Nvwal

Kim, Wook-Hee; Kim, Jinwoong; Baek, Woongki; Nam, Beomseok; Won, Youjip

doi:10.1145/2954680.2872392

Cited by 11 publications

(7 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Current studies (Kim et al 2016;Mohan et al 1992;Chatzistergiou et al 2015;Malviya et al 2014;Stonebraker et al 2007) adopt logging and checkpointing to achieve persistence. Most systems apply write-ahead logging (keeping redo or undo information) in persistent storage to guarantee the atomicity and durability of the transactional updates being committed.…”

Section: Background and Challenges 21 Data Persistence Techniquesmentioning

confidence: 99%

“…Most systems apply write-ahead logging (keeping redo or undo information) in persistent storage to guarantee the atomicity and durability of the transactional updates being committed. Some systems (Kim et al 2016;Mohan et al 1992;Chatzistergiou et al 2015) use value logging; they store all modified data within a transaction in persistent value logs. If the application or system crashes, then it can apply modifications in value logs on the last checkpoint to recover the persistent data to be the state just after the last committed transaction.…”

Section: Background and Challenges 21 Data Persistence Techniquesmentioning

confidence: 99%

“…As a result, persistence of every log via flush and memory fence operations on NVRAM brings about considerable runtime overhead. NVRAM Write-Ahead Logging (NVWAL) (Kim et al 2016) proposes a transaction-aware lazy synchronization methodology for NVRAM. It works by group-based flushing and memory fencing of the pending logs in the transaction on commit.…”

Section: Runtime Logging Overhead and Abort Windows Formentioning

confidence: 99%

“…Despite growing research efforts to design persistence systems exploiting NVRAM, there are still open challenges regarding performance. First, most persistence systems (Volos et al 2011;Coburn et al 2011;Kim et al 2016;Huang et al 2014) are using transactional semantics to execute programs. A transaction usually contains multiple modifications, each of which generates a log.…”

Section: Introductionmentioning

confidence: 99%

“…Further research on reducing the overhead of persisting logs and minimizing the abort probability is necessary for tapping into the real performance advantage of NVRAM. Second, many systems (Kim et al 2016;Chatzistergiou et al 2015;Huang et al 2014) employ a combination of modified value logging (a.k.a. ARIES-type physiological logging) (Mohan et al 1992) and checkpointing to achieve persistence.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Simpo

Zhang

Lam

Yao

et al. 2018

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

While CPU architectures are incorporating many more cores to meet ever-bigger workloads, advance in fault-tolerance support is indispensable for sustaining system performance under reliability constraints. Emerging non-volatile memory technologies are yielding fast, dense, and energy-efficient NVRAM that can dethrone SSD drives for persisting data. Research on using NVRAM to enable fast in-memory data persistence is ongoing. In this work, we design and implement a persistent object framework, dubbed scalable in-memory persistent object (SIMPO) , which exploits NVRAM, alongside DRAM, to support efficient object persistence in highly threaded big data applications. Based on operation logging, we propose a new programming model that classifies functions into instant and deferrable groups. SIMPO features a streamlined execution model, which allows lazy evaluation of deferrable functions and is well suited to big data computing workloads that would see improved data locality and concurrency. Our log recording and checkpointing scheme is effectively optimized towards NVRAM, mitigating its long write latency through write-combining and consolidated flushing techniques. Efficient persistent object management with features including safe references and memory leak prevention is also implemented and tailored to NVRAM. We evaluate a wide range of SIMPO-enabled applications with machine learning, high-performance computing, and database workloads on an emulated hybrid memory architecture and a real hybrid memory machine with NVDIMM. Compared with native applications without persistence, experimental results show that SIMPO incurs less than 5% runtime overhead on both platforms and even gains up to 2.5× speedup and 84% increase in throughput in highly threaded situations on the two platforms, respectively, thanks to the streamlined execution model.

show abstract