Flamingo: Enabling Evolvable HDD-based Near-Line Storage

Slides:



Advertisements
Similar presentations
Pelican: A Building Block for Exascale Cold Data Storage
Advertisements

Availability in Globally Distributed Storage Systems
GREEN CLOUD By Sphoorthy. LOGO WHAT IS CLOUD COMPUTING? Cloud computing is a model for enabling convenient, on- demand network access to a shared pool.
Reconfigurable Network Topologies at Rack Scale
1 Exploring Data Reliability Tradeoffs in Replicated Storage Systems NetSysLab The University of British Columbia Abdullah Gharaibeh Matei Ripeanu.
SOFTWARE AS A SERVICE PLATFORM AS A SERVICE INFRASTRUCTURE AS A SERVICE.
Simplify your Job – Automatic Storage Management Angelo Session id:
1 Exploring Data Reliability Tradeoffs in Replicated Storage Systems NetSysLab The University of British Columbia Abdullah Gharaibeh Advisor: Professor.
Challenges of Storage in an Elastic Infrastructure. May 9, 2014 Farid Yavari, Storage Solutions Architect and Technologist.
Introduction To Windows Azure Cloud
A Survey of Mobile Cloud Computing Application Models
PARAID: The Gear-Shifting Power-Aware RAID Charles Weddle, Mathew Oldham, An-I Andy Wang – Florida State University Peter Reiher – University of California,
Midori Life after windows Microsoft Research’s. Singularity  Midori is a stem off of this operating system  A research project started in 2003 to build.
At A Glance VOLT is a freeware, platform independent tool set that coordinates cross-mission observation planning and scheduling among one or more space.
1EMC CONFIDENTIAL—INTERNAL USE ONLY Why EMC for SQL Performance Optimization.
School of EECS, Peking University Microsoft Research Asia UStore: A Low Cost Cold and Archival Data Storage System for Data Centers Quanlu Zhang †, Yafei.
1 © Copyright 2010 EMC Corporation. All rights reserved.  Consolidation  Create economies of scale through standardization  Reduce IT costs  Deliver.
Hadoop Hardware Infrastructure considerations ©2013 OpalSoft Big Data.
Mayuresh Varerkar ECEN 5613 Current Topics Presentation March 30, 2011.
Neil Sanderson 24 October, Early days for virtualisation Virtualization Adoption x86 servers used for virtualization Virtualization adoption.
1 PARAID: A Gear-Shifting Power-Aware RAID Charles Weddle, Mathew Oldham, Jin Qian, An-I Andy Wang – Florida St. University Peter Reiher – University of.
PARAID: A Gear-Shifting Power-Aware RAID Charles Weddle, Mathew Oldham, Jin Qian, An-I Andy Wang – Florida St. University Peter Reiher – University of.
MIDORI The Post Windows Operating System Microsoft Research’s.
A dynamic optimization model for power and performance management of virtualized clusters Vinicius Petrucci, Orlando Loques Univ. Federal Fluminense Niteroi,
1 © Copyright 2011 EMC Corporation. All rights reserved. BIG DATA & Storage Automation George Kokkinakis Enterprise Account Manager.
1 © Copyright 2010 EMC Corporation. All rights reserved. The Virtualization BenefitThe Physical Challenge Virtualizing Microsoft Applications Aging, Inefficient.
PaaSport Introduction on Cloud Computing PaaSport training material.
Network design Topic 6 Testing and documentation.
Modeling Billion-Node Torus Networks Using Massively Parallel Discrete-Event Simulation Ning Liu, Christopher Carothers 1.
OPERATING SYSTEMS CS 3530 Summer 2014 Systems and Models Chapter 03.
Virtualization A brief introduction Virtualization1.
CSE791 COURSE PRESENTATION QIUWEN CHEN Workload-aware Storage.
Improving System Availability in Distributed Environments Sam Malek with Marija Mikic-Rakic Nels.
Course 03 Basic Concepts assist. eng. Jánó Rajmond, PhD
© 2012 Eucalyptus Systems, Inc. Cloud Computing Introduction Eucalyptus Education Services 2.
Monitoreo y Administración de Infraestructura Fisica (DCIM). StruxureWare for Data Centers 2.0 Arturo Maqueo Business Development Data Centers LAM.
SizeCap: Efficiently Handling Power Surges for Fuel Cell Powered Data Centers Yang Li, Di Wang, Saugata Ghose, Jie Liu, Sriram Govindan, Sean James, Eric.
XFabric: a Reconfigurable In-Rack Network for Rack-Scale Computers Sergey Legtchenko, Nicholas Chen, Daniel Cletheroe, Antony Rowstron, Hugh Williams,
Journey to the HyperConverged Agile Infrastructure
The Post Windows Operating System
Optimizing Distributed Actor Systems for Dynamic Interactive Services
Volume Licensing Readiness: Level 100
MERANTI Caused More Than 1.5 B$ Damage
OPERATING SYSTEMS CS 3502 Fall 2017
Seth Pugsley, Jeffrey Jestes,
How Much SSD Is Useful For Resilience In Supercomputers
Volume Licensing Readiness: Level 200
Volume Licensing Readiness: Level 200
Presentation on Software Requirements Submitted by
Volume Licensing Readiness: Level 100
Green cloud computing 2 Cs 595 Lecture 15.
Pelican: A building block for exascale cold data storage
Software Defined Storage
Sebastian Solbach Consulting Member of Technical Staff
Microsoft SAM for Hosting (SPLA)
Volume Licensing Readiness: Level 100
AWS Batch Overview A highly-efficient, dynamically-scaled, batch computing service May 2017.
Volume Licensing Readiness: Level 200
Server Allocation for Multiplayer Cloud Gaming
Gabor Madl Ph.D. Candidate, UC Irvine Advisor: Nikil Dutt
Windows Server 2016 Software Defined Storage
reFresh SSDs: Enabling High Endurance, Low Cost Flash in Datacenters
Explore the Azure Cosmos DB with .NET Core 2.0
Performance Evaluation of Computer Networks
Performance Evaluation of Computer Networks
Unit 2. Day 16..
On the Role of Burst Buffers in Leadership-Class Storage Systems
Unit 2. Day 17..
Area Coverage Problem Optimization by (local) Search
Modeling Event-Based Systems in Ptolemy II EE249 Project Status Report
Presentation transcript:

Flamingo: Enabling Evolvable HDD-based Near-Line Storage Sergey Legtchenko, Xiaozhou Li, Antony Rowstron, Austin Donnelly, Richard Black

Storing Cold Data in the Cloud Pelican rack (Microsoft, [OSDI 14]) 8% HDDs active 576 HDDs/server Open Compute Cold Storage rack (Facebook) 6% HDDs active 240 HDDs/server Custom racks: trading latency for cost Only fraction of HDDs concurrently active Reduced #servers (1 or 2 per rack) Cold data: rarely accessed data Challenge: storing cold data at low cost Benefits Lower capital cost Capped resource consumption Higher storage density

Designing Cold Storage Racks is Hard Resources are constrained in the rack Resource constraints: 1 HDD / cooling column 2 HDDs / tray Vibration, bandwidth… Power Cooling Software: co-designed, constraint-aware Data Layout IO Scheduler Experience from building Pelican Design complexity Storage stack is brittle to design changes Impact of resource provisioning on end performance? Pelican: 8% HDDs active

Flamingo: a Tool to Help Cold Storage Rack Design In the rest of the talk Input Generic Storage Stack Data layout IO scheduler parameters Online Constraint Solver Data layout, IO scheduler + Storage stack configuration Set of rack descriptions: same topology varying resource provisioning Perf. Analysis (simulator) Rack description Constraints Hardware properties Resource provisioning exploration Performance goals Rack Resource provisioning specification

Rack Description Resource Domain: set of HDDs sharing a limited resource {D1,D2,D3,D4}: 40 {D5,D6,D7,D8}: 40 {D4,D8}: 1 {D3,D7}: 1 {D1,D5}: 1 {D2,D6}: 1 Domain A: type: power, budget: 40W Domain B: type: vibration, budget: 1HDD D1 D2 D3 D4 D5 D6 D7 D8 Standby Active Spin up power: 10W vibration: 1 power: 20W power: 2W vibration: 0 HDD: operating states + resource consumption IO-capable state Expresses constraints: e.g. only 1 HDD can be spinning up in A 40W 2 + 20+ 20 = 44W 40W 2 + 20 = 26 can be hard or soft

Storage Stack Configuration Data layout Groups of HDDs that concurrently transition state Minimize inter-group conflicts Constraint Solver Rack description

Storage Stack Configuration Data layout Groups of HDDs that concurrently transition state Minimize inter-group conflicts Generic Storage Stack Data Layout Spin ups/downs, IOs   Blob-store API IO Scheduler Blob Group of HDDs Constraint Solver D1 D2 D4 D5 D6 D7 D8 D3 Conflict Conflicts minimized vibration, budget: 1 power budget: 40 G1: {D3, D8} G2: {D4, D7} G3: {D1, D6} G4: {D2, D5} Group definition D1 D2 D4 D5 D6 D7 D8 D3 vibration, budget: 1 power budget: 40 Each group conflicts with 2 Conflict D1 D6 D1 D5 D4 D2 D7 D2 D3 D4 D8 D4 D5 D6 D7 D8 Sd A Su 10, 1 20, 1 2, 0 power budget: 40 vibration, budget: 1 Rack description Conflicts between groups {D4,D8}: 1 {D3,D7}: 1 {D2,D6}: 1 {D1,D5}: 1 {D1,D2,D3,D4}: 40 {D5,D6,D7,D8}: 40 Rack description Inter-group constraints {G1,G2,G3,G4}: 40 {G1,G2}: 1 {G3,G4}: 1 {G2,G1}: 1 {G1,G2}: 1 {G4,G3}: 1 {G3,G4}: 1 {G3,G4,G1,G2}: 40 {G4,G3,G2,G1}: 40

Resource Provisioning Exploration N resource types: N-dimensional space of rack descriptions Fully provisioned Resource 1 (e.g. power) D1 D2 D4 D5 D6 D7 D8 D3 40 1 Fully provisioned Resource 2 (e.g. vibration)

Resource Provisioning Exploration N resource types: N-dimensional space of rack descriptions Fully-provisioned rack (JBOD), per domain: all HDDs in most resource-consuming state Sd A Su 10, 1 20, 1 2, 0 D1 D2 D4 D5 D6 D7 D8 D3 vibration, budget: 2 power budget: 80 Fully provisioned Bottleneck resource constraint Relaxing Resource 1 (e.g. power) Least-provisioned rack, per domain: 1 HDD in IO-capable state n-1 in lowest resource-consuming D1 D2 D4 D5 D6 D7 D8 D3 vibration, budget: 1 power budget: 26 Fully provisioned Resource 2 (e.g. vibration)

Resource Provisioning Exploration N resource types: N-dimensional space of rack descriptions Fully-provisioned rack (JBOD), per domain: all HDDs in most resource-consuming state Fully provisioned Resource 1 (e.g. power) Discrete surface in the N-dimensional space For Pelican: 747 rack descriptions Bottleneck resource: vibration Bottleneck resource: power Fully provisioned Resource 2 (e.g. vibration)

Evaluation - Pelican Pelican Simulator – Poisson workload, 1GB reads.

Execution Time - Pelican 9 minutes

Execution Time - Pelican 9 minutes Rack name OCP Pelican Rack_A Rack_B Rack_C Rack_D Rack_E #rack descriptions 1921 747 1421 1152 973 649 683

Execution Time for Different Racks 9 minutes 3 hours Rack name OCP Pelican Rack_A Rack_B Rack_C Rack_D Rack_E #rack descriptions 1921 747 1421 1152 973 649 683

Conclusion Cold storage racks: Co-design: resource-constrained hardware + constraint-aware software low cost but hard to (re)design Flamingo simplifies design of cold storage racks Synthesizes Data Layout and IO Scheduler parameters Explores impact of resource provisioning on end performance Redesign in days vs months manually