the proceedings contain 23 papers. the topics discussed include: lock-free collaboration support for cloud storage services with operation inference and transformation;carver: finding important parameters for storage ...
ISBN:
(纸本)9781939133120
the proceedings contain 23 papers. the topics discussed include: lock-free collaboration support for cloud storage services with operation inference and transformation;carver: finding important parameters for storage system tuning;uncovering access, reuse, and sharing characteristics of I/O-intensive files on large-scale production HPC systems;scalable parallel flash firmware for many-core architectures;an empirical guide to the behavior and use of scalable persistent memory;DC-store: eliminating noisy neighbor containers using deterministic I/O performance and resource isolation;and InfiniCache: exploiting ephemeral serverless functions to build a cost-effective memory cache.
the proceedings contain 27 papers. the topics discussed include: a logic of file systems;accurate and efficient replaying of file system traces;MicroHash: an efficient index structure for flash-based sensor devices;ad...
the proceedings contain 27 papers. the topics discussed include: a logic of file systems;accurate and efficient replaying of file system traces;MicroHash: an efficient index structure for flash-based sensor devices;adaptive data placement for wide-area sensing services;ursa minor: versatile cluster-based storage;zodiac: efficient impact analysis for storage area networks;DULO: an effective buffer cache management scheme to exploit both temporal and spatial localities;and a security model for full-text file system search in multi-user environments.
the proceedings contain 28 papers. the topics discussed include: ROART: range-query optimized persistent ART;SpanDB: a fast, cost-effective LSM-tree based KV store on hybrid storage;evolution of development priorities...
ISBN:
(纸本)9781939133205
the proceedings contain 28 papers. the topics discussed include: ROART: range-query optimized persistent ART;SpanDB: a fast, cost-effective LSM-tree based KV store on hybrid storage;evolution of development priorities in key-value stores serving large-scale applications: the RocksDB experience;high velocity kernel file systems with bento;scalable persistent memory file system with kernel-userspace collaboration;rethinking file mapping for persistent memory;pattern-guided file compression with user-experience enhancement for log-structured file system on mobile devices;ArchTM: architecture-aware, high performance transaction for persistent memory;the dilemma between deduplication and locality: can both be achieved?;and remap-SSD: safely and efficiently exploiting SSD address remapping to eliminate duplicate writes.
the proceedings contain 28 papers. the topics discussed include: NyxCache: flexible and efficient multi-tenant persistent memory caching;HTMFS: strong consistency comes for free with hardware transactional memory in p...
ISBN:
(纸本)9781939133267
the proceedings contain 28 papers. the topics discussed include: NyxCache: flexible and efficient multi-tenant persistent memory caching;HTMFS: strong consistency comes for free with hardware transactional memory in persistent memory file systems;ctFS: replacing file indexing with hardware memory translation through contiguous file allocation for persistent memory;FORD: fast one-sided RDMA-based distributed transactions for disaggregated persistent memory;closing the B+-tree vs. LSM-tree write amplification gap on modern storage hardware with built-in transparent compression;TVStore: automatically bounding time series storage via time-varying compression;removing double-logging with passive data persistence in LSM-tree based relational databases;hardware/software co-programmable framework for computational SSDs to accelerate deep learning service on large-scale graphs;and Aurogon: taming aborts in all phases for distributed in-memory transactions.
the proceedings contain 20 papers. the topics discussed include: a study of practical deduplication;tradeoffs in scalable data routing for deduplication clusters;capo: recapitulating storage for virtual desktops;consi...
ISBN:
(纸本)9781931971829
the proceedings contain 20 papers. the topics discussed include: a study of practical deduplication;tradeoffs in scalable data routing for deduplication clusters;capo: recapitulating storage for virtual desktops;consistent and durable data structures for non-volatile byte-addressable memory;CAFTL: a content-aware flash translation layer enhancing the lifespan of flash memory based solid state drives;leveraging value locality in optimizing NAND flash-based SSDs;reliably erasing data from flash-based solid state drives;a scheduling framework that makes any disk schedulers non-work-conserving solely based on request characteristics;and the SCADS director: scaling a distributed storage system under stringent performance requirements.
Traditional file system optimizations typically use a one-to-one mapping of logical files to their physical metadata representations. this mapping results in missed opportunities for a class of optimizations in which ...
详细信息
ISBN:
(纸本)9781931971287
Traditional file system optimizations typically use a one-to-one mapping of logical files to their physical metadata representations. this mapping results in missed opportunities for a class of optimizations in which such coupling is removed. We have designed, implemented, and evaluated a composite-filefile system, which allows many-to-one mappings of files to metadata, and we have explored the design space of different mapping strategies. Under webserver and software development workloads, our empirical evaluation shows up to a 27% performance improvement. this result demonstrates the promise of composite files.
We present WiscKey, a persistent LSM-tree-based key-value store with a performance-oriented data layout that separates keys from values to minimize I/O amplification. the design of WiscKey is highly SSD optimized, lev...
详细信息
ISBN:
(纸本)9781931971287
We present WiscKey, a persistent LSM-tree-based key-value store with a performance-oriented data layout that separates keys from values to minimize I/O amplification. the design of WiscKey is highly SSD optimized, leveraging boththe sequential and random performance characteristics of the device. We demonstrate the advantages of WiscKey with both microbenchmarks and YCSB workloads. Microbenchmark results show that WiscKey is 2.5x-111x faster than LevelDB for loading a database and 1.6x-14x faster for random lookups. WiscKey is faster than both LevelDB and RocksDB in all six YCSB workloads.
In flash storage, an FTL is a complex piece of code that resides completely inside the storage device and is provided by the manufacturer. Its principal virtue is providing interoperability with conventional HDDs. How...
详细信息
ISBN:
(纸本)9781931971287
In flash storage, an FTL is a complex piece of code that resides completely inside the storage device and is provided by the manufacturer. Its principal virtue is providing interoperability with conventional HDDs. However, this virtue is also its biggest impediment in reaching the full performance of the underlying flash storage. We propose to refactor the flash storage architecture so that it relies on a new block I/O interface which does not permit overwriting of data without intervening erasures. We demonstrate how high-level applications, in particular file systems, can deal withthis restriction efficiently by employing append-only segments. this refactoring dramatically reduces flash management overhead and improves performance of applications, such as file systems and databases, by permitting them to directly manage flash storage. Our experiments on a machine withthe new block I/O interface show that DRAM in the flash controller is reduced by 128X and the performance of the file system improves by 80% over conventional SSDs.
Fast non-volatile memories (NVMs) will soon appear on the processor memory bus alongside DRAM. the resulting hybrid memory systems will provide software with submicrosecond, high-bandwidth access to persistent data, b...
详细信息
ISBN:
(纸本)9781931971287
Fast non-volatile memories (NVMs) will soon appear on the processor memory bus alongside DRAM. the resulting hybrid memory systems will provide software with submicrosecond, high-bandwidth access to persistent data, but managing, accessing, and maintaining consistency for data stored in NVM raises a host of challenges. Existing file systems built for spinning or solid-state disks introduce software overheads that would obscure the performance that NVMs should provide, but proposed file systems for NVMs either incur similar overheads or fail to provide the strong consistency guarantees that applications require. We present NOVA, a file system designed to maximize performance on hybrid memory systems while providing strong consistency guarantees. NOVA adapts conventional log-structured file system techniques to exploit the fast random access that NVMs provide. In particular, it maintains separate logs for each inode to improve concurrency, and stores file data outside the log to minimize log size and reduce garbage collection costs. NOVA's logs provide metadata, data, and mmap atomicity and focus on simplicity and reliability, keeping complex metadata structures in DRAM to accelerate lookup operations. Experimental results show that in write-intensive workloads, NOVA provides 22% to 216x throughput improvement compared to state-of-the-art file systems, and 3.1x to 13.5x improvement compared to file systems that provide equally strong data consistency guarantees.
暂无评论