Seagate Kinetic Open Storage Platform. James Hughes and many others



Similar documents
Hadoop Architecture. Part 1

Can High-Performance Interconnects Benefit Memcached and Hadoop?

Reference Design: Scalable Object Storage with Seagate Kinetic, Supermicro, and SwiftStack

Accelerating Enterprise Applications and Reducing TCO with SanDisk ZetaScale Software

Distributed File System. MCSN N. Tonellotto Complements of Distributed Enabling Platforms

Enabling Technologies for Distributed Computing

SOLUTION BRIEF AUGUST All-Flash Server-Side Storage for Oracle Real Application Clusters (RAC) on Oracle Linux

CURSO: ADMINISTRADOR PARA APACHE HADOOP

THE HADOOP DISTRIBUTED FILE SYSTEM

Cisco UCS and Fusion- io take Big Data workloads to extreme performance in a small footprint: A case study with Oracle NoSQL database

StarWind Virtual SAN Best Practices

Enabling Technologies for Distributed and Cloud Computing

Introduction to Cloud Computing

Hadoop: Embracing future hardware

Hadoop Distributed File System. Dhruba Borthakur Apache Hadoop Project Management Committee

VMware Virtual SAN Hardware Guidance. TECHNICAL MARKETING DOCUMENTATION v 1.0

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

How To Run Apa Hadoop 1.0 On Vsphere Tmt On A Hyperconverged Network On A Virtualized Cluster On A Vspplace Tmter (Vmware) Vspheon Tm (

Scaling Cloud-Native Virtualized Network Services with Flash Memory

Enabling High performance Big Data platform with RDMA

Certified Big Data and Apache Hadoop Developer VS-1221

System Requirements Table of contents

Operating Systems. Cloud Computing and Data Centers

Network Virtualization for Large-Scale Data Centers

Benchmarking Hadoop & HBase on Violin

Hadoop Distributed File System. Jordan Prosch, Matt Kipps

New Storage System Solutions

Welcome to the unit of Hadoop Fundamentals on Hadoop architecture. I will begin with a terminology review and then cover the major components

DEPLOYING AND MONITORING HADOOP MAP-REDUCE ANALYTICS ON SINGLE-CHIP CLOUD COMPUTER

Data-Intensive Programming. Timo Aaltonen Department of Pervasive Computing

Distributed File Systems

Commoditisation of the High-End Research Storage Market with the Dell MD3460 & Intel Enterprise Edition Lustre

A virtual SAN for distributed multi-site environments

Converged storage architecture for Oracle RAC based on NVMe SSDs and standard x86 servers

Hadoop Distributed File System. Dhruba Borthakur June, 2007

State of the Art Cloud Infrastructure

EMC E Exam Name: Virtualized Data Center and Cloud Infrastructure Design Specialist

Data Center Storage Solutions

Big Data Trends and HDFS Evolution

Purchase of High Performance Computing (HPC) Central Compute Resources by Northwestern Researchers

Sriram Krishnan, Ph.D.

Optimize VMware and Hyper-V Protection with HP and Veeam

The OpenStack TM Object Storage system

CSE 590: Special Topics Course ( Supercomputing ) Lecture 10 ( MapReduce& Hadoop)

AMD SEAMICRO OPENSTACK BLUEPRINTS CLOUD- IN- A- BOX OCTOBER 2013

N /150/151/160 RAID Controller. N MegaRAID CacheCade. Feature Overview

Big Data Use Case. How Rackspace is using Private Cloud for Big Data. Bryan Thompson. May 8th, 2013

WHITE PAPER BRENT WELCH NOVEMBER

An Oracle White Paper November Backup and Recovery with Oracle s Sun ZFS Storage Appliances and Oracle Recovery Manager

Solution for private cloud computing

EXPERIMENTATION. HARRISON CARRANZA School of Computer Science and Mathematics

Mobile Cloud Computing for Data-Intensive Applications

Building & Optimizing Enterprise-class Hadoop with Open Architectures Prem Jain NetApp

Aqua Connect Load Balancer User Manual (Mac)

Cost-Effective Business Intelligence with Red Hat and Open Source

Technical Note. Dell PowerVault Solutions for Microsoft SQL Server 2005 Always On Technologies. Abstract

Online Remote Data Backup for iscsi-based Storage Systems

Hadoop on OpenStack Cloud. Dmitry Mescheryakov Software

SALSA Flash-Optimized Software-Defined Storage

SMB Direct for SQL Server and Private Cloud

GraySort and MinuteSort at Yahoo on Hadoop 0.23

The Comprehensive Performance Rating for Hadoop Clusters on Cloud Computing Platform

CSE-E5430 Scalable Cloud Computing Lecture 2

Comparative analysis of mapreduce job by keeping data constant and varying cluster size technique

Hyperscale Use Cases for Scaling Out with Flash. David Olszewski

HDFS Architecture Guide

Deploying Cloudera CDH (Cloudera Distribution Including Apache Hadoop) with Emulex OneConnect OCe14000 Network Adapters

OBSERVEIT DEPLOYMENT SIZING GUIDE

Ignify ecommerce. Item Requirements Notes

Cisco Nexus 1000V Switch for Microsoft Hyper-V

Parallels Cloud Server 6.0

Accelerating and Simplifying Apache

Hadoop Lab - Setting a 3 node Cluster. Java -

Cost Efficient VDI. XenDesktop 7 on Commodity Hardware

Comparing SMB Direct 3.0 performance over RoCE, InfiniBand and Ethernet. September 2014

How To Write An Article On An Hp Appsystem For Spera Hana

Maximizing Hadoop Performance and Storage Capacity with AltraHD TM

Notes on Transferring 100 TB of Data Using Globus. William E. Mihalo; Anton Verlygo; Ryan K. Sisk Northwestern University

Open source Google-style large scale data analysis with Hadoop

Pivotal Clustering Concepts Guide

Implementing a Digital Video Archive Based on XenData Software

AirWave 7.7. Server Sizing Guide

Scientific Computing Data Management Visions

The Future of Computing Cisco Unified Computing System. Markus Kunstmann Channels Systems Engineer

Research Article Hadoop-Based Distributed Sensor Node Management System

LSI SAS inside 60% of servers. 21 million LSI SAS & MegaRAID solutions shipped over last 3 years. 9 out of 10 top server vendors use MegaRAID

General system requirements

Zadara Storage Cloud A

System Requirements for Microsoft Dynamics GP 9.0

What s New in VMware vsphere Flash Read Cache TECHNICAL MARKETING DOCUMENTATION

Development of Monitoring and Analysis Tools for the Huawei Cloud Storage

Transcription:

eagate Kinetic Open torage Platform James Hughes and many others

2

2

A 3

A D 3

A D 3

A D 3

D A 3

D No. 77103, LibKinetic effective Jan. 18, 2009, A 3

D No. 77103, LibKinetic effective Jan. 18, 2009, ProtoBuf TCP/IP/GbE A 3

lication Clustering Management Interconnect D No. 77103, LibKinetic effective Jan. 18, 2009, ProtoBuf TCP/IP/GbE Proprietary to ystem Vendor GPL tandard torage A Proprietary to eagate 3

lication Clustering Management Interconnect D No. 77103, LibKinetic effective Jan. 18, 2009, ProtoBuf TCP/IP/GbE C++, Java, Python, Erlang, DIY Proprietary to ystem Vendor GPL tandard torage A Proprietary to eagate 3

lication Clustering Management Interconnect D No. 77103, LibKinetic effective Jan. 18, 2009, ProtoBuf TCP/IP/GbE Proprietary to ystem Vendor GPL tandard torage A Proprietary to eagate 3

lication Clustering Management Interconnect LibKinetic D No. 77103, LibKinetic effective Jan. 18, LibKinetic 2009, ProtoBuf TCP/IP/GbE Proprietary to ystem Vendor GPL tandard torage A Proprietary to eagate 3

lication Clustering Management Interconnect LibKinetic D No. 77103, LibKinetic effective Jan. 18, LibKinetic 2009, ProtoBuf TCP/IP/GbE Proprietary to ystem Vendor GPL tandard torage A D A D A D A D A A D A D A D A D Proprietary to eagate 3

A versus Kinetic Open torage Object torage ystems wift Basho Riak Ceph HDF!!!!! : ubject to NDA!!!!!! 4

Typical HA High Density Intel server Double ocket 48GB Ram 1000w Core A tray Connected to the server A A 5

Low cost HA Configuration Each drive talks to both switches Each switch has 2 by 10Gb/s Ethernet Kinetic Tray talks directly to ToR No servers Eth Eth Core 6

ystem Hardware Typical JBOD architecture Does not require a server, just JBODs to the ToR witch 10 JBOD 60 drives 4TB = 2.4PB/Rack 7

ystem Hardware 8

Kinetic Drive Provides RPC to Key/Value database Data is pre-indexed Compression and other value is easy and transparent P2P (Drive to Drive) copy of key ranges Communicate using existing Data Center Plumbing (TCP/IP) Multiple masters - Data sharing between machines Configurable caching per command WriteThrough, WriteBack, Flush Local space management 9

Kinetic ystems Clustering (performance, reliability, management) Compatibility with large scale applications (3, etc.) Centralized Management Reliability, availability, durability 10

Existing Traffic Flow ToR ToR Data Center : Core Router ubject to NDA Data Node Client A helf A helf 11

Kinetic Traffic Flow ToR ToR Data Center : Core Router ubject to NDA Kinetic helf Client 12

Conventional HDF ystem Client Namenode Datanode Datanode Datanode Datanode... Disk Disk... Disk Disk Disk... Disk Disk Disk... Disk Disk Disk... Disk Conventional Drives 13

HDF on Kinetic Client Namenode DataNode Daemon Datanode Datanode... K K K K K K........................ Disk Disk Disk Disk Disk Disk... Disk Disk... Disk Disk Disk... Disk Conventional Drives Kinetic Drives 14

Cumulative operations ordered by length 100% 80% operations 92% of the operations data Cumulative percentage 60% 40% 20% 32KB 0.5% of the data 0% 1.00 10.00 100.00 1000.00 10000.00 100000.00 Length (KB) 15

Map of Operations 0 512KB Length 0 1 2 3 4 Time (minutes) Location (TB) 0 1 2 3 16

Performance Metrics ame normal performance expectations equential Write: 50MB/s Random Write: 50MB/s equential Read: 50MB/s Random Read: 20% slower than traditional drives 17

Write Performance Results 120 MB/s 6000 Puts/s 90 4500 60 30 3000 1500 0 0 2 4 6 8 0 0 2 4 6 8 1MB values put rate (MB/s) 10B value put rate 18

Goals of API Data movement Get/put/delete/getnext/getprevious Versioned (== for success), options Range operations Multiple masters Authentication/Integrity/Authorization Cluster-able imple cluster configuration version enforcement P2P copy Management 19

Bootstrapping devices DHCP erver DHCP erver BMC 20

Bootstrapping devices Mgmt erver DHCP erver DHCP erver BMC 20

Bootstrapping devices Mgmt erver DHCP erver DHCP erver BMC 20

Bootstrapping devices M Client Object torage Client Object M torage Mgmt erver DHCP erver DHCP erver BMC 20

Conclusion Next Generation torage Devices Dis-intermediates cloud applications to drive Enable innovation in hardware and software ecosystem Lower TCO Integration with: wift HDF Basho Riak Ceph cality 21

More information http://seagate.com/www/kinetic https://developers.seagate.com http://github.com/eagate 22