Storage Class Memory (SCM) is emerging as a viable alternative to traditional DRAM, alleviating its scalability limits, both in terms of capacity and energy consumption, while being non-volatile. Hence, SCM has the potential to become a universal memory, blurring well-known storage hierarchies. However, along with opportunities, SCM brings many challenges. In this tutorial we will dissect SCM challenges and provide an in-depth view of existing programming models that circumvent them, as well as novel data structures that stem from these models. We will also elaborate on fail-safety testing challenges -- an often overlooked, yet important topic. Finally, we will discuss SCM emulation techniques for end-to-end testing of SCM-based software components. In contrast to surveys investigating the use of SCM in database systems, this tutorial is designed as a programming guide for researchers and professionals interested in leveraging SCM in database systems.

Storage Class Memory (SCM) has the potential to significantly improve database performance. This potential has been well documented for throughput [4] and response time [25, 22]. In this paper we show that SCM has also the potential to significantly improve restart performance, a shortcoming of traditional main memory database systems. We present SOFORT, a hybrid SCM-DRAM storage engine that leverages full capabilities of SCM by doing away with a traditional log and updating the persisted data in place in small increments. We show that we can achieve restart times of a few seconds independent of instance size and transaction volume without significantly impacting transaction throughput.

This thesis presents a method for enhancing performance of Storage Class Memory File System (SCMFS) and an implementation of SCMFS on Android platform. It focuses on analyzing performance influencing factors of memory file systems and the differences in implementation of SCMFS on Android and Linux kernels. SCMFS allocates memory pages as file blocks and employs virtual memory addresses as file block addresses. SCMFS utilizes processor's memory management unit and TLB (Translation Lookaside Buffer) during file accesses. TLB is an expensive resource and has a limited number of entries to cache virtual to physical address translations. TLB miss results in expensive page walks through memory page table. Thus TLB misses play an important role in determining SCMFS performance. In this thesis, SCMFS is designed to support both 4KB and 2MB page sizes in order to reduce TLB misses and to avoid significant internal fragmentation. By comparing SCMFS with YAFFS2 and EXT4 using popular benchmarks, both advantages and disadvantages of SCMFS huge-page version and small-page version are revealed. In the second part of this thesis, an implementation of SCMFS on Android platform is presented. At the time of working on this research project, Android kernel was not merged into Linux kernel yet. Two main changes of SCMFS kernel code: memory zoning and inode functions, are made to be compatible with Android kernel. AndroSH, a file system benchmark for SCMFS on Android, is developed based on shell script. Evaluations are made from three perspectives to compare SCMFS with YAFFS2 and EXT4: I/O throughput, user data access latency, and application execution latency. SCMFS shows a performance advantage because of its small instruction footprint and its pre-allocation mechanism. However, the singly linked list used by SCMFS to store subdirectories is less efficient than HTree index used by EXT4. The future work can improve lookup efficiency of SCMFS.

abstract: As persistent non-volatile memory solutions become integrated in the computing ecosystem and landscape, traditional commodity file systems architected and developed for traditional block I/O based memory solutions must be reevaluated. A majority of commodity file systems have been architected and designed with the goal of managing data on non-volatile storage devices such as hard disk drives (HDDs) and solid state drives (SSDs). HDDs and SSDs are attached to a computing system via a controller or I/O hub, often referred to as the southbridge. The point of HDD and SSD attachment creates multiple levels of translation for any data managed by the CPU that must be stored in non-volatile memory (NVM) on an HDD or SSD. Storage Class Memory (SCM) devices provide the ability to store data at the CPU and DRAM level of a computing system. A novel set of modifications to the ext2 and ext4 commodity file systems to address the needs of SCM will be presented and discussed. An in-depth analysis of many existing file systems, from multiple sources, will be presented along with an analysis to identify key modifications and extensions that would be necessary to execute file system on SCM devices. From this analysis, modifications and extensions have been applied to the FAT commodity file system for key functional tests that will be presented to demonstrate the operation and execution of the file system extensions. / Dissertation/Thesis / Masters Thesis Computer Science 2016

Storage Class Memory (SCM) is a novel class of memory technologies that promise to revolutionize database architectures. SCM is byte-addressable and exhibits latencies similar to those of DRAM, while being non-volatile. Hence, SCM could replace both main memory and storage, enabling a novel single-level database architecture without the traditional I/O bottleneck. Fail-safe persistent SCM allocation can be considered conditio sine qua non for enabling this novel architecture paradigm for database management systems. In this paper we present PAllocator, a fail-safe persistent SCM allocator whose design emphasizes high concurrency and capacity scalability. Contrary to previous works, PAllocator thoroughly addresses the important challenge of persistent memory fragmentation by implementing an efficient defragmentation algorithm. We show that PAllocator outperforms state-of-the-art persistent allocators by up to one order of magnitude, both in operation throughput and recovery time, and enables up to 2.39x higher operation throughput on a persistent B-Tree.

Leveraging Storage Class Memory (SCM) as a universal memory--i.e. as memory and storage at the same time--has deep implications on database architectures. It becomes possible to store a single copy of the data in SCM and directly operate on it at a fine granularity. However, exposing the whole database with direct access to the application dramatically increases the risk of data corruption. In this paper we propose a lightweight on-line testing framework that helps find and debug SCM-related errors that can occur upon software or power failures. Our testing framework simulates failures in critical code paths and achieves fast code coverage by leveraging call stack information to limit duplicate testing. It also partially covers the errors that might arise as a result of reordered memory operations. We show through an experimental evaluation that our testing framework is fast enough to be used with large software systems and discuss its use during the development of our in-house persistent SCM allocator.

Submitted by Caroline Xavier (caroline.xavier@pucrs.br) on 2017-07-03T15:03:25Z No. of bitstreams: 1 TES_TACIANO_DRECKMANN_PEREZ_COMPLETO.pdf: 1779781 bytes, checksum: 2111ccea963be0eea76bdbb7d24321d9 (MD5) / Made available in DSpace on 2017-07-03T15:03:25Z (GMT). No. of bitstreams: 1 TES_TACIANO_DRECKMANN_PEREZ_COMPLETO.pdf: 1779781 bytes, checksum: 2111ccea963be0eea76bdbb7d24321d9 (MD5) Previous issue date: 2017-05-03 / Sistemas computacionais da atualidade tradicionalmente separam mem?ria e armazenamento. Linguagens de programa??o tipicamente refletem essa distin??o usando diferentes representa??es para dados em mem?ria (ex. estruturas de dados, objetos) e armazenamento (ex. arquivos, bancos de dados). A movimenta??o de dados entre esses dois n?veis e representa??es, bidirecionalmente, compromete tanto a efici?ncia do programador quanto de execu??o dos programas. Tecnologias recentes de memoria n?o-vol?til, tais como mem?ria de transi??o de fase, resistiva e magnetoresistiva, possibilitam combinar mem?ria principal e armazenamento em uma ?nica entidade de mem?ria persistente, abrindo caminho para abstra??es mais eficientes para lidar com persist?ncia de dados. Essa tese de doutorado introduz uma abordagem de projeto para o ambiente de execu??o de linguagens com ger?ncia autom?tica de mem?ria, baseado numa combina??o original de persist?ncia ortogonal, programa??o para mem?ria persistente, persist?ncia por alcance, e transa??es com atomicidade em caso de falha. Esta abordagem pode melhorar significativamente a produtividade do programador e a efici?ncia de execu??o dos programas, uma vez que estruturas de dados em mem?ria passam a ser persistentes de forma transparente, sem a necessidade de programar explicitamente o armazenamento, e removendo a necessidade de cruzar fronteiras sem?nticas. De forma a validar e demonstrar a abordagem proposta, esse trabalho tamb?m apresenta JaphaVM, a primeira M?quina Virtual Java especificamente projetada para mem?ria persistente. Resultados experimentais usando benchmarks e aplica??es reais demonstram que a JaphaVM, na maioria dos casos, executa as mesmas opera??es cerca de uma a duas ordens de magnitude mais rapidamente do que implementa??es equivalentes usando bancos de dados ou arquivos, e, ao mesmo tempo, requer significativamente menos linhas de c?digo. / Current computer systems separate main memory from storage. Programming languages typically reflect this distinction using different representations for data in memory (e.g. data structures, objects) and storage (e.g. files, databases). Moving data back and forth between these different layers and representations compromise both programming and execution efficiency. Recent nonvolatile memory technologies, such as Phase-Change Memory, Resistive RAM, and Magnetoresistive RAM make it possible to collapse main memory and storage into a single layer of persistent memory, opening the way for simpler and more efficient programming abstractions for handling persistence. This Ph.D. thesis introduces a design for the runtime environment for languages with automatic memory management, based on an original combination of orthogonal persistence, persistent memory programming, persistence by reachability, and lock-based failure-atomic transactions. Such design can significantly increase programming and execution efficiency, as in-memory data structures are transparently persistent, without the need for programmatic persistence handling, and removing the need for crossing semantic boundaries. In order to validate and demonstrate the proposed concepts, this work also presents JaphaVM, the first Java Virtual Machine specifically designed for persistent memory. In experimental results using benchmarks and real-world applications, JaphaVM in most cases executed the same operations between one and two orders of magnitude faster than database- and file-based implementations, while requiring significantly less lines of code.

The advent of Storage Class Memory (SCM) is driving a rethink of storage systems towards a single-level architecture where memory and storage are merged. In this context, several works have investigated how to design persistent trees in SCM as a fundamental building block for these novel systems. However, these trees are significantly slower than DRAM-based counterparts since trees are latency-sensitive and SCM exhibits higher latencies than DRAM. In this paper we propose a novel hybrid SCM-DRAM persistent and concurrent B-Tree, named Fingerprinting Persistent Tree (FPTree) that achieves similar performance to DRAM-based counterparts. In this novel design, leaf nodes are persisted in SCM while inner nodes are placed in DRAM and rebuilt upon recovery. The FPTree uses Fingerprinting, a technique that limits the expected number of in-leaf probed keys to one. In addition, we propose a hybrid concurrency scheme for the FPTree that is partially based on Hardware Transactional Memory. We conduct a thorough performance evaluation and show that the FPTree outperforms state-of-the-art persistent trees with different SCM latencies by up to a factor of 8.2. Moreover, we show that the FPTree scales very well on a machine with 88 logical cores. Finally, we integrate the evaluated trees in memcached and a prototype database. We show that the FPTree incurs an almost negligible performance overhead over using fully transient data structures, while significantly outperforming other persistent trees.