Garbage Collection for Multiversion Index in Flash-Based Embedded Databases

Abstract

Recently, flash-based embedded databases have gained their momentum in various control and monitoring systems, such as cyber-physical systems (CPSes). To support the functionality to access the historical data, a multiversion index is adopted to simultaneously maintain multiple versions of data items, as well as their index information. However, maintaining a multiversion index on flash memory incurs considerable performance overheads on garbage collection, which is to reclaim the spaces occupied by the outdated/invalid data items and their index information on flash memory. In this work, we propose an efficient garbage collection strategy to solve the garbage collection issues of flash-based multiversion databases. In particular, a version-tracking method is proposed to accelerate the performance on the process on identifying/reclaiming the space of invalid data and their indexes, and a pre-summary method is also designed to solve the cascading update problem that is caused by the write-once nature of flash memory and is worsened when more versions refer to the same data item. The capability of the proposed strategy is then verified by analytical and experimental studies.

Lazy-RTGC

“…Many schemes have been proposed to optimize the management of address mapping information [Wu et al 2007;Lee et al 2007;] and to solve endurance and power failure problems [Chang 2007;Guo et al 2013;Chang et al 2013bChang et al , 2015Zhang et al 2014;Huang et al 2014a]. Only few works are about realtime NAND flash memory storage systems.…”

“…A lot of work has been conducted on storage system architecture design Chang et al 2013aHsieh et al 2013Hsieh et al , 2014Huang et al 2013], while others studied the flash translation-layer design Lee et al 2008;Chung et al 2009;Wu and Lin 2012]. Several techniques have been proposed to improve the system performance of NAND flash memory storage systems [Hu et al 2010;Jung et al 2010;Guan et al 2013;Huang et al 2014a]. Different from prior works, our scheme aims to guarantee the worst system response time and, at the same time, optimize the average system performance in NAND flash memory storage systems.…”

Garbage collection and wear leveling for flash memory: Past and future

“…Meanwhile, some recent research concentrated on the maintenance of multiversion data on flash memory, so as to improve the performance and reliability of embedded databases. In particular, based on the well-known excellent multiversion B + -tree (MVBT) [9] for hard disk, several related work suggest efficient approaches to data item/index management and space allocation/reclamation on flash memory [39], [57]. However, there are still some missing parts in the allocation/reclamation/management of multiversion data over flash memory that could be great research topics in the future.…”

A Reliability-Aware Address Mapping Strategy for NAND Flash Memory Storage Systems

“…Most existing approaches aim to minimize either the amount of critical data or/and the number of block erase counts [3], [4]. Reliability is enhanced through the modification of different components in NAND flash architecture, such as file system [5], [6], parallelism of solid-state drive (SSD) [7], [8], flash translation layer (FTL) [9], [10], or the optimization of garbage collection operations [11], [12]. These approaches can provide good solutions in terms of endurance, wear-leveling, memory usage, energy consumption, and response time.…”