Nisha Talagala

HEC: improving endurance of high performance flash-based cache devices

Flash memory is widely used for its fast random I/O access performance in a gamut of enterprise storage applications. However, due to the limited endurance and asymmetric write performance of flash memory, minimizing writes to a flash device is critical for both performance and endurance. Previous studies have focused on flash memory as a candidate for primary storage devices; little is known about its behavior as a Solid State Cache (SSC) device. In this paper, we propose HEC, a High Endurance Cache that aims to improve overall device endurance via reduced media writes and erases while maximizing cache hit rate performance. We analyze the added write pressures that cache workloads place on flash devices and propose optimizations at both the cache and flash management layers to improve endurance while maintaining or increasing cache hit rate. We demonstrate the individual and cumulative contributions of cache admission policy, cache eviction policy, flash garbage collection policy, and flash device configuration on a) hit rate, b) overall writes, and c) erases as seen by the SSC device. Through our improved cache and flash optimizations, 83% of the analyzed workload ensembles achieved increased or maintained hit rate with write reductions up to 20x, and erase count reductions up to 6x.

Snapshots in a flash with ioSnap

Snapshots are a common and heavily relied upon feature in storage systems. The high performance of flash-based storage systems brings new, more stringent, requirements for this classic capability. We present ioSnap, a flash optimized snapshot system. Through careful design exploiting common snapshot usage patterns and flash oriented optimizations, including leveraging native characteristics of Flash Translation Layers, ioSnap delivers low-overhead snapshots with minimal disruption to foreground traffic. Through our evaluation, we show that ioSnap incurs negligible performance overhead during normal operation, and that common-case operations such as snapshot creation and deletion incur little cost. We also demonstrate techniques to mitigate the performance impact on foreground I/O during intensive snapshot operations such as activation. Overall, ioSnap represents a case study of how to integrate snapshots into a modern, well-engineered flash-based storage system.

Usage patterns of a Web-based image collection

This paper presents a study of user access patterns to a large, Web-based, image collection. The images are the entire collection of the Fine Arts Museums of San Francisco, the largest on-line collection of high resolution art images in the world. The images are served using a tile-based solution that allows a user to zoom-in and navigate within an image. We studied five months of web log data for this collection. Our analysis revealed the following: less than 10% of all available documents on the site were accessed in the five month period and document popularity appears to follow a Zipf distribution. Also, images have interesting areas which are viewed more than others, some image resolutions are viewed far more than others, and user navigation patterns vary between resolutions and are sensitive to download time. The paper discusses these results and their implications for the design of caches and archival storage systems to support this type of workload.

Mjölnir: collecting trash in a demanding new world

As flash devices become ubiquitous in data centers and cost per gigabyte drops, flash systems need to provide data services similar to those of traditional storage. We present Mjölnir, a powerful and scalable engine that addresses the core problems that make efficient flash based data services challenging: multi-reference management and garbage collection. Additionally, by providing powerful primitives for address remapping, Mjölnir enables redesign of the I/O stack for greater efficiency and performance with flash. Mjölnir uses techniques from language runtimes for reference management and garbage collection; we show via prototype and experimental evaluation that this design can deliver predictable performance even with varied user workloads across a range of capacity and reference-count scales.

NVM aware MariaDB database system

MariaDB is a community-developed fork of the MySQL relational database management system and originally designed and implemented to use traditional spinning disk architecture. Now that devices with Non-Volatile memory (NVM) technologies are available, MariaDB addresses this challenge by adding support for NVM devices and introduces NVM Compression method. NVM Compression is a novel hybrid technique that combines application level compression with flash awareness for optimal performance and storage efficiency. Utilizing new interface primitives exported by Flash Translation Layers (FTLs), we leverage the garbage collection available in flash devices to optimize the capacity management required by compression systems. We implement NVM Compression in the popular MariaDB database and use variants of commonly available POSIX file system interfaces to provide the extended FTL capabilities to the user space application. The experimental results show that the hybrid approach of NVM Compression can improve compression performance by 2-7x, deliver compression performance for flash devices that is within 5% of uncompressed performance, improves storage efficiency by 19% over legacy Row-Compression, reduce data writes by up to 4x when combined with other flash aware techniques such as Atomic Writes, and deliver further advantages in power efficiency and CPU utilization.

Towards software defined persistent memory: rethinking software support for heterogenous memory architectures

The emergence of persistent memories promises a sea-change in application and data center architectures, with efficiencies and performance not possible with todays volatile DRAM and persistent slow storage. We present Software Defined Persistent Memory, an approach that enables applications to use persistent memory in a variety of local and remote configurations. The heterogeneity is managed by a middleware that manages hardware specific needs and optimizations. We present the first ever design and implementation of such an architecture, and illustrate the key abstractions that are needed to hide hardware specific details from applications while exposing necessary characteristics for performance optimization. We evaluate the performance of our implementation on a set of microbenchmarks and database workloads using the MySQL database. Through our evaluation, we show that it is possible to apply Software Defined concepts to persistent memory, to improve performance while retaining functionality and optimizing for different hardware architectures.

Failure Characteristics and Soft Error Behavior in a Large Storage System

This chapter analyzes the error behavior of a 3.2TB disk storage system. We report reliability data for 18 months of the prototype’s operation, and analyze 6 months of error logs from nodes in the prototype. We found that the disks drives were among the most reliable components in the system. We were also able to divide errors into eleven categories, comprising disk errors, network errors and SCSI errors that appeared repeatedly across all nodes. We also gained insight into the types of error messages reported by devices in various conditions, and the effects of these events on the operating system. We also present data from four cases of disk drive failures. These results and insights should be useful to any designer of a fault tolerant storage system.