To satisfy increasing storage demands in both capacity and performance, industry has turned to multiple storage technologies, including Flash SSDs and SMR disks. These devices employ a translation layer that conceals the idiosyncrasies of their mediums and enables random access. Device translation layers are, however, inherently constrained: resources on the drive are scarce, they cannot be adapted to application requirements, and lack visibility across multiple devices. As a result, performance and durability of many storage devices is severely degraded. In this paper, we present SALSA: a translation layer that executes on the host and allows unmodified applications to better utilize commodity storage. SALSA supports a wide range of single-and multi-device optimizations and, because is implemented in software, can adapt to specific workloads. We describe SALSA's design, and demonstrate its significant benefits using microbenchmarks and case studies based on three applications: MySQL, the Swift object store, and a video server.

Despite its widespread use in consumer devices and enterprise storage systems, NAND flash faces a growing number of challenges. While technology advances have helped to increase the storage density and reduce costs, they have also led to reduced endurance and larger block variations, which cannot be compensated solely by stronger ECC or read-retry schemes but have to be addressed holistically. Our goal is to enable low-cost NAND flash in enterprise storage for cost efficiency. We present novel flash-management approaches that reduce write amplification, achieve better wear leveling, and enhance endurance without sacrificing performance. We introduce block calibration, a technique to determine optimal read-threshold voltage levels that minimize error rates, and novel garbage-collection as well as data-placement schemes that alleviate the effects of block health variability and show how these techniques complement one another and thereby achieve enterprise storage requirements. By combining the proposed schemes, we improve endurance by up to 15× compared to the baseline endurance of NAND flash without using a stronger ECC scheme. The flash-management algorithms presented herein were designed and implemented in simulators, hardware test platforms, and eventually in the flash controllers of production enterprise all-flash arrays. Their effectiveness has been validated across thousands of customer deployments since 2015.

During the past decade, network and storage devices have undergone rapid performance improvements, delivering ultra-low latency and several Gbps of bandwidth. Nevertheless, current network and storage stacks fail to deliver this hardware performance to the applications, often due to the loss of I/O efficiency from stalled CPU performance. While many efforts attempt to address this issue solely on either the network or the storage stack, achieving high-performance for networked-storage applications requires a holistic approach that considers both. In this article, we present FlashNet, a software I/O stack that unifies high-performance network properties with flash storage access and management. FlashNet builds on RDMA principles and abstractions to provide a direct, asynchronous, end-to-end data path between a client and remote flash storage. The key insight behind FlashNet is to co-design the stack’s components (an RDMA controller, a flash controller, and a file system) to enable cross-stack optimizations and maximize I/O efficiency. In micro-benchmarks, FlashNet improves 4kB network I/O operations per second (IOPS by 38.6% to 1.22M, decreases access latency by 43.5% to 50.4μs, and prolongs the flash lifetime by 1.6-5.9× for writes. We illustrate the capabilities of FlashNet by building a Key-Value store and porting a distributed data store that uses RDMA on it. The use of FlashNet’s RDMA API improves the performance of KV store by 2× and requires minimum changes for the ported data store to access remote flash devices.

During the past decade, network and storage devices have undergone rapid performance improvements, delivering ultra-low latency and several Gbps of bandwidth. Nevertheless, current network and storage stacks fail to deliver this hardware performance to the applications, often due to the loss of IO efficiency from stalled CPU performance. While many efforts attempt to address this issue solely on either the network or the storage stack, achieving high-performance for networked-storage applications requires a holistic approach that considers both.

The explosion of data volumes in enterprise environments and limited budgets have triggered the need for multi-tiered storage systems. With the bulk of the data being extremely infrequently accessed, tape is a natural fit for storing such data. In this paper we present our approach to a file storage system that seamlessly integrates disk and tape, enabling a bottomless and cost-effective storage architecture that can scale to accommodate Big Data requirements. The proposed system offers access to data through a POSIX filesystem interface under a single global namespace, optimizing the placement of data across disk and tape tiers. Using a self-contained, standardized and open filesystem format on the removable tape media, the proposed system avoids dependence on proprietary software and external metadata servers to access the data stored on tape. By internally managing the tape tier resources, such as tape drives and cartridges, the system relieves the user from the burden of dealing with the complexities of tape storage. Our implementation, which is based on the GPFS and LTFS filesystems, demonstrates the applicability of the proposed architecture in real-world environments. Our experimental evaluation has shown that this is a very promising approach in terms scalability, performance and manageability. The proposed system has been productized by IBM as LTFS Enterprise Edition.

DMA-capable interconnects, providing ultra-low latency and high bandwidth, are increasingly being used in the context of distributed storage and data processing systems. However, the deployment of such systems in virtualized data centers is currently inhibited by the lack of a flexible and high-performance virtualization solution for RDMA network interfaces.

Modern workstations and servers implicitly trust hard disks to act as well-behaved block devices. This paper analyzes the catastrophic loss of security that occurs when hard disks are not trustworthy. First, we show that it is possible to compromise the firmware of a commercial off-the-shelf hard drive, by resorting only to public information and reverse engineering. Using such a compromised firmware, we present a stealth rootkit that replaces arbitrary blocks from the disk while they are written, providing a data replacement back-door. The measured performance overhead of the compromised disk drive is less than 1% compared with a normal, non-malicious disk drive. We then demonstrate that a remote attacker can even establish a communication channel with a compromised disk to infiltrate commands and to ex-filtrate data. In our example, this channel is established over the Internet to an unmodified web server that relies on the compromised drive for its storage, passing through the original webserver, database server, database storage engine, filesystem driver, and block device driver. Additional experiments, performed in an emulated disk-drive environment, could automatically extract sensitive data such as /etc/shadow (or a secret key file) in less than a minute. This paper claims that the difficulty of implementing such an attack is not limited to the area of government cyber-warfare; rather, it is well within the reach of moderately funded criminals, botnet herders and academic researchers.

Flash SSDs are quickly becoming mainstream and emerge as alternatives to magnetic disks. It is therefore imperative to incorporate them seamlessly into the enterprise. We present the salient results of research in the area, touching all aspects of the data management stack: from the fundamentals of flash technology, through storage for database systems and the manipulation of SSD-resident data, to query processing.

Database archiving is important for the retrieval of old versions of a database and for temporal queries over the history of data. We demonstrate XArch, a management system for maintaining, populating, and querying archives of hierarchical data. XArch is based on a nested merge approach that efficiently stores multiple versions of hierarchical data in a compact archive. By merging elements into one data structure, any specific version is retrievable from the archive in a single pass over the data and efficient tracking of object history is possible. XArch implements this approach and extends it in two important ways. First, in order to merge large hierarchical data sets, elements need to be sorted according to their key values. We developed an efficient algorithm for sorting hierarchical data in secondary storage and modified the nested merge algorithm accordingly. Second, we designed and implemented a declarative query language that enables one both to view data from particular versions and to track the history of objects. We demonstrate this using both molecular biology and demographic reference data as examples.

Sorting hierarchical data in external memory is necessary for a wide variety of applications including archiving scientific data and dealing with large XML datasets. The topic of sorting hierarchical data, however, has received little attention from the research community so far. In this paper we focus on sorting arbitrary hierarchical data that far exceed the size of physical memory. We propose HErMeS, an algorithm that generalizes the most widely-used techniques for sorting flat data in external memory. HErMeS efficiently exploits the hierarchical structure to minimize the number of disk accesses and optimize the use of available memory. We extract the theoretical bounds of the algorithm with respect to the structure of the hierarchical dataset. We then show how the algorithm can be used to support efficient archiving. We have conducted an experimental study using several workloads and comparing HErMeS to the state-of-the-art approaches. Our results show that our algorithm ( a ) meets its theoretical expectations, ( b ) allows for scalable database archiving, and ( c ) outperforms the competition by a significant factor. These results, we believe, prove our technique to be a viable and scalable solution to the problem of sorting hierarchical data in external memory.

In the near future, commodity hardware is expected to incorporate both flash and magnetic disks. In this paper we study how the storage layer of a database system can benefit from the presence of both kinds of disk. We propose using the flash and the magnetic disk at the same level of the memory hierarchy and placing a data page to only one of these disks according to the workload of the page. Pages with a read-intensive workload are placed on the flash disk, while pages with a write-intensive workload are placed on the magnetic disk. We present a family of on-line algorithms to decide the optimal placement of a page and study their theoretical properties. Our system is self-tuning, i.e. , our algorithms adapt page placement to changing workloads. We also present a buffer replacement policy that takes advantage of the asymmetric I/O properties of the two types of storage media to reduce the total I/O cost. Our experimental evaluation shows remarkable I/O performance improvement over both flash-only and magnetic-only systems. These results, we believe, exhibit both the potential and necessity of such algorithms in future database systems.