SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery

Size: px
Start display at page:

Download "SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery"

Transcription

1 SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery Ismail Oukid*, Daniel Booss, Wolfgang Lehner*, Peter Bumbulis, and Thomas Willhalm + *Dresden University of Technology SAP AG + Intel GmbH DaMoN 2014, Snowbird, Utah, USA, June 23, 2014 Prof. Dr.-Ing. Wolfgang Lehner Diese Zeile ersetzt man über: Einfügen > Kopf- und

2 What is Storage Class Memory? Non-Volatile Memory Byteaddressable (load/store semantics) Latency close to DRAM s, with writes slower than reads More energy efficient than DRAM Storage Class Memory Denser than DRAM (i.e. better scaling) e.g. PCM, STTRAM, MRAM and Memristors Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 2

3 SCM Compared to Other Technologies Latency Bandwidth Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 3

4 SCM Compared to Other Technologies Latency HDD Bandwidth Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 4

5 SCM Compared to Other Technologies Latency HDD DRAM Bandwidth Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 5

6 SCM Compared to Other Technologies Latency HDD Bridging the gap between memory and storage DRAM Bandwidth Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 6

7 SCM Compared to Other Technologies Latency HDD SSD Bridging the gap between memory and storage DRAM Bandwidth Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 7

8 SCM Compared to Other Technologies Latency HDD SSD Bridging the gap between memory and storage SCM DRAM Bandwidth Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 8

9 SCM Compared to Other Technologies Latency HDD SSD Bridging the gap between memory and storage SCM DRAM Bandwidth Storage Class Memory is a merging point between memory and storage Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 9

10 Main Memory DB Restart Time Availability guarantees are an important part of many SLAs Availability of 99% downtime of 3.65 days/year! Many database crash conditions are transient A reasonable approach to recovery: restarting the database Database restart time impacts directly database availability Long restart times: a shortcoming of main memory databases Improve database restart time using SCM Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 10

11 Overview of SOFORT (1) SOFORT: A hybrid SCM-DRAM storage engine for fast data recovery Requirements: Fast data recovery Mixed OLTP and OLAP A hybrid SCM-DRAM memory environment Architecture: Single-level column-store Dictionary encoded Multi version concurrency control (MVCC) No transactional log Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 11

12 Overview of SOFORT (2) SOFORT Table MVCC Entry CTS DTS Appendonly TX Array Tables MVCC Array Columns Column On DRAM to enable better performance Dictionary Encoded Persisted on SCM Volatile in DRAM Dict. Index Dict. Index (optional) Value IDs Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 12

13 Persisting Data on the Fly SCM accessed via the CPU cache: Writes not guaranteed to have reached SCM Need to flush data from cache Code reordered at compilation time & Complex CPU out-of-order execution Need to order memory operations Persistence primitives: flushing instructions (CLFLUSH), memory barriers (MFENCE, SFENCE, LFENCE) and non-temporal stores () Data may still be held in memory controller buffers Need to flush memory controller buffers on power failures Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 13

14 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; Flush var; persisted = true; Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 14

15 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; Flush var; persisted = true; Cache SCM var 0 0 persisted False False Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 15

16 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; Flush var; persisted = true; Cache SCM var 0 0 persisted True True Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 16

17 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; Flush var; persisted = true; Cache SCM var 1 0 persisted True True Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 17

18 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; Flush var; MFENCE; persisted = true; Cache SCM var 0 0 persisted False False Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 18

19 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; Flush var; MFENCE; persisted = true; Cache SCM var 0 0 persisted False False Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 19

20 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; MFENCE; Flush var; MFENCE; persisted = true; Cache SCM var 0 0 persisted False False Flush persisted; Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 20

21 Persisting Data on the Fly: Example Int var = 0; Bool persisted = false; var = 1; MFENCE; Flush var; MFENCE; persisted = true; MFENCE; Flush persisted; MFENCE; Cache SCM var 0 0 persisted False False Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 21

22 Persistent Memory Allocator Persistent Memory File System (PMFS): SCM-aware file system No buffering in DRAM on mmap direct access to SCM PMAllocator: Huge PMFS files as memory pages Pages cut into segments for allocation Persistent counter of memory pages Mapping from persistent memory to VM PMFS file = Memory page Segment File name = Unique page ID Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 22

23 Persistent Memory Pointers Regular pointers are bound to the program s address space Cannot be used for recovery We propose Persistent Memory Pointers (PMPtrs): Struct PMPtr { int64_t m_baseid; ptrdiff_t m_offset; } Persistent memory page ID Offset indicating the start of the allocated block Can be converted (swizzled) to regular pointers PMPtrs stay valid across failures Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 23

24 Recovery Mechanism PM Entry Point Reload PMAllocator Reload SOFORT A PMPtr to the PMAllocator A PMPtr to the storage engine Reload memory pages Reconstruct mapping to VM Rollback unfinished transactions Allocate ressources for recovery Reload Table M Reload Table 1 Data structures sanity check Initiate columns recovery Reload column N Reload column 1 Data structures sanity check Reconstruct dictionary indexes Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 24

25 Evaluation: Setup Hardware-based SCM simulation: Intel Xeon 20MB L3 cache, 8 physical cores/socket Avoiding NUMA effects: benchmark run on a single socket Limitation: symmetric instead of asymmetric read/write latency Socket 0 Socket 1 QPI Link CPU (8 cores) CPU (8 cores) Memory Bus Memory Bus DRAM SCM (PMFS mount) DRAM Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 25

26 Evaluation: Setup TATP benchmark: Simple but realistic OLTP benchmark (20% write, 80% reads) Simulates a telecommunication application Initial population of 1 million subscribers (~2GB data size) Shore-MT: A storage manager designed for scalability in multicores Run on Ramdisk (tempfs) Restart time: Single, 4 column table SCM latency set to 200ns Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 26

27 Mio. Tx/s Evaluation: Throughput 2 1,8 1,6 GetSubData Throughput (Read Only) Realistic case Upper bound 1,4 1,2 shore-read-shm 1 0,8 0,6 0,4 0,2 User contention. Solved with PLP. Worst case sofort-read-shm sofort-read-200ns sofort-read-300ns sofort-read-700ns # users Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 27

28 Mio. Tx/s Evaluation: Throughput 1,2 1 0,8 0,6 0,4 0,2 TATP Throughput (80% Read, 20% Write) ~900K Tx/s ~550K Tx/s ~250K Tx/s Upper bound Worst case Realistic case shore-shm sofort-shm sofort-200ns sofort-300ns sofort-700ns # users SOFORT outperforms Shore-MT even in a high latency SCM environment Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 28

29 Mio Tx/s Mio Tx/s Restart Time (s) Restart Time (s) Evaluation: Restart Time Dict. Size Fixed Mio. Rows/Table SOFORT on SCM ~2sec Time (s) Restart time is independent of data size time-dbfixedhashmap time-dictfixedskiplist time-dictfixedhashmap SOFORT recovers nearly instantly Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June Data Size Fixed Mio. Distinct Values Restart time is independent of instance size and depends only on dictionaries size SOFORT on DISK ~18sec Time (s) Volatile structures drive restart time time-dbfixedskiplist No persistence guarantees

30 Conclusion and Future Work We showed that SCM can help getting rid of the transactional log and significantly improving database restart times Latency study based on hardware simulation showed that SOFORT exhibits competitive performance even in a high latency SCM environment Future Work: Extend SOFORT to support long transactions Explore new recovery schemes Design persistent index structures Experiment with write-through caching policy Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery DaMoN 14, June 23 30

31 Takeaway: SCM is a game changer in the database industry Make the change happen! Thank You! Questions? Comments? Ismail Oukid SOFORT: A Hybrid SCM-DRAM Storage Engine for Fast Data Recovery Ismail Oukid, Daniel Booss, Wolfgang Lehner, Peter Bumbulis, and Thomas Willhalm + DaMoN 2014, Snowbird, Utah, USA, June 23, 2014 Prof. Dr.-Ing. Wolfgang Lehner Diese Zeile ersetzt man über: Einfügen > Kopf- und

Instant Recovery for Main-Memory Databases

Instant Recovery for Main-Memory Databases Persistent Storage Non-Volatile Main Memory Transient Main Memory Transient Main Memory runtime data moving the persistency bar Instant Recovery for Main-Memory Databases Ismail Oukid Technische Universität

More information

A Study of Application Performance with Non-Volatile Main Memory

A Study of Application Performance with Non-Volatile Main Memory A Study of Application Performance with Non-Volatile Main Memory Yiying Zhang, Steven Swanson 2 Memory Storage Fast Slow Volatile In bytes Persistent In blocks Next-Generation Non-Volatile Memory (NVM)

More information

NVRAM-aware Logging in Transaction Systems. Jian Huang

NVRAM-aware Logging in Transaction Systems. Jian Huang NVRAM-aware Logging in Transaction Systems Jian Huang Karsten Schwan Moinuddin K. Qureshi Logging Support for Transactions 2 Logging Support for Transactions ARIES: Disk-Based Approach (TODS 92) Write-ahead

More information

Computer Engineering and Systems Group Electrical and Computer Engineering SCMFS: A File System for Storage Class Memory

Computer Engineering and Systems Group Electrical and Computer Engineering SCMFS: A File System for Storage Class Memory SCMFS: A File System for Storage Class Memory Xiaojian Wu, Narasimha Reddy Texas A&M University What is SCM? Storage Class Memory Byte-addressable, like DRAM Non-volatile, persistent storage Example: Phase

More information

NVRAM-aware Logging in Transaction Systems

NVRAM-aware Logging in Transaction Systems NVRAM-aware Logging in Transaction Systems Jian Huang jhuang95@cc.gatech.edu Karsten Schwan schwan@cc.gatech.edu Moinuddin K. Qureshi moin@ece.gatech.edu Georgia Institute of Technology ABSTRACT Emerging

More information

DIABLO TECHNOLOGIES MEMORY CHANNEL STORAGE AND VMWARE VIRTUAL SAN : VDI ACCELERATION

DIABLO TECHNOLOGIES MEMORY CHANNEL STORAGE AND VMWARE VIRTUAL SAN : VDI ACCELERATION DIABLO TECHNOLOGIES MEMORY CHANNEL STORAGE AND VMWARE VIRTUAL SAN : VDI ACCELERATION A DIABLO WHITE PAPER AUGUST 2014 Ricky Trigalo Director of Business Development Virtualization, Diablo Technologies

More information

NV-DIMM: Fastest Tier in Your Storage Strategy

NV-DIMM: Fastest Tier in Your Storage Strategy NV-DIMM: Fastest Tier in Your Storage Strategy Introducing ArxCis-NV, a Non-Volatile DIMM Author: Adrian Proctor, Viking Technology [email: adrian.proctor@vikingtechnology.com] This paper reviews how Non-Volatile

More information

Tips and Tricks for Using Oracle TimesTen In-Memory Database in the Application Tier

Tips and Tricks for Using Oracle TimesTen In-Memory Database in the Application Tier Tips and Tricks for Using Oracle TimesTen In-Memory Database in the Application Tier Simon Law TimesTen Product Manager, Oracle Meet The Experts: Andy Yao TimesTen Product Manager, Oracle Gagan Singh Senior

More information

SAP HANA - Main Memory Technology: A Challenge for Development of Business Applications. Jürgen Primsch, SAP AG July 2011

SAP HANA - Main Memory Technology: A Challenge for Development of Business Applications. Jürgen Primsch, SAP AG July 2011 SAP HANA - Main Memory Technology: A Challenge for Development of Business Applications Jürgen Primsch, SAP AG July 2011 Why In-Memory? Information at the Speed of Thought Imagine access to business data,

More information

Data Growth. Von 2000 bis 2002 sind mehr Daten generiert worden als in den 40.000 Jahren davor

Data Growth. Von 2000 bis 2002 sind mehr Daten generiert worden als in den 40.000 Jahren davor Prof. Dr.-Ing. Wolfgang Lehner 200. Datenbankstammtisch Data Growth One Minute of Internet Von 2000 bis 2002 sind mehr Daten generiert worden als in den 40.000 Jahren davor Von 2004-2005 hat sich diese

More information

Parallel Processing and Software Performance. Lukáš Marek

Parallel Processing and Software Performance. Lukáš Marek Parallel Processing and Software Performance Lukáš Marek DISTRIBUTED SYSTEMS RESEARCH GROUP http://dsrg.mff.cuni.cz CHARLES UNIVERSITY PRAGUE Faculty of Mathematics and Physics Benchmarking in parallel

More information

nvm malloc: Memory Allocation for NVRAM

nvm malloc: Memory Allocation for NVRAM nvm malloc: Memory Allocation for NVRAM David Schwalb, Tim Berning, Martin Faust, Markus Dreseler, Hasso Plattner Hasso-Plattner-Institute, Germany SAP SE, Walldorf, Germany Contact: david.schwalb@hpi.de

More information

Intel Xeon Processor 5560 (Nehalem EP)

Intel Xeon Processor 5560 (Nehalem EP) SAP NetWeaver Mobile 7.1 Intel Xeon Processor 5560 (Nehalem EP) Prove performance to synchronize 10,000 devices in ~60 mins Intel SAP NetWeaver Solution Management Intel + SAP Success comes from maintaining

More information

Accelerating Server Storage Performance on Lenovo ThinkServer

Accelerating Server Storage Performance on Lenovo ThinkServer Accelerating Server Storage Performance on Lenovo ThinkServer Lenovo Enterprise Product Group April 214 Copyright Lenovo 214 LENOVO PROVIDES THIS PUBLICATION AS IS WITHOUT WARRANTY OF ANY KIND, EITHER

More information

A client side persistent block cache for the data center. Vault Boston 2015 - Luis Pabón - Red Hat

A client side persistent block cache for the data center. Vault Boston 2015 - Luis Pabón - Red Hat PBLCACHE A client side persistent block cache for the data center Vault Boston 2015 - Luis Pabón - Red Hat ABOUT ME LUIS PABÓN Principal Software Engineer, Red Hat Storage IRC, GitHub: lpabon QUESTIONS:

More information

Why Hybrid Storage Strategies Give the Best Bang for the Buck

Why Hybrid Storage Strategies Give the Best Bang for the Buck JANUARY 28, 2014, SAN JOSE, CA Tom Coughlin, Coughlin Associates & Jim Handy, Objective Analysis PRESENTATION TITLE GOES HERE Why Hybrid Storage Strategies Give the Best Bang for the Buck 1 Outline Different

More information

PSAM, NEC PCIe SSD Appliance for Microsoft SQL Server (Reference Architecture) September 11 th, 2014 NEC Corporation

PSAM, NEC PCIe SSD Appliance for Microsoft SQL Server (Reference Architecture) September 11 th, 2014 NEC Corporation PSAM, NEC PCIe SSD Appliance for Microsoft SQL Server (Reference Architecture) September 11 th, 2014 NEC Corporation 1. Overview of NEC PCIe SSD Appliance for Microsoft SQL Server Page 2 NEC Corporation

More information

Blurred Persistence in Transactional Persistent Memory

Blurred Persistence in Transactional Persistent Memory Blurred Persistence in Transactional Youyou Lu, Jiwu Shu, Long Sun Department of Computer Science and Technology, Tsinghua University, Beijing, China luyouyou@tsinghua.edu.cn, shujw@tsinghua.edu.cn, sun-l12@mails.tsinghua.edu.cn

More information

Emerging storage and HPC technologies to accelerate big data analytics Jerome Gaysse JG Consulting

Emerging storage and HPC technologies to accelerate big data analytics Jerome Gaysse JG Consulting Emerging storage and HPC technologies to accelerate big data analytics Jerome Gaysse JG Consulting Introduction Big Data Analytics needs: Low latency data access Fast computing Power efficiency Latest

More information

In-Memory Data Management for Enterprise Applications

In-Memory Data Management for Enterprise Applications In-Memory Data Management for Enterprise Applications Jens Krueger Senior Researcher and Chair Representative Research Group of Prof. Hasso Plattner Hasso Plattner Institute for Software Engineering University

More information

Implications of Storage Class Memories (SCM) on Software Architectures

Implications of Storage Class Memories (SCM) on Software Architectures Implications of Storage Class Memories (SCM) on Software Architectures C. Mohan, IBM Almaden Research Center, San Jose mohan@almaden.ibm.com http://www.almaden.ibm.com/u/mohan Suparna Bhattacharya, IBM

More information

Non-Volatile Memory. Non-Volatile Memory & its use in Enterprise Applications. Contents

Non-Volatile Memory. Non-Volatile Memory & its use in Enterprise Applications. Contents Non-Volatile Memory Non-Volatile Memory & its use in Enterprise Applications Author: Adrian Proctor, Viking Technology [email: adrian.proctor@vikingtechnology.com] This paper reviews different memory technologies,

More information

Caching Mechanisms for Mobile and IOT Devices

Caching Mechanisms for Mobile and IOT Devices Caching Mechanisms for Mobile and IOT Devices Masafumi Takahashi Toshiba Corporation JEDEC Mobile & IOT Technology Forum Copyright 2016 Toshiba Corporation Background. Unified concept. Outline The first

More information

Non-Volatile Memory and Its Use in Enterprise Applications

Non-Volatile Memory and Its Use in Enterprise Applications Non-Volatile Memory and Its Use in Enterprise Applications Contributor: Viking Technology January 2014 About the SNIA The Storage Networking Industry Association (SNIA) is a not for profit global organization,

More information

Benchmarking Cassandra on Violin

Benchmarking Cassandra on Violin Technical White Paper Report Technical Report Benchmarking Cassandra on Violin Accelerating Cassandra Performance and Reducing Read Latency With Violin Memory Flash-based Storage Arrays Version 1.0 Abstract

More information

Let s Talk About Storage & Recovery Methods for Non-Volatile Memory Database Systems

Let s Talk About Storage & Recovery Methods for Non-Volatile Memory Database Systems Let s Talk About Storage & Recovery Methods for Non-Volatile Database Systems Joy Arulraj jarulraj@cs.cmu.edu Andrew Pavlo pavlo@cs.cmu.edu Subramanya R. Dulloor subramanya.r.dulloor@intel.com Carnegie

More information

IS IN-MEMORY COMPUTING MAKING THE MOVE TO PRIME TIME?

IS IN-MEMORY COMPUTING MAKING THE MOVE TO PRIME TIME? IS IN-MEMORY COMPUTING MAKING THE MOVE TO PRIME TIME? EMC and Intel work with multiple in-memory solutions to make your databases fly Thanks to cheaper random access memory (RAM) and improved technology,

More information

Speed and Persistence for Real-Time Transactions

Speed and Persistence for Real-Time Transactions Speed and Persistence for Real-Time Transactions by TimesTen and Solid Data Systems July 2002 Table of Contents Abstract 1 Who Needs Speed and Persistence 2 The Reference Architecture 3 Benchmark Results

More information

Scaling Database Performance in Azure

Scaling Database Performance in Azure Scaling Database Performance in Azure Results of Microsoft-funded Testing Q1 2015 2015 2014 ScaleArc. All Rights Reserved. 1 Test Goals and Background Info Test Goals and Setup Test goals Microsoft commissioned

More information

HP ProLiant BL660c Gen9 and Microsoft SQL Server 2014 technical brief

HP ProLiant BL660c Gen9 and Microsoft SQL Server 2014 technical brief Technical white paper HP ProLiant BL660c Gen9 and Microsoft SQL Server 2014 technical brief Scale-up your Microsoft SQL Server environment to new heights Table of contents Executive summary... 2 Introduction...

More information

Distributed Architecture of Oracle Database In-memory

Distributed Architecture of Oracle Database In-memory Distributed Architecture of Oracle Database In-memory Niloy Mukherjee, Shasank Chavan, Maria Colgan, Dinesh Das, Mike Gleeson, Sanket Hase, Allison Holloway, Hui Jin, Jesse Kamp, Kartik Kulkarni, Tirthankar

More information

Boosting Database Batch workloads using Flash Memory SSDs

Boosting Database Batch workloads using Flash Memory SSDs Boosting Database Batch workloads using Flash Memory SSDs Won-Gill Oh and Sang-Won Lee School of Information and Communication Engineering SungKyunKwan University, 27334 2066, Seobu-Ro, Jangan-Gu, Suwon-Si,

More information

Rethinking SIMD Vectorization for In-Memory Databases

Rethinking SIMD Vectorization for In-Memory Databases SIGMOD 215, Melbourne, Victoria, Australia Rethinking SIMD Vectorization for In-Memory Databases Orestis Polychroniou Columbia University Arun Raghavan Oracle Labs Kenneth A. Ross Columbia University Latest

More information

bla bla OPEN-XCHANGE Open-Xchange Hardware Needs

bla bla OPEN-XCHANGE Open-Xchange Hardware Needs bla bla OPEN-XCHANGE Open-Xchange Hardware Needs OPEN-XCHANGE: Open-Xchange Hardware Needs Publication date Wednesday, 8 January version. . Hardware Needs with Open-Xchange.. Overview The purpose of this

More information

Assessing the Performance of OpenMP Programs on the Intel Xeon Phi

Assessing the Performance of OpenMP Programs on the Intel Xeon Phi Assessing the Performance of OpenMP Programs on the Intel Xeon Phi Dirk Schmidl, Tim Cramer, Sandra Wienke, Christian Terboven, and Matthias S. Müller schmidl@rz.rwth-aachen.de Rechen- und Kommunikationszentrum

More information

NV-Heaps: Making Persistent Objects Fast and Safe with Next-Generation, Non-Volatile Memories Joel Coburn

NV-Heaps: Making Persistent Objects Fast and Safe with Next-Generation, Non-Volatile Memories Joel Coburn NV-Heaps: Making Persistent Objects Fast and Safe with Next-Generation, Non-Volatile Memories Joel Coburn Work done at UCSD with Adrian M. Caulfield, Ameen Akel, Laura M. Grupp, Rajesh K. Gupta, Ranjit

More information

<Insert Picture Here> Oracle In-Memory Database Cache Overview

<Insert Picture Here> Oracle In-Memory Database Cache Overview Oracle In-Memory Database Cache Overview Simon Law Product Manager The following is intended to outline our general product direction. It is intended for information purposes only,

More information

Parallel Programming Survey

Parallel Programming Survey Christian Terboven 02.09.2014 / Aachen, Germany Stand: 26.08.2014 Version 2.3 IT Center der RWTH Aachen University Agenda Overview: Processor Microarchitecture Shared-Memory

More information

VERITAS Database Edition 2.1.2 for Oracle on HP-UX 11i. Performance Report

VERITAS Database Edition 2.1.2 for Oracle on HP-UX 11i. Performance Report VERITAS Database Edition 2.1.2 for Oracle on HP-UX 11i Performance Report V E R I T A S W H I T E P A P E R Table of Contents Introduction.................................................................................1

More information

The Methodology Behind the Dell SQL Server Advisor Tool

The Methodology Behind the Dell SQL Server Advisor Tool The Methodology Behind the Dell SQL Server Advisor Tool Database Solutions Engineering By Phani MV Dell Product Group October 2009 Executive Summary The Dell SQL Server Advisor is intended to perform capacity

More information

Multi-Threading Performance on Commodity Multi-Core Processors

Multi-Threading Performance on Commodity Multi-Core Processors Multi-Threading Performance on Commodity Multi-Core Processors Jie Chen and William Watson III Scientific Computing Group Jefferson Lab 12000 Jefferson Ave. Newport News, VA 23606 Organization Introduction

More information

In-memory database systems, NVDIMMs and data durability

In-memory database systems, NVDIMMs and data durability In-memory database systems, NVDIMMs and data durability Steve Graves - July 23, 2014 Database management system (DBMS) software is increasingly common in electronics, spurred by growing data management

More information

Measuring Cache and Memory Latency and CPU to Memory Bandwidth

Measuring Cache and Memory Latency and CPU to Memory Bandwidth White Paper Joshua Ruggiero Computer Systems Engineer Intel Corporation Measuring Cache and Memory Latency and CPU to Memory Bandwidth For use with Intel Architecture December 2008 1 321074 Executive Summary

More information

ThyNVM: Enabling Software-Transparent Crash Consistency in Persistent Memory Systems

ThyNVM: Enabling Software-Transparent Crash Consistency in Persistent Memory Systems ThyNVM: Enabling Software-Transparent Crash Consistency in Persistent Memory Systems Jinglei Ren Jishen Zhao Samira Khan Jongmoo Choi + Yongwei Wu Onur Mutlu Carnegie Mellon University Tsinghua University

More information

WITH A FUSION POWERED SQL SERVER 2014 IN-MEMORY OLTP DATABASE

WITH A FUSION POWERED SQL SERVER 2014 IN-MEMORY OLTP DATABASE WITH A FUSION POWERED SQL SERVER 2014 IN-MEMORY OLTP DATABASE 1 W W W. F U S I ON I O.COM Table of Contents Table of Contents... 2 Executive Summary... 3 Introduction: In-Memory Meets iomemory... 4 What

More information

Data Warehousing & Data Mining

Data Warehousing & Data Mining Data Warehousing & Data Mining Wolf-Tilo Balke Silviu Homoceanu Institut für Informationssysteme Technische Universität Braunschweig http://www.ifis.cs.tu-bs.de 8. Real-Time DW 8. Real-Time Data Warehouses

More information

C. Mohan, IBM Almaden Research Center, San Jose, CA 95120 mohan@almaden.ibm.com http://www.almaden.ibm.com/u/mohan

C. Mohan, IBM Almaden Research Center, San Jose, CA 95120 mohan@almaden.ibm.com http://www.almaden.ibm.com/u/mohan Implications of Storage Class Memories (SCM) on Software Architectures C. Mohan, IBM Almaden Research Center, San Jose, CA 95120 mohan@almaden.ibm.com http://www.almaden.ibm.com/u/mohan New England Database

More information

Virtuoso and Database Scalability

Virtuoso and Database Scalability Virtuoso and Database Scalability By Orri Erling Table of Contents Abstract Metrics Results Transaction Throughput Initializing 40 warehouses Serial Read Test Conditions Analysis Working Set Effect of

More information

Accelerate SQL Server 2014 AlwaysOn Availability Groups with Seagate. Nytro Flash Accelerator Cards

Accelerate SQL Server 2014 AlwaysOn Availability Groups with Seagate. Nytro Flash Accelerator Cards Accelerate SQL Server 2014 AlwaysOn Availability Groups with Seagate Nytro Flash Accelerator Cards Technology Paper Authored by: Mark Pokorny, Database Engineer, Seagate Overview SQL Server 2014 provides

More information

Oracle Database Reliability, Performance and scalability on Intel Xeon platforms Mitch Shults, Intel Corporation October 2011

Oracle Database Reliability, Performance and scalability on Intel Xeon platforms Mitch Shults, Intel Corporation October 2011 Oracle Database Reliability, Performance and scalability on Intel platforms Mitch Shults, Intel Corporation October 2011 1 Intel Processor E7-8800/4800/2800 Product Families Up to 10 s and 20 Threads 30MB

More information

System Requirements Table of contents

System Requirements Table of contents Table of contents 1 Introduction... 2 2 Knoa Agent... 2 2.1 System Requirements...2 2.2 Environment Requirements...4 3 Knoa Server Architecture...4 3.1 Knoa Server Components... 4 3.2 Server Hardware Setup...5

More information

The Data Placement Challenge

The Data Placement Challenge The Data Placement Challenge Entire Dataset Applications Active Data Lowest $/IOP Highest throughput Lowest latency 10-20% Right Place Right Cost Right Time 100% 2 2 What s Driving the AST Discussion?

More information

COS 318: Operating Systems

COS 318: Operating Systems COS 318: Operating Systems File Performance and Reliability Andy Bavier Computer Science Department Princeton University http://www.cs.princeton.edu/courses/archive/fall10/cos318/ Topics File buffer cache

More information

Diablo and VMware TM powering SQL Server TM in Virtual SAN TM. A Diablo Technologies Whitepaper. May 2015

Diablo and VMware TM powering SQL Server TM in Virtual SAN TM. A Diablo Technologies Whitepaper. May 2015 A Diablo Technologies Whitepaper Diablo and VMware TM powering SQL Server TM in Virtual SAN TM May 2015 Ricky Trigalo, Director for Virtualization Solutions Architecture, Diablo Technologies Daniel Beveridge,

More information

In-Memory Databases Algorithms and Data Structures on Modern Hardware. Martin Faust David Schwalb Jens Krüger Jürgen Müller

In-Memory Databases Algorithms and Data Structures on Modern Hardware. Martin Faust David Schwalb Jens Krüger Jürgen Müller In-Memory Databases Algorithms and Data Structures on Modern Hardware Martin Faust David Schwalb Jens Krüger Jürgen Müller The Free Lunch Is Over 2 Number of transistors per CPU increases Clock frequency

More information

An Approach for Hybrid-Memory Scaling Columnar In-Memory Databases

An Approach for Hybrid-Memory Scaling Columnar In-Memory Databases An Approach for Hybrid-Memory Scaling Columnar In-Memory Databases Bernhard Höppner SAP SE Potsdam, Germany bernhard.hoeppner@ sap.com Ahmadshah Waizy Fujitsu Technology Solutions GmbH Paderborn, Germany

More information

RAID 5 rebuild performance in ProLiant

RAID 5 rebuild performance in ProLiant RAID 5 rebuild performance in ProLiant technology brief Abstract... 2 Overview of the RAID 5 rebuild process... 2 Estimating the mean-time-to-failure (MTTF)... 3 Factors affecting RAID 5 array rebuild

More information

RAMCloud and the Low- Latency Datacenter. John Ousterhout Stanford University

RAMCloud and the Low- Latency Datacenter. John Ousterhout Stanford University RAMCloud and the Low- Latency Datacenter John Ousterhout Stanford University Most important driver for innovation in computer systems: Rise of the datacenter Phase 1: large scale Phase 2: low latency Introduction

More information

The SAP HANA Database An Architecture Overview

The SAP HANA Database An Architecture Overview The SAP HANA Database An Architecture Overview Franz Färber and Norman May and Wolfgang Lehner and Philipp Große and Ingo Müller and Hannes Rauhe and Jonathan Dees Abstract Requirements of enterprise applications

More information

An Overview of Flash Storage for Databases

An Overview of Flash Storage for Databases An Overview of Flash Storage for Databases Vadim Tkachenko Morgan Tocker http://percona.com MySQL CE Apr 2010 -2- Introduction Vadim Tkachenko Percona Inc, CTO and Lead of Development Morgan Tocker Percona

More information

Solid State Drive Architecture

Solid State Drive Architecture Solid State Drive Architecture A comparison and evaluation of data storage mediums Tyler Thierolf Justin Uriarte Outline Introduction Storage Device as Limiting Factor Terminology Internals Interface Architecture

More information

Leveraging EMC Fully Automated Storage Tiering (FAST) and FAST Cache for SQL Server Enterprise Deployments

Leveraging EMC Fully Automated Storage Tiering (FAST) and FAST Cache for SQL Server Enterprise Deployments Leveraging EMC Fully Automated Storage Tiering (FAST) and FAST Cache for SQL Server Enterprise Deployments Applied Technology Abstract This white paper introduces EMC s latest groundbreaking technologies,

More information

Bigdata High Availability (HA) Architecture

Bigdata High Availability (HA) Architecture Bigdata High Availability (HA) Architecture Introduction This whitepaper describes an HA architecture based on a shared nothing design. Each node uses commodity hardware and has its own local resources

More information

Flash for Databases. September 22, 2015 Peter Zaitsev Percona

Flash for Databases. September 22, 2015 Peter Zaitsev Percona Flash for Databases September 22, 2015 Peter Zaitsev Percona In this Presentation Flash technology overview Review some of the available technology What does this mean for databases? Specific opportunities

More information

Accelerating Enterprise Applications and Reducing TCO with SanDisk ZetaScale Software

Accelerating Enterprise Applications and Reducing TCO with SanDisk ZetaScale Software WHITEPAPER Accelerating Enterprise Applications and Reducing TCO with SanDisk ZetaScale Software SanDisk ZetaScale software unlocks the full benefits of flash for In-Memory Compute and NoSQL applications

More information

In-Memory Databases MemSQL

In-Memory Databases MemSQL IT4BI - Université Libre de Bruxelles In-Memory Databases MemSQL Gabby Nikolova Thao Ha Contents I. In-memory Databases...4 1. Concept:...4 2. Indexing:...4 a. b. c. d. AVL Tree:...4 B-Tree and B+ Tree:...5

More information

A High-Throughput In-Memory Index, Durable on Flash-based SSD

A High-Throughput In-Memory Index, Durable on Flash-based SSD A High-Throughput In-Memory Index, Durable on Flash-based SSD Insights into the Winning Solution of the SIGMOD Programming Contest 2011 Thomas Kissinger, Benjamin Schlegel, Matthias Boehm, Dirk Habich,

More information

Microsoft SQL Server 2014 Fast Track

Microsoft SQL Server 2014 Fast Track Microsoft SQL Server 2014 Fast Track 34-TB Certified Data Warehouse 103-TB Maximum User Data Tegile Systems Solution Review 2U Design: Featuring Tegile T3800 All-Flash Storage Array http:// www.tegile.com/solutiuons/sql

More information

Capacity Management for Oracle Database Machine Exadata v2

Capacity Management for Oracle Database Machine Exadata v2 Capacity Management for Oracle Database Machine Exadata v2 Dr. Boris Zibitsker, BEZ Systems NOCOUG 21 Boris Zibitsker Predictive Analytics for IT 1 About Author Dr. Boris Zibitsker, Chairman, CTO, BEZ

More information

NUMA obliviousness through memory mapping

NUMA obliviousness through memory mapping NUMA obliviousness through memory mapping Mrunal Gawade CWI, Amsterdam mrunal.gawade@cwi.nl Martin Kersten CWI, Amsterdam martin.kersten@cwi.nl ABSTRACT With the rise of multi-socket multi-core CPUs a

More information

Safe Harbor Statement

Safe Harbor Statement Safe Harbor Statement "Safe Harbor" Statement: Statements in this presentation relating to Oracle's future plans, expectations, beliefs, intentions and prospects are "forward-looking statements" and are

More information

Hardware Configuration Guide

Hardware Configuration Guide Hardware Configuration Guide Contents Contents... 1 Annotation... 1 Factors to consider... 2 Machine Count... 2 Data Size... 2 Data Size Total... 2 Daily Backup Data Size... 2 Unique Data Percentage...

More information

Software Support for Atomicity and Persistence in Non-volatile Memory

Software Support for Atomicity and Persistence in Non-volatile Memory Software Support for Atomicity and Persistence in Non-volatile Memory Ellis Giles Rice University erg@rice.edu Kshitij Doshi Intel Corp. kshitij.a.doshi@intel.com Peter Varman Rice University pjv@rice.edu

More information

System Architecture. In-Memory Database

System Architecture. In-Memory Database System Architecture for Are SSDs Ready for Enterprise Storage Systems In-Memory Database Anil Vasudeva, President & Chief Analyst, Research 2007-13 Research All Rights Reserved Copying Prohibited Contact

More information

IN-MEMORY DATABASE SYSTEMS. Prof. Dr. Uta Störl Big Data Technologies: In-Memory DBMS - SoSe 2015 1

IN-MEMORY DATABASE SYSTEMS. Prof. Dr. Uta Störl Big Data Technologies: In-Memory DBMS - SoSe 2015 1 IN-MEMORY DATABASE SYSTEMS Prof. Dr. Uta Störl Big Data Technologies: In-Memory DBMS - SoSe 2015 1 Analytical Processing Today Separation of OLTP and OLAP Motivation Online Transaction Processing (OLTP)

More information

3. Memory Persistency Goals. 2. Background

3. Memory Persistency Goals. 2. Background Memory Persistency Steven Pelley Peter M. Chen Thomas F. Wenisch University of Michigan {spelley,pmchen,twenisch}@umich.edu Abstract Emerging nonvolatile memory technologies (NVRAM) promise the performance

More information

Storage Management in the NVRAM Era

Storage Management in the NVRAM Era Storage Management in the NVRAM Era ABSTRACT Steven Pelley University of Michigan spelley@umich.edu Brian T. Gold Oracle Corporation brian.t.gold@gmail.com Emerging nonvolatile memory technologies (NVRAM)

More information

CONSOLIDATING MICROSOFT SQL SERVER OLTP WORKLOADS ON THE EMC XtremIO ALL FLASH ARRAY

CONSOLIDATING MICROSOFT SQL SERVER OLTP WORKLOADS ON THE EMC XtremIO ALL FLASH ARRAY Reference Architecture CONSOLIDATING MICROSOFT SQL SERVER OLTP WORKLOADS ON THE EMC XtremIO ALL FLASH ARRAY An XtremIO Reference Architecture Abstract This Reference architecture examines the storage efficiencies

More information

The Benefits of Virtualizing

The Benefits of Virtualizing T E C H N I C A L B R I E F The Benefits of Virtualizing Aciduisismodo Microsoft SQL Dolore Server Eolore in Dionseq Hitachi Storage Uatummy Environments Odolorem Vel Leveraging Microsoft Hyper-V By Heidi

More information

St eps t o w ards HW/SW- DB- Co De sign. Wolfgang Lehner

St eps t o w ards HW/SW- DB- Co De sign. Wolfgang Lehner St eps t o w ards HW/SW- DB- Co De sign Wolfgang Lehner Modern Hardware All over the place 2 A Look at Hardware Trends 201x 3 A Look at Hardware Trends - 2015 Extreme NUMA Effects System Level 2 Component

More information

Resolving Journaling of Journal Anomaly via Weaving Recovery Information into DB Page. Beomseok Nam

Resolving Journaling of Journal Anomaly via Weaving Recovery Information into DB Page. Beomseok Nam NVRAMOS 14 10.30. 2014 Resolving Journaling of Journal Anomaly via Weaving Recovery Information into DB Page Beomseok Nam UNIST Outline Motivation Journaling of Journal Anomaly How to resolve Journaling

More information

Cloud Storage. Parallels. Performance Benchmark Results. White Paper. www.parallels.com

Cloud Storage. Parallels. Performance Benchmark Results. White Paper. www.parallels.com Parallels Cloud Storage White Paper Performance Benchmark Results www.parallels.com Table of Contents Executive Summary... 3 Architecture Overview... 3 Key Features... 4 No Special Hardware Requirements...

More information

Increase Database Performance by Implementing Cirrus Data Solutions DCS SAN Caching Appliance With the Seagate Nytro Flash Accelerator Card

Increase Database Performance by Implementing Cirrus Data Solutions DCS SAN Caching Appliance With the Seagate Nytro Flash Accelerator Card Implementing Cirrus Data Solutions DCS SAN Caching Appliance With the Seagate Nytro Technology Paper Authored by Rick Stehno, Principal Database Engineer, Seagate Introduction Supporting high transaction

More information

Price/performance Modern Memory Hierarchy

Price/performance Modern Memory Hierarchy Lecture 21: Storage Administration Take QUIZ 15 over P&H 6.1-4, 6.8-9 before 11:59pm today Project: Cache Simulator, Due April 29, 2010 NEW OFFICE HOUR TIME: Tuesday 1-2, McKinley Last Time Exam discussion

More information

Scaling Objectivity Database Performance with Panasas Scale-Out NAS Storage

Scaling Objectivity Database Performance with Panasas Scale-Out NAS Storage White Paper Scaling Objectivity Database Performance with Panasas Scale-Out NAS Storage A Benchmark Report August 211 Background Objectivity/DB uses a powerful distributed processing architecture to manage

More information

Managing Storage Space in a Flash and Disk Hybrid Storage System

Managing Storage Space in a Flash and Disk Hybrid Storage System Managing Storage Space in a Flash and Disk Hybrid Storage System Xiaojian Wu, and A. L. Narasimha Reddy Dept. of Electrical and Computer Engineering Texas A&M University IEEE International Symposium on

More information

Binary search tree with SIMD bandwidth optimization using SSE

Binary search tree with SIMD bandwidth optimization using SSE Binary search tree with SIMD bandwidth optimization using SSE Bowen Zhang, Xinwei Li 1.ABSTRACT In-memory tree structured index search is a fundamental database operation. Modern processors provide tremendous

More information

The Technologies & Architectures. President, Demartek

The Technologies & Architectures. President, Demartek Deep Dive on Solid State t Storage The Technologies & Architectures Dennis Martin Dennis Martin President, Demartek Demartek Company Overview Industry analysis with on-site test lab Lab includes servers,

More information

COS 318: Operating Systems. Virtual Memory and Address Translation

COS 318: Operating Systems. Virtual Memory and Address Translation COS 318: Operating Systems Virtual Memory and Address Translation Today s Topics Midterm Results Virtual Memory Virtualization Protection Address Translation Base and bound Segmentation Paging Translation

More information

Application-Tier In-Memory Analytics Best Practices and Use Cases

Application-Tier In-Memory Analytics Best Practices and Use Cases Application-Tier In-Memory Analytics Best Practices and Use Cases Susan Cheung Vice President Product Management Oracle, Server Technologies Oct 01, 2014 Guest Speaker: Kiran Tailor Senior Oracle DBA and

More information

Analysis of VDI Storage Performance During Bootstorm

Analysis of VDI Storage Performance During Bootstorm Analysis of VDI Storage Performance During Bootstorm Introduction Virtual desktops are gaining popularity as a more cost effective and more easily serviceable solution. The most resource-dependent process

More information

SWISSBOX REVISITING THE DATA PROCESSING SOFTWARE STACK

SWISSBOX REVISITING THE DATA PROCESSING SOFTWARE STACK 3/2/2011 SWISSBOX REVISITING THE DATA PROCESSING SOFTWARE STACK Systems Group Dept. of Computer Science ETH Zürich, Switzerland SwissBox Humboldt University Dec. 2010 Systems Group = www.systems.ethz.ch

More information

Achieving a High Performance OLTP Database using SQL Server and Dell PowerEdge R720 with Internal PCIe SSD Storage

Achieving a High Performance OLTP Database using SQL Server and Dell PowerEdge R720 with Internal PCIe SSD Storage Achieving a High Performance OLTP Database using SQL Server and Dell PowerEdge R720 with This Dell Technical White Paper discusses the OLTP performance benefit achieved on a SQL Server database using a

More information

Azure VM Performance Considerations Running SQL Server

Azure VM Performance Considerations Running SQL Server Azure VM Performance Considerations Running SQL Server Your company logo here Vinod Kumar M @vinodk_sql http://blogs.extremeexperts.com Session Objectives And Takeaways Session Objective(s): Learn the

More information

Enhancing Recovery Using an SSD Buffer Pool Extension

Enhancing Recovery Using an SSD Buffer Pool Extension Enhancing Recovery Using an SSD Buffer Pool Extension Bishwaranjan Bhattacharjee IBM T.J.Watson Research Center bhatta@us.ibm.com Christian Lang* Acelot Inc. clang@acelot.com George A Mihaila* Google Inc.

More information

Memory Channel Storage ( M C S ) Demystified. Jerome McFarland

Memory Channel Storage ( M C S ) Demystified. Jerome McFarland ory nel Storage ( M C S ) Demystified Jerome McFarland Principal Product Marketer AGENDA + INTRO AND ARCHITECTURE + PRODUCT DETAILS + APPLICATIONS THE COMPUTE-STORAGE DISCONNECT + Compute And Data Have

More information

Non Volatile Memory Invades the Memory Bus: Performance and Versatility is the Result

Non Volatile Memory Invades the Memory Bus: Performance and Versatility is the Result Non Volatile Memory Invades the Memory Bus: Performance and Versatility is the Result A discussion on why and how Non Volatile Memory is moving onto the DDR bus and what it means for the industry, as well

More information

Tushar Joshi Turtle Networks Ltd

Tushar Joshi Turtle Networks Ltd MySQL Database for High Availability Web Applications Tushar Joshi Turtle Networks Ltd www.turtle.net Overview What is High Availability? Web/Network Architecture Applications MySQL Replication MySQL Clustering

More information

Solid State Technology What s New?

Solid State Technology What s New? Solid State Technology What s New? Dennis Martin, President, Demartek www.storagedecisions.com Agenda: Solid State Technology What s New? Demartek About Us Solid-state storage overview Types of NAND flash

More information

Removing Performance Bottlenecks in Databases with Red Hat Enterprise Linux and Violin Memory Flash Storage Arrays. Red Hat Performance Engineering

Removing Performance Bottlenecks in Databases with Red Hat Enterprise Linux and Violin Memory Flash Storage Arrays. Red Hat Performance Engineering Removing Performance Bottlenecks in Databases with Red Hat Enterprise Linux and Violin Memory Flash Storage Arrays Red Hat Performance Engineering Version 1.0 August 2013 1801 Varsity Drive Raleigh NC

More information