Qingsong Wei

CDRM: A Cost-Effective Dynamic Replication Management Scheme for Cloud Storage Cluster

Data replication has been widely used as a mean of increasing the data availability of large-scale cloud storage systems where failures are normal. Aiming to provide cost-effective availability, and improve performance and load-balancing of cloud storage, this paper presents a cost-effective dynamic replication management scheme referred to as CDRM. A novel model is proposed to capture the relationship between availability and replica number. CDRM leverages this model to calculate and maintain minimal replica number for a given availability requirement. Replica placement is based on capacity and blocking probability of data nodes. By adjusting replica number and location according to workload changing and node capacity, CDRM can dynamically redistribute workloads among data nodes in the heterogeneous cloud. We implemented CDRM in Hadoop Distributed File System (HDFS) and experiment results conclusively demonstrate that our CDRM is cost effective and outperforms default replication management of HDFS in terms of performance and load balancing for large-scale cloud storage.

Improving flash-based disk cache with Lazy Adaptive Replacement

The increasing popularity of flash memory has changed storage systems. Flash-based solid state drive(SSD) is now widely deployed as cache for magnetic hard disk drives(HDD) to speed up data intensive applications. However, existing cache algorithms focus exclusively on performance improvements and ignore the write endurance of SSD. In this paper, we proposed a novel cache management algorithm for flash-based disk cache, named Lazy Adaptive Replacement Cache(LARC). LARC can filter out seldom accessed blocks and prevent them from entering cache. This avoids cache pollution and keeps popular blocks in cache for a longer period of time, leading to higher hit rate. Meanwhile, LARC reduces the amount of cache replacements thus incurs less write traffics to SSD, especially for read dominant workloads. In this way, LARC improves performance and extends SSD lifetime at the same time. LARC is self-tuning and low overhead. It has been extensively evaluated by both trace-driven simulations and a prototype implementation in flashcache. Our experiments show that LARC outperforms state-of-art algorithms and reduces write traffics to SSD by up to 94.5% for read dominant workloads, 11.2-40.8% for write dominant workloads.

WAFTL: A workload adaptive flash translation layer with data partition

Current FTL schemes have inevitable limitations in terms of memory requirement, performance, garbage collection overhead, and scalability. To overcome these limitations, we propose a workload adaptive flash translation layer referred to as WAFTL. WAFTL explores either page-level or block-level address mapping for normal data block based on access patterns. Page Mapping Block (PMB) is used to store random data and handle large number of partial updates. Block Mapping Block (BMB) is utilized to store sequential data and lower overall mapping table. PMB or BMB is allocated on demand and the number of PMB or BMB eventually depends on workload. An efficient address mapping is designed to reduce overall mapping table and quickly conduct address translation. WAFTL explores a small part of flash space as Buffer Zone to log writes sequentially and migrate data into BMB or PMB based on threshold. Static and dynamic threshold setting are proposed to balance performance and mapping table size. WAFTL has been extensively evaluated under various enterprise workloads. Benchmark results conclusively demonstrate that proposed WAFTL is workload adaptive and achieves up to 80% performance improvement, 83% garbage collection overhead reduction and 50% mapping table reduction compared to existing FTL schemes.

Improving Flash-Based Disk Cache with Lazy Adaptive Replacement

The increasing popularity of flash memory has changed storage systems. Flash-based solid state drive(SSD) is now widely deployed as cache for magnetic hard disk drives(HDD) to speed up data intensive applications. However, existing cache algorithms focus exclusively on performance improvements and ignore the write endurance of SSD. In this paper, we proposed a novel cache management algorithm for flash-based disk cache, named Lazy Adaptive Replacement Cache(LARC). LARC can filter out seldom accessed blocks and prevent them from entering cache. This avoids cache pollution and keeps popular blocks in cache for a longer period of time, leading to higher hit rate. Meanwhile, LARC reduces the amount of cache replacements thus incurs less write traffics to SSD, especially for read dominant workloads. In this way, LARC improves performance and extends SSD lifetime at the same time. LARC is self-tuning and low overhead. It has been extensively evaluated by both trace-driven simulations and a prototype implementation in flashcache. Our experiments show that LARC outperforms state-of-art algorithms and reduces write traffics to SSD by up to 94.5% for read dominant workloads, 11.2-40.8% for write dominant workloads.

FSMAC: A file system metadata accelerator with non-volatile memory

File system performance is dominated by metadata access because it is small and popular. Metadata is stored as block in the file system. Partial metadata update results in whole block read and write which amplifies disk I/O. Huge performance gap between CPU and disk aggravates this problem. In this paper, a file system metadata accelerator (referred as FSMAC) is proposed to optimize metadata access by efficiently exploiting the advantages of Nonvolatile Memory (NVM). FSMAC decouples data and metadata I/O path, putting data on disk and metadata on NVM at runtime. Thus, data is accessed in block from I/O bus and metadata is accessed in byte-addressable manner from memory bus. Metadata access is significantly accelerated and metadata I/O is eliminated because metadata in NVM is not flushed back to disk periodically anymore. A light-weight consistency mechanism combining fine-grained versioning and transaction is introduced in the FSMAC. The FSMAC is implemented on the basis of Linux Ext4 file system and intensively evaluated under different workloads. Evaluation results show that the FSMAC accelerates file system up to 49.2 times for synchronized I/O and 7.22 times for asynchronized I/O.

HRAID6ML: A hybrid RAID6 storage architecture with mirrored logging

The RAID6 provides high reliability using double-parity-update at cost of high write penalty. In this paper, we propose HRAID6ML, a new logging architecture for RAID6 systems for enhanced energy efficiency, performance and reliability. HRAID6ML explores a group of Solid State Drives (SSDs) and Hard Disk Drives (HDDs): Two HDDs (parity disks) and several SSDs form RAID6. The free space of the two parity disks is used as mirrored log region of the whole system to absorb writes. The mirrored logging policy helps to recover system from parity disk failure. Mirrored logging operation does not introduce noticeable performance overhead to the whole system. HRAID6ML eliminates the additional hardware and energy costs, potential single point of failure and performance bottleneck. Furthermore, HRAID6ML prolongs the lifecycle of the SSDs and improves the systems energy efficiency by reducing the SSDs write frequency. We have implemented proposed HRAID6ML. Extensive trace-driven evaluations demonstrate the advantages of the HRAID6ML system over both traditional SSD-based RAID6 system and HDD-based RAID6 system.

Accelerating File System Metadata Access with Byte-Addressable Nonvolatile Memory

File system performance is dominated by small and frequent metadata access. Metadata is stored as blocks on the hard disk drive. Partial metadata update results in whole-block read or write, which significantly amplifies disk I/O. Furthermore, a huge performance gap between the CPU and disk aggravates this problem. In this article, a file system metadata accelerator (referred to as FSMAC) is proposed to optimize metadata access by efficiently exploiting the persistency and byte-addressability of Nonvolatile Memory (NVM). The FSMAC decouples data and metadata access path, putting data on disk and metadata in byte-addressable NVM at runtime. Thus, data is accessed in a block from I/O the bus and metadata is accessed in a byte-addressable manner from the memory bus. Metadata access is significantly accelerated and metadata I/O is eliminated because metadata in NVM is no longer flushed back to the disk periodically. A lightweight consistency mechanism combining fine-grained versioning and transaction is introduced in the FSMAC. The FSMAC is implemented on a real NVDIMM platform and intensively evaluated under different workloads. Evaluation results show that the FSMAC accelerates the file system up to 49.2 times for synchronized I/O and 7.22 times for asynchronized I/O. Moreover, it can achieve significant performance speedup in network storage and database environment, especially for metadata-intensive or write-dominated workloads.

Fine-grained metadata journaling on NVM

Journaling file systems have been widely used where data consistency must be assured. However, we observed that the overhead of journaling can cause up to 48.2% performance drop under certain kinds of workloads. On the other hand, the emerging high-performance, byte-addressable Non-volatile Memory (NVM) has the potential to minimize such overhead by being used as the journal device. The traditional journaling mechanism based on block devices is nevertheless unsuitable for NVM due to the write amplification of metadata journal we observed. In this paper, we propose a fine-grained metadata journal mechanism to fully utilize the low-latency byte-addressable NVM so that the overhead of journaling can be significantly reduced. Based on the observation that conventional block-based metadata journal contains up to 90% clean metadata that is unnecessary to be journalled, we design a fine-grained journal format for byte-addressable NVM which contains only modified metadata. Moreover, we redesign the process of transaction committing, checkpointing and recovery in journaling file systems utilizing the new journal format. Therefore, thanks to the reduced amount of ordered writes to NVM, the overhead of journaling can be reduced without compromising the file system consistency. Experimental results show that our NVM-based fine-grained metadata journaling is up to 15.8× faster than the traditional approach under FileBench workloads.

CBM: A cooperative buffer management for SSD

Random writes significantly limit the application of Solid State Drive (SSD) in the I/O intensive applications such as scientific computing, Web services, and database. While several buffer management algorithms are proposed to reduce random writes, their ability to deal with workloads mixed with sequential and random accesses is limited. In this paper, we propose a cooperative buffer management scheme referred to as CBM, which coordinates write buffer and read cache to fully exploit temporal and spatial localities among I/O intensive workload. To improve both buffer hit rate and destage sequentiality, CBM divides write buffer space into Page Region and Block Region. Randomly written data is put in the Page Region at page granularity, while sequentially written data is stored in the Block Region at block granularity. CBM leverages threshold-based migration to dynamically classify random write from sequential writes. When a block is evicted from write buffer, CBM merges the dirty pages in write buffer and the clean pages in read cache belonging to the evicted block to maximize the possibility of forming full block write. CBM has been extensively evaluated with simulation and real implementation on OpenSSD. Our testing results conclusively demonstrate that CBM can achieve up to 84% performance improvement and 85% garbage collection overhead reduction compared to existing buffer management schemes.

WPAR: A Weight-Based Metadata Management Strategy for Petabyte-Scale Object Storage Systems

In an object-based network storage system, metadata access is decoupled from the data transferring path to improve system performance, scalability, and management. Designing an efficient metadata management scheme in such a system is critically important to the overall system performance and poses lots of challenges to the system designers. Traditional metadata management schemes, either partitioning the metadata statically, or using pure hashing methods, are not able to adapt to the dynamic metadata workload, and hence suffer from scalability and hotspot problems. In this paper, we propose a new metadata management strategy, referred to as Weighted Partitioning and Adaptive Replication (WPAR) scheme, taking into account the weight (CPUpower, memory capacity, network bandwidth, etc.) of each metadata server. Our objective is to efficiently distribute and dynamically replicate the metadata, based on the weights, to balance the workload and reduce hotspots within the metadata server cluster.