BZPEERPapers, people, evidence and research paths
Menu

Discover

SearchFind papers, people, topics, institutions and opportunities from one place.GraphSee how papers, authors, topics and evidence connect.

Research tools

HomeContinue from the papers, people and topics that matter now.CollectionsTurn saved papers into reading lists, evidence maps and shareable context.ResearchReconstruct the path that led you from search to paper to question.CollaborationMove from a paper, topic or expert request into a focused thread.PublishCreate full-text research articles with metadata, formulas and versions.

Network

ExpertsFind specialists whose work and reviews explain why they can help.CommunitiesJoin research groups organized around papers, topics and shared evidence.

Opportunities

ListingsBrowse roles, calls and collaborations connected to real research context.

Account

Log inReturn to your saved papers, graph views and active threads.Create accountStart saving papers, building a research trail and joining teams.
Log inCreate account

Source author record

Patrick P. C. Lee

Patrick P. C. Lee appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record
Distributed, Parallel, and Cluster ComputingInformation Theorymath.ITPerformanceArtificial IntelligenceCryptography and SecurityDatabasesNetworking and Internet ArchitectureOperating Systems

Catalog footprint

What is connected

14works
9topics
4close collaborators

Actions

Connect this record

Log in to claim
Open graphBrowse works

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

14 published item(s)

preprint2026arXiv

TierCheck: Tiered Checkpointing for Fault Tolerance in Large Language Model Training

Large Language Model (LLM) training is frequently interrupted by a heterogeneous spectrum of failures, from common GPU crashes to catastrophic cluster-wide outages. Existing checkpointing systems rely on monolithic, single-tier storage backend, forcing a trade-off between state-saving overhead and recovery speed. We propose TierCheck, a cluster-aware tiered checkpointing system that aligns storage placement with failure heterogeneity. TierCheck adopts a three-tier design that maintains lightweight differential checkpoints in local and peer memory for fast localized recovery, while asynchronously migrating heavyweight base checkpoints to remote persistent storage. It also ensures strict global consistency across tiers without stalling training, and achieves fast cluster-aware checkpoint restoration during recovery. Evaluations on models up to 40 billion parameters show that TierCheck achieves low training overhead, reduces end-to-end checkpointing time to under 10s, and supports high-frequency checkpointing, ultimately striking an optimal balance between low-overhead persistence and fast recovery.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A Generalization of Array Codes with Local Properties and Efficient Encoding/Decoding

A maximum distance separable (MDS) array code is composed of $m\times (k+r)$ arrays such that any $k$ out of $k+r$ columns suffice to retrieve all the information symbols. Expanded-Blaum-Roth (EBR) codes and Expanded-Independent-Parity (EIP) codes are two classes of MDS array codes that can repair any one symbol in a column by locally accessing some other symbols within the column, where the number of symbols $m$ in a column is a prime number. By generalizing the constructions of EBR and EIP codes, we propose new MDS array codes, such that any one symbol can be locally recovered and the number of symbols in a column can be not only a prime number but also a power of an odd prime number. Also, we present an efficient encoding/decoding method for the proposed generalized EBR (GEBR) and generalized EIP (GEIP) codes based on the LU factorization of a Vandermonde matrix. We show that the proposed decoding method has less computational complexity than existing methods. Furthermore, we show that the proposed GEBR codes have both a larger minimum symbol distance and a larger recovery ability of erased lines for some parameters when compared to EBR codes. We show that EBR codes can recover any $r$ erased lines of a slope for any parameter $r$, which was an open problem in [2].

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Efficient LSM-Tree Key-Value Data Management on Hybrid SSD/HDD Zoned Storage

Zoned storage devices, such as zoned namespace (ZNS) solid-state drives (SSDs) and host-managed shingled magnetic recording (HM-SMR) hard-disk drives (HDDs), expose interfaces for host-level applications to support fine-grained, high-performance storage management. Combining ZNS SSDs and HM-SMR HDDs into a unified hybrid storage system is a natural direction to scale zoned storage at low cost, yet how to effectively incorporate zoned storage awareness into hybrid storage is a non-trivial issue. We make a case for key-value (KV) stores based on log-structured merge trees (LSM-trees) as host-level applications, and present HHZS, a middleware system that bridges an LSM-tree KV store with hybrid zoned storage devices based on hints. HHZS leverages hints issued by the flushing, compaction, and caching operations of the LSM-tree KV store to manage KV objects in placement, migration, and caching in hybrid ZNS SSD and HM-SMR HDD zoned storage. Experiments show that our HHZS prototype, when running on real ZNS SSD and HM-SMR HDD devices, achieves the highest throughput compared with all baselines under various settings.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Separating Data via Block Invalidation Time Inference for Write Amplification Reduction in Log-Structured Storage

Log-structured storage has been widely deployed in various domains of storage systems, yet its garbage collection incurs write amplification (WA) due to the rewrites of live data. We show that there exists an optimal data placement scheme that minimizes WA using the future knowledge of block invalidation time (BIT) of each written block, yet it is infeasible to realize in practice. We propose a novel data placement algorithm for reducing WA, SepBIT, that aims to infer the BITs of written blocks from storage workloads and separately place the blocks into groups with similar estimated BITs. We show via both mathematical and production trace analyses that SepBIT effectively infers the BITs by leveraging the write skewness property in practical storage workloads. Trace analysis and prototype experiments show that SepBIT reduces WA and improves I/O throughput, respectively, compared with state-of-the-art data placement schemes. SepBIT is currently deployed to support the log-structured block storage management at Alibaba Cloud.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Two New Piggybacking Designs with Lower Repair Bandwidth

Piggybacking codes are a special class of MDS array codes that can achieve small repair bandwidth with small sub-packetization by first creating some instances of an $(n,k)$ MDS code, such as a Reed-Solomon (RS) code, and then designing the piggyback function. In this paper, we propose a new piggybacking coding design which designs the piggyback function over some instances of both $(n,k)$ MDS code and $(n,k')$ MDS code, when $k\geq k'$. We show that our new piggybacking design can significantly reduce the repair bandwidth for single-node failures. When $k=k'$, we design piggybacking code that is MDS code and we show that the designed code has lower repair bandwidth for single-node failures than all existing piggybacking codes when the number of parity node $r=n-k\geq8$ and the sub-packetization $α<r$. Moreover, we propose another piggybacking codes by designing $n$ piggyback functions of some instances of $(n,k)$ MDS code and adding the $n$ piggyback functions into the $n$ newly created empty entries with no data symbols. We show that our code can significantly reduce repair bandwidth for single-node failures at a cost of slightly more storage overhead. In addition, we show that our code can recover any $r+1$ node failures for some parameters. We also show that our code has lower repair bandwidth than locally repairable codes (LRCs) under the same fault-tolerance and redundancy for some parameters.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A Fast and Compact Invertible Sketch for Network-Wide Heavy Flow Detection

Fast detection of heavy flows (e.g., heavy hitters and heavy changers) in massive network traffic is challenging due to the stringent requirements of fast packet processing and limited resource availability. Invertible sketches are summary data structures that can recover heavy flows with small memory footprints and bounded errors, yet existing invertible sketches incur high memory access overhead that leads to performance degradation. We present MV-Sketch, a fast and compact invertible sketch that supports heavy flow detection with small and static memory allocation. MV-Sketch tracks candidate heavy flows inside the sketch data structure via the idea of majority voting, such that it incurs small memory access overhead in both update and query operations, while achieving high detection accuracy. We present theoretical analysis on the memory usage, performance, and accuracy of MV-Sketch in both local and network-wide scenarios. We further show how MV-Sketch can be implemented and deployed on P4-based programmable switches subject to hardware deployment constraints. We conduct evaluation in both software and hardware environments. Trace-driven evaluation in software shows that MV-Sketch achieves higher accuracy than existing invertible sketches, with up to 3.38x throughput gain. We also show how to boost the performance of MV-Sketch with SIMD instructions. Furthermore, we evaluate MV-Sketch on a Barefoot Tofino switch and show how MV-Sketch achieves line-rate measurement with limited hardware resource overhead.

0 citations0 reviewsNo code linkedOpen access
preprint2015arXiv

CDStore: Toward Reliable, Secure, and Cost-Efficient Cloud Storage via Convergent Dispersal

We present CDStore, which disperses users' backup data across multiple clouds and provides a unified multi-cloud storage solution with reliability, security, and cost-efficiency guarantees. CDStore builds on an augmented secret sharing scheme called convergent dispersal, which supports deduplication by using deterministic content-derived hashes as inputs to secret sharing. We present the design of CDStore, and in particular, describe how it combines convergent dispersal with two-stage deduplication to achieve both bandwidth and storage savings and be robust against side-channel attacks. We evaluate the performance of our CDStore prototype using real-world workloads on LAN and commercial cloud testbeds. Our cost analysis also demonstrates that CDStore achieves a monetary cost saving of 70% over a baseline cloud storage solution using state-of-the-art secret sharing.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

Efficient Hybrid Inline and Out-of-line Deduplication for Backup Storage

Backup storage systems often remove redundancy across backups via inline deduplication, which works by referring duplicate chunks of the latest backup to those of existing backups. However, inline deduplication degrades restore performance of the latest backup due to fragmentation, and complicates deletion of ex- pired backups due to the sharing of data chunks. While out-of-line deduplication addresses the problems by forward-pointing existing duplicate chunks to those of the latest backup, it introduces additional I/Os of writing and removing duplicate chunks. We design and implement RevDedup, an efficient hybrid inline and out-of-line deduplication system for backup storage. It applies coarse-grained inline deduplication to remove duplicates of the latest backup, and then fine-grained out-of-line reverse deduplication to remove duplicates from older backups. Our reverse deduplication design limits the I/O overhead and prepares for efficient deletion of expired backups. Through extensive testbed experiments using synthetic and real-world datasets, we show that RevDedup can bring high performance to the backup, restore, and deletion operations, while maintaining high storage efficiency comparable to conventional inline deduplication.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

STAIR Codes: A General Family of Erasure Codes for Tolerating Device and Sector Failures

Practical storage systems often adopt erasure codes to tolerate device failures and sector failures, both of which are prevalent in the field. However, traditional erasure codes employ device-level redundancy to protect against sector failures, and hence incur significant space overhead. Recent sector-disk (SD) codes are available only for limited configurations. By making a relaxed but practical assumption, we construct a general family of erasure codes called \emph{STAIR codes}, which efficiently and provably tolerate both device and sector failures without any restriction on the size of a storage array and the numbers of tolerable device failures and sector failures. We propose the \emph{upstairs encoding} and \emph{downstairs encoding} methods, which provide complementary performance advantages for different configurations. We conduct extensive experiments on STAIR codes in terms of space saving, encoding/decoding speed, and update cost. We demonstrate that STAIR codes not only improve space efficiency over traditional erasure codes, but also provide better computational efficiency than SD codes based on our special code construction. Finally, we present analytical models that characterize the reliability of STAIR codes, and show that the support of a wider range of configurations by STAIR codes is critical for tolerating sector failure bursts discovered in the field.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Analysis and Construction of Functional Regenerating Codes with Uncoded Repair for Distributed Storage Systems

Modern distributed storage systems apply redundancy coding techniques to stored data. One form of redundancy is based on regenerating codes, which can minimize the repair bandwidth, i.e., the amount of data transferred when repairing a failed storage node. Existing regenerating codes mainly require surviving storage nodes encode data during repair. In this paper, we study functional minimum storage regenerating (FMSR) codes, which enable uncoded repair without the encoding requirement in surviving nodes, while preserving the minimum repair bandwidth guarantees and also minimizing disk reads. Under double-fault tolerance settings, we formally prove the existence of FMSR codes, and provide a deterministic FMSR code construction that can significantly speed up the repair process. We further implement and evaluate our deterministic FMSR codes to show the benefits. Our work is built atop a practical cloud storage system that implements FMSR codes, and we provide theoretical validation to justify the practicality of FMSR codes.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

CORE: Augmenting Regenerating-Coding-Based Recovery for Single and Concurrent Failures in Distributed Storage Systems

Data availability is critical in distributed storage systems, especially when node failures are prevalent in real life. A key requirement is to minimize the amount of data transferred among nodes when recovering the lost or unavailable data of failed nodes. This paper explores recovery solutions based on regenerating codes, which are shown to provide fault-tolerant storage and minimum recovery bandwidth. Existing optimal regenerating codes are designed for single node failures. We build a system called CORE, which augments existing optimal regenerating codes to support a general number of failures including single and concurrent failures. We theoretically show that CORE achieves the minimum possible recovery bandwidth for most cases. We implement CORE and evaluate our prototype atop a Hadoop HDFS cluster testbed with up to 20 storage nodes. We demonstrate that our CORE prototype conforms to our theoretical findings and achieves recovery bandwidth saving when compared to the conventional recovery approach based on erasure codes.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

RevDedup: A Reverse Deduplication Storage System Optimized for Reads to Latest Backups

Scaling up the backup storage for an ever-increasing volume of virtual machine (VM) images is a critical issue in virtualization environments. While deduplication is known to effectively eliminate duplicates for VM image storage, it also introduces fragmentation that will degrade read performance. We propose RevDedup, a deduplication system that optimizes reads to latest VM image backups using an idea called reverse deduplication. In contrast with conventional deduplication that removes duplicates from new data, RevDedup removes duplicates from old data, thereby shifting fragmentation to old data while keeping the layout of new data as sequential as possible. We evaluate our RevDedup prototype using microbenchmark and real-world workloads. For a 12-week span of real-world VM images from 160 users, RevDedup achieves high deduplication efficiency with around 97% of saving, and high backup and read throughput on the order of 1GB/s. RevDedup also incurs small metadata overhead in backup/read operations.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Stochastic Analysis on RAID Reliability for Solid-State Drives

Solid-state drives (SSDs) have been widely deployed in desktops and data centers. However, SSDs suffer from bit errors, and the bit error rate is time dependent since it increases as an SSD wears down. Traditional storage systems mainly use parity-based RAID to provide reliability guarantees by striping redundancy across multiple devices, but the effectiveness of RAID in SSDs remains debatable as parity updates aggravate the wearing and bit error rates of SSDs. In particular, an open problem is that how different parity distributions over multiple devices, such as the even distribution suggested by conventional wisdom, or uneven distributions proposed in recent RAID schemes for SSDs, may influence the reliability of an SSD RAID array. To address this fundamental problem, we propose the first analytical model to quantify the reliability dynamics of an SSD RAID array. Specifically, we develop a "non-homogeneous" continuous time Markov chain model, and derive the transient reliability solution. We validate our model via trace-driven simulations and conduct numerical analysis to provide insights into the reliability dynamics of SSD RAID arrays under different parity distributions and subject to different bit error rates and array configurations. Designers can use our model to decide the appropriate parity distribution based on their reliability requirements.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Stochastic Modeling of Large-Scale Solid-State Storage Systems: Analysis, Design Tradeoffs and Optimization

Solid state drives (SSDs) have seen wide deployment in mobiles, desktops, and data centers due to their high I/O performance and low energy consumption. As SSDs write data out-of-place, garbage collection (GC) is required to erase and reclaim space with invalid data. However, GC poses additional writes that hinder the I/O performance, while SSD blocks can only endure a finite number of erasures. Thus, there is a performance-durability tradeoff on the design space of GC. To characterize the optimal tradeoff, this paper formulates an analytical model that explores the full optimal design space of any GC algorithm. We first present a stochastic Markov chain model that captures the I/O dynamics of large-scale SSDs, and adapt the mean-field approach to derive the asymptotic steady-state performance. We further prove the model convergence and generalize the model for all types of workload. Inspired by this model, we propose a randomized greedy algorithm (RGA) that can operate along the optimal tradeoff curve with a tunable parameter. Using trace-driven simulation on DiskSim with SSD add-ons, we demonstrate how RGA can be parameterized to realize the performance-durability tradeoff.

0 citations0 reviewsNo code linkedOpen access