a node to use part of memory in remote nodes in the same cluster. Recent studies have improved RDMA-based memory disaggregation systems, supporting lower latency and higher bandwidth than the prior generation of disaggregated memory. However, the current disaggregated memory systems manage remote memory only at coarse granularity due to the limitation of the access validation mechanism of RDMA. In such systems, to support fine-grained remote page allocation, the trustworthiness of all participating systems needs to be assumed, and thus a security breach in a node can propagate to the entire cluster. From the security perspective, the memory-providing node must protect its memory from memory-requesting nodes. On the other hand, the memory-requesting node requires the confidentiality and integrity protection of its memory contents even if they are stored in remote nodes. To address the weak isolation support in the current system, this study proposes a novel hardware-assisted memory disaggregation system. Based on the security features of FPGA, the logic in each per-node FPGA board provides a secure memory disaggregation engine. With its own networks, a set of FPGA-based engines form a trusted memory disaggregation system, which is isolated from the privileged software of each participating node. The secure memory disaggregation system allows fine-grained memory management in memory-providing nodes, while the access validation is guaranteed with the hardware-hardened mechanism. In addition, the proposed system hides the memory access patterns observable from remote nodes, supporting obliviousness. Our evaluation with FPGA implementation shows that such fine-grained secure disaggregated memory is feasible with comparable performance to the latest software-based techniques.

show abstract

Section: Macrobenchmark Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Hardware-assisted Trusted Memory Disaggregation for Secure Far Memory

Heo,

Kang,

Lee

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Time-based sampling is periodically scanning if each memory page was accessed [31,40,41,57] via scanning accessed bits in page tables. This does not require source code or binary modifications and incurs much less overhead compared to tracing.…”

Section: Data Access Monitoringmentioning

confidence: 99%

“…Space-based sampling is an alternative. It refers to increasing the granularity of the monitoring space unit [31,61] by splitting the monitored regions into small sub-regions that are larger than a single page. Monitoring only one page in each region can mitigate the unbounded overhead problem.…”

Section: Data Access Monitoringmentioning

confidence: 99%

“…Program execution context-based access profiling for the dynamic access pattern changes have been used for SSD write grouping [28,38] and data item placement on heterogeneous memory systems [54]. For data item prioritization, recency only (LRU) [19,23], frequency only (LFU) [27,35], and adaptive approaches [11,31,43,47,51] have been explored.…”

Section: Access-aware Optimizationmentioning

confidence: 99%

See 1 more Smart Citation

Daos

Park

Bhowmik

Uta³

2022

Proceedings of the 31st International Symposium on High-Performance Parallel and Distributed Computing

View full text Add to dashboard Cite

In data-intensive workloads, data placement and memory management are inherently difficult: the programmer and the operating system have to choose between (combinations of) DRAM and storage, replacement policies, as well as paging sizes. Efficient memory management is based on fine-grained data access patterns driving placement decisions. Current solutions in this space cannot be applied to general workloads and production systems due to either unrealistic assumptions or prohibitive monitoring overheads.To overcome these issues, we introduce DAOS, an open-source system for general data access-aware memory management. DAOS provides a data access monitoring framework that utilizes practical best-effort trade-offs between overhead and accuracy. The memory management engine of DAOS allows users to implement their access-aware management with no code, just simple configuration schemes. For system administrators, DAOS provides a runtime system that auto-tunes the schemes for user-defined objectives in a finite time. We evaluated DAOS on commercial service production systems as well as state-of-the-art benchmarks. DAOS achieves up to 12% performance improvement and 91% memory saving. DAOS is upstreamed and available in the Linux kernel.

show abstract

Intelligent Page Migration on Heterogeneous Memory by Using Transformer

Pei,

Qin,

et al. 2024

Int J Parallel Prog

View full text Add to dashboard Cite

Adaptive Page Migration Policy With Huge Pages in Tiered Memory Systems

Cited by 10 publications

References 57 publications

Hardware-assisted Trusted Memory Disaggregation for Secure Far Memory

Hardware-assisted Trusted Memory Disaggregation for Secure Far Memory

Daos

Intelligent Page Migration on Heterogeneous Memory by Using Transformer

Contact Info

Product

Resources

About