A simple object storage system for web applications Dan Pollack AOL



Similar documents
Distributed File System Choices: Red Hat Storage, GFS2 & pnfs

OPTIMIZING PRIMARY STORAGE WHITE PAPER FILE ARCHIVING SOLUTIONS FROM QSTAR AND CLOUDIAN

Building Storage as a Service with OpenStack. Greg Elkinbard Senior Technical Director

Hadoop Distributed File System. T Seminar On Multimedia Eero Kurkela

Diagram 1: Islands of storage across a digital broadcast workflow

The Panasas Parallel Storage Cluster. Acknowledgement: Some of the material presented is under copyright by Panasas Inc.

Service Description Cloud Storage Openstack Swift

Object Storage: A Growing Opportunity for Service Providers. White Paper. Prepared for: 2012 Neovise, LLC. All Rights Reserved.

Designing a Cloud Storage System

References. Introduction to Database Systems CSE 444. Motivation. Basic Features. Outline: Database in the Cloud. Outline

Introduction to Database Systems CSE 444

Neil Stobart Cloudian Inc. CLOUDIAN HYPERSTORE Smart Data Storage

How Comcast Built An Open Source Content Delivery Network National Engineering & Technical Operations

OBIEE 11g Scaleout & Clustering

Building Storage-as-a-Service Businesses

SWIFT. Page:1. Openstack Swift. Object Store Cloud built from the grounds up. David Hadas Swift ATC. HRL 2012 IBM Corporation

Introduction to OpenStack

EDG Project: Database Management Services

Lab Validation Report

TRANSFORMING DATA PROTECTION

(Scale Out NAS System)

3 Techniques for Database Scalability with Hibernate. Geert Bevin - SpringOne 2009

21 st Century Storage What s New and What s Changing

Building Storage Service in a Private Cloud

Cloud Based Application Architectures using Smart Computing

No.1 IT Online training institute from Hyderabad URL: sriramtechnologies.com

Tushar Joshi Turtle Networks Ltd

Petabyte Scale Data at Facebook. Dhruba Borthakur, Engineer at Facebook, SIGMOD, New York, June 2013

The deployment of OHMS TM. in private cloud

STORAGE CENTER. The Industry s Only SAN with Automated Tiered Storage STORAGE CENTER

Comparison of the High Availability and Grid Options

Introduction to NetApp Infinite Volume

Introduction to Cloud : Cloud and Cloud Storage. Lecture 2. Dr. Dalit Naor IBM Haifa Research Storage Systems. Dalit Naor, IBM Haifa Research

Glassfish Architecture.

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

Hitachi Content Platform as a Continuous Integration Build Artifact Storage System

Scale out NAS on the outside, Object storage on the inside

Where We Are. References. Cloud Computing. Levels of Service. Cloud Computing History. Introduction to Data Management CSE 344

Flash Databases: High Performance and High Availability

Cloud computing - Architecting in the cloud

1. Comments on reviews a. Need to avoid just summarizing web page asks you for:

STORIANT TECHNOLOGY DEEP DIVE. Data Storage for a New Generation

Hadoop: Embracing future hardware

Casper Suite. Security Overview

Alfresco Enterprise on AWS: Reference Architecture

Getting performance & scalability on standard platforms, the Object vs Block storage debate. Copyright 2013 MPSTOR LTD. All rights reserved.

SQL Server 2012/2014 AlwaysOn Availability Group

Storage Made Easy Enterprise File Share and Sync (EFSS) Cloud Control Gateway Architecture

The Design and Implementation of the Zetta Storage Service. October 27, 2009

Using Oracle NoSQL Database

3. PGCluster. There are two formal PGCluster Web sites.

Implementing Multi-Tenanted Storage for Service Providers with Cloudian HyperStore. The Challenge SOLUTION GUIDE

An Approach to Implement Map Reduce with NoSQL Databases

SwiftStack Filesystem Gateway Architecture

Google File System. Web and scalability

ZooKeeper. Table of contents

How Cisco IT Built a Storage Cloud for Big Data

Distributed File Systems

Intro to AWS: Storage Services

HDFS Architecture Guide

Request Routing, Load-Balancing and Fault- Tolerance Solution - MediaDNS

Enterprise Linux Business Continuity Solutions for Critical Applications

Data Backup and Restore (DBR) Overview Detailed Description Pricing... 5 SLAs... 5 Service Matrix Service Description

Tier Architectures. Kathleen Durant CS 3200

ENTERPRISE STORAGE WITH THE FUTURE BUILT IN

High Availability with Elixir

Oracle BI Publisher Enterprise Cluster Deployment. An Oracle White Paper August 2007

Database Monitoring Requirements. Salvatore Di Guida (CERN) On behalf of the CMS DB group

Migration and Building of Data Centers in IBM SoftLayer with the RackWare Management Module

WOS OBJECT STORAGE PRODUCT BROCHURE DDN.COM Full Spectrum Object Storage

Configure AlwaysOn Failover Cluster Instances (SQL Server) using InfoSphere Data Replication Change Data Capture (CDC) on Windows Server 2012

THE HADOOP DISTRIBUTED FILE SYSTEM

In Memory Accelerator for MongoDB

IBM Spectrum Protect in the Cloud

Product Overview. UNIFIED COMPUTING Managed Hosting - Storage Data Sheet

Amazon Elastic Beanstalk

Ecomm Enterprise High Availability Solution. Ecomm Enterprise High Availability Solution (EEHAS) Page 1 of 7

Resource control in ATLAS distributed data management: Rucio Accounting and Quotas

XTM Web 2.0 Enterprise Architecture Hardware Implementation Guidelines. A.Zydroń 18 April Page 1 of 12

Cluster Computing. ! Fault tolerance. ! Stateless. ! Throughput. ! Stateful. ! Response time. Architectures. Stateless vs. Stateful.

Storage Systems Autumn Chapter 6: Distributed Hash Tables and their Applications André Brinkmann

Fax Server Cluster Configuration

Understanding Enterprise NAS

Robert Honeyman Honeyman IT Consulting.

Non-Stop Hadoop Paul Scott-Murphy VP Field Techincal Service, APJ. Cloudera World Japan November 2014

CompTIA Cloud+ 9318; 5 Days, Instructor-led

Oracle s Big Data solutions. Roger Wullschleger. <Insert Picture Here>

Scalability and Reliability Features of MySQL Connector/J

Transcription:

A simple object storage system for web applications Dan Pollack AOL

AOL Leading edge web services company AOL s business spans the internet 2

Motivation Most web content is static and shared Traditional NAS systems inefficient and costly for content distribution Every interface to content is unique per application 3

Background circa 2006 Google file system Cluster file systems Gluster Lustre IBrix Scalable NAS Isilon Onstor Parallel file systems pnfs Oceanstore 4

First attempt IBrix Commodity hardware Scalable metadata Scalable cluster Good resilience Problems Hierarchical metadata Weak metadata replication Client software required Client and server version mismatches 5

Second attempt Object store Purpose built Commodity hardware Open source software components Linux Tomcat JAVA MySQL Simple external API Manageability prioritized 6

Requirements Shared nothing components Scalable metadata Separate metadata and data system components Asymmetric components allowed Multi-site capable RESTful external API POST GET DELETE 7

Requirements Multi-tenant Strong data protection Availability Durability Background checking and recovery External security but internal access control Extended object metadata Modular Performance monitoring external system Hardware monitoring internal and external together 8

Implementation User/Application Clients HSS Load Balancer VIP HTTP Requests HTTP Return HSS Storage Nodes HTTP Requests HTTP Return Admin Console Admin Tasks HTTP Requests HTTP Return HTTP Requests HTTP Return HSS RW MySQL ATOMICS Load Balancer VIP HSS RO MySQL ATOMICS Load Balancer VIP HTTP Requests HSS Admin MySQL ATOMICS Load Balancer VIP MySQL Replication MySQL Replication 9

Write example POST request to VIP from client Load balancer selects storage server Calculate OID Write file locally Update DB with new OID and server owner Create second replica copy Update DB with OID and second server owner Return OID to client Set replication flag in DB to create third replica 10

Read example GET request to VIP from client Load balancer selects storage server Storage server checks local cache for OID Cache miss causes OID lookup in DB DB returns location of all replicas Storage server retrieves one of the replicas Storage server returns the file to the requestor If the file is above the redirect threshold send 302 redirect 11

Common failures DB unavailable for write 502 server error Write failure of initial file 500 server error Write failure of second replica retry File not in DB 404 not found File retrieved corrupt or unavailable Use different replica Schedule replication to proper number of required replicas 12

Features Automatic file expiration configurable by application OID can be specified for application flexibility Frequently accessed files are cached on all servers Usage accounting 13

Some statistics 98% of all requests take less than 100ms 99.5% of all requests take less than 200ms Over 200M requests in a single day Over 400M objects managed 165TB of objects served per month 20+ applications storing files 14

Future enhancements Containers for objects improve performance and reliability Better geographic awareness location affinity and latency improvements Storage tiers better resource allocation and performance Improved modularity different storage and metadata backends 15

Questions? 16