SlideShare a Scribd company logo
!
Presented by Silvius Rus, Director, Big Data Platforms!
December 2013!
!
!
!
Quantcast

Petabyte Storage at Half Price with QFS!
12-131
Big Data Gurus Meetup!
© Quantcast 2012!
Quantcast File System (QFS)!
A high performance
alternative to the Hadoop
Distributed File System
(HDFS).!
!
12-13!Quantcast File System! 2!
Manages multi-petabyte Hadoop workloads
with significantly faster I/O than HDFS and uses
only half the disk space.!
Offers massive cost savings to large scale
Hadoop users (fewer disks = fewer machines).!
Production hardened at Quantcast under
massive processing loads (multi exabyte).!
Fully Compatible with Apache Hadoop.!
100% Open Source.!
!
!
© Quantcast 2012!
Quantcast 

Technology Innovation Timeline!
12-13!Quantcast File System! 3!
Quantcast!
Measurement!
Launched!
!
Quantcast!
Advertising!
Launched!
!
!
Launch!
QFS!
Started!
using!
Hadoop!
Using and!
sponsoring!
KFS!
Turned!
off!
HDFS!
Receiving!
1TB/day!
Receiving!
10TB/day!
Receiving!
20TB/day!
Receiving!
40TB/day!
Processing!
1PB/day!
Processing!
10PB/day!
Processing!
20PB/day!
2006! 2007! 2008! 2009! 2010! 2011! 2012! 2013!
© Quantcast 2012!
Architecture!
12-13!Quantcast File System! 4!
Rack	
  1
Client
Metaserver
Rack	
  2
Chunk	
  servers
Chunk	
  servers
Chunk	
  Server
·∙ 	
  Handles	
  IO	
  to	
  locally	
  stored	
  
64MB	
  chunks
·∙ 	
  Monitors	
  host	
  file	
  system	
  health
·∙ 	
  Replicates	
  and	
  recovers	
  chunks	
  
as	
  metaserver	
  directsMetaserver
·∙ 	
  Maps	
  /file/paths	
  to	
  chunk	
  ids
·∙ 	
  Manages	
  chunk	
  locations
·∙ 	
  Directs	
  clients	
  to	
  chunk	
  servers
Client
·∙ 	
  Implements	
  high	
  level	
  file	
  
interface	
  (read/write/delete)
·∙ 	
  On	
  write,	
  RS	
  encodes	
  chunks	
  
and	
  distributes	
  stripes	
  to	
  nine	
  
chunk	
  servers.
·∙ 	
  On	
  read,	
  collects	
  RS	
  stripes	
  
from	
  six	
  chunk	
  servers	
  and	
  
recomposes	
  chunk.
Locate	
  or	
  
allocate	
  chunks
Read/write	
  
RS	
  encoded	
  
data	
  from/to	
  
chunk	
  servers
Chunk	
  replication
and	
  rebalancing
instructions
Copy/Recover
chunks
© Quantcast 2012!
QFS vs. HDFS!
Broadly comparable feature
set, with significant storage
efficiency advantages.!
12-13!Quantcast File System! 5!
Feature! QFS! HDFS!
Scalable, distributed storage designed
for efficient batch processing! ü! ü!
Open source! ü! ü!
Hadoop compatible! ü! ü!
Unix style file permissions! ü! ü!
Error Recovery mechanism!
Reed-Solomon
encoding!
Multiple data
copies!
Disk space required

(as a multiple of raw data)!
1.5x! 3x!
© Quantcast 2012!
Reed-Solomon Error Correction

Leveraging high-speed modern networks!
HDFS optimizes toward data
locality for older networks.!
10Gbps networks are now
common, making disk I/O 

a more critical bottleneck.!
QFS leverages faster networks 

to achieve better parallelism 

and encoding efficiency.!
Result: higher error tolerance,
faster performance, with half 

the disk space.!
12-13!Quantcast File System! 6!
1.  Break original data into
64K stripes.!
2.  Reed-Solomon generates
three parity stripes for
every six data strips!
!
3.  Write those to nine
different drives.!
4.  Up to three stripes can
become unreadable...!
5.  …yet the original data can
still be recovered !
Every write parallelized
across 9 drives, every
read across 6!
Reed-Solomon Parallel Data I/O!
© Quantcast 2012!
MapReduce on 6+3 Erasure Coded Files

versus 3x Replicated Files!
Positives!
Writing is ½ off, both in terms of space
and time!
Any 3 broken or slow devices will be
tolerated vs. any 2 with 3-way replication!
Re-executed stragglers run faster due to
reading from multiple devices (striping)!
12-13!Quantcast File System! 7!
Negatives!
There is no locality, reading will require the network!
On read failure, recovery is needed – however it’s
lightning fast on modern CPUs (2 GB/s per core)!
Writes don’t achieve network line rate as original +
parity data is written by a single client!
© Quantcast 2012!
Read/Write Benchmarks!
12-13!Quantcast File System! 8!
0
2
4
6
8
10
12
14
16
18
Write Read
End-to-endtime(minutes)
HDFS 64 MB
HDFS 2.5 GB
QFS 64 MB
End-to-end 20 TB write test
End-to-end 20 TB read test
8,000 workers * 2.5 GB each
Tests ran as Hadoop MapReduce jobs
© Quantcast 2012!
Read/Write Benchmarks!
12-13!Quantcast File System! 9!
0
2
4
6
8
10
12
14
16
18
Write Read
End-to-endtime(minutes)
HDFS 64 MB
HDFS 2.5 GB
QFS 64 MB
End-to-end 20 TB write test
End-to-end 20 TB read test
8,000 workers * 2.5 GB each
Tests ran as Hadoop MapReduce jobs
Host network behavior during tests
QFS write = ½ disk I/O of HDFS write
QFS write à network/disk = 8/9
HDFS write à network/disk = 6/9
QFS read à network/disk = 1
HDFS read à network/disk = very small
© Quantcast 2012!
0 50 100 150 200 250 300
ls
mkdir
rmdir
stat
Operations per second (thousands)
QFS HDFS
Metaserver Performance!
12-13!Quantcast File System! 10!
Intel E5-2670
64 GB RAM
70 million directories
© Quantcast 2012!
Production Hardening for Petascale!
Continuous I/O Balancing!
•  Full feedback loop!
•  Metaserver knows the I/O
queue size of every device!
•  Activity biased towards
under-loaded chunkservers!
•  Direct I/O = short loop!
12-13!Quantcast File System! 11!
Optimization!
•  Direct I/O and fixed buffer
space = predictable RAM
and storage device usage!
•  C++, own memory
allocation and layout!
•  Vector instructions for Reed
Solomon coding!
Operations!
•  Hibernation!
•  Evacuation through
recovery!
•  Continuous space/integrity
rebalancing!
•  Monitoring and alerts!
© Quantcast 2012!
Fast and Efficient MapReduce

Quantsort:All I/O over QFS!
12-13!Quantcast File System! 12!
Concurrent append.
10,000 writers append
to same file at once.
http://qc.st/QCQuantsort
Largest sort = 1 PB
Daily = 1 to 2 PB, max = 3 PB
© Quantcast 2012!
How Well Does It Work!
Reliable at Scale!
Hundreds of days of
metaserver uptime common!
Quantcast MapReduce sorter
uses QFS as distributed
virtualized store instead of local
disk!
8 petabytes of compressed
data!
Close to 1 billion chunks!
7,500 I/O devices!
! 12-13!Quantcast File System! 13!
© Quantcast 2012!
How Well Does It Work!
Reliable at Scale!
Hundreds of days of
metaserver uptime common!
Quantcast MapReduce sorter
uses QFS as distributed
virtualized store instead of local
disk!
8 petabytes of compressed
data!
Close to 1 billion chunks!
7,500 I/O devices!
! 12-13!Quantcast File System! 14!
Fast and Large!
Ran petabyte sort last
weekend.!
Direct I/O not hurting fast
scans: Sawzall query
performance similar to Presto:!
! Presto/
HDFS
Turbo/
QFS
Seconds 16 16
Rows 920 M 970 M
Bytes 31 G 294 G
Rows/sec 57.5 M 60.6 M
Bytes/sec 2.0 G 18.4 G
© Quantcast 2012!
How Well Does It Work!
Reliable at Scale!
Hundreds of days of
metaserver uptime common!
Quantcast MapReduce sorter
uses QFS as distributed
virtualized store instead of local
disk!
8 petabytes of compressed
data!
Close to 1 billion chunks!
7,500 I/O devices!
! 12-13!Quantcast File System! 15!
Fast and Large!
Petabyte sort.!
Direct I/O not hurting fast
scans: Sawzall query
performance similar to Presto:!
!
Easy to Use!
1 Ops Engineer for QFS and
MapReduce on 1,000+ node
cluster!
Neustar set up multi petabyte
instance without help from
Quantcast!
Migrate from HDFS using
hadoop distcp!
Hadoop MapReduce “just
works” on QFS!
Presto/
HDFS
Turbo/
QFS
Seconds 16 16
Rows 920 M 970 M
Bytes 31 G 294 G
Rows/sec 57.5 M 60.6 M
Bytes/sec 2.0 G 18.4 G
© Quantcast 2012!
Metaserver Statistics in Production!
12-13!Quantcast File System! 16!
QFS metaserver statistics over Quantcast production file systems in July 2013.
•  High Availability is nice to have but not a must-have for MapReduce. There are certainly
other use cases where High Availability is a must.
•  Federation may be needed to support file systems beyond 10 PB, depending on file size
© Quantcast 2012!
Chunkserver
12-13!Quantcast File System! 17!
Other Features

Tiered Storage!
RAM
2 SSDs
10 Disks
Chunkserver
RAM
2 SSDs
10 Disks
And 450 more just like them.
Tier Range as File Attribute.
Use tier across 450 machines.
Used in production to
accelerate MapReduce fanout
© Quantcast 2012!12-13!Quantcast File System! 18!
94.5
16.7
8.5
4.8
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
HDFS Default HDFS Small Blocks QFS on Disk QFS in RAM
BroadcastTime(s)
Configuration
Other Features

Fast Broadcast through Wide Striping!
© Quantcast 2012!12-13!Quantcast File System! 19!
700
7
0
100
200
300
400
500
600
700
800
HDFS QFS
Time (msec)
Time (msec)
Refreshingly Fast Command Line Tool

hadoop fs -ls / versus qfs –ls /!
© Quantcast 2012!
Who will find QFS valuable?!
Likely to benefit from QFS!
Existing Hadoop users with large-scale
data clusters.!
Data heavy, tech savvy organizations for
whom performance and efficient use of
hardware are high priorities.!
12-13!Quantcast File System! 20!
May find HDFS a better fit!
Small or new Hadoop deployments, as
HDFS has been deployed in a broader
variety of production environments.!
Clusters with slow or unpredictable 

network connectivity.!
Environments needing specific HDFS
features such as head node federation 

or hot standby.!
!
© Quantcast 2012!
Summary!
Key Benefits of QFS!
Delivers stable high performance alternative to 

HDFS in a production-hardened 1.0 release!
Offers high performance management of 

multi-petabyte workloads!
Faster I/O than HDFS with half the disk space.!
Fully Compatible with Apache Hadoop!
100% Open Source!
© Quantcast 2012!
Future Work!
What QFS Doesn’t Have Just Yet!
Kerberos Security – under development!
HA – No strong case at Quantcast, but nice to have!
Federation – Not a strong case either at Quantcast!
Contributions welcome!!
New York

432 ParkAvenue South

New York, NY 10016!
San Francisco

201 Third Street

San Francisco, CA94103!
London

48 Charlotte Street

London, W1T 2NS!
© Quantcast 2012!
Thank You. 

Questions? !
!
Download QFS for free at:

github.com/quantcast/qfs!
12-13! 23!Quantcast File System!

More Related Content

What's hot

Introduction to Hadoop
Introduction to HadoopIntroduction to Hadoop
Introduction to Hadoop
Dr. C.V. Suresh Babu
 
Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...
Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...
Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...
Databricks
 
Introduction to Apache Spark
Introduction to Apache SparkIntroduction to Apache Spark
Introduction to Apache Spark
Anastasios Skarlatidis
 
Spark 의 핵심은 무엇인가? RDD! (RDD paper review)
Spark 의 핵심은 무엇인가? RDD! (RDD paper review)Spark 의 핵심은 무엇인가? RDD! (RDD paper review)
Spark 의 핵심은 무엇인가? RDD! (RDD paper review)
Yongho Ha
 
Optimizing Apache Spark SQL Joins
Optimizing Apache Spark SQL JoinsOptimizing Apache Spark SQL Joins
Optimizing Apache Spark SQL Joins
Databricks
 
RAPIDS – Open GPU-accelerated Data Science
RAPIDS – Open GPU-accelerated Data ScienceRAPIDS – Open GPU-accelerated Data Science
RAPIDS – Open GPU-accelerated Data Science
Data Works MD
 
PySpark Best Practices
PySpark Best PracticesPySpark Best Practices
PySpark Best Practices
Cloudera, Inc.
 
Dynamic Partition Pruning in Apache Spark
Dynamic Partition Pruning in Apache SparkDynamic Partition Pruning in Apache Spark
Dynamic Partition Pruning in Apache Spark
Databricks
 
Data Source API in Spark
Data Source API in SparkData Source API in Spark
Data Source API in Spark
Databricks
 
Introduction to Apache Spark
Introduction to Apache SparkIntroduction to Apache Spark
Introduction to Apache Spark
Datio Big Data
 
Spark overview
Spark overviewSpark overview
Spark overview
Lisa Hua
 
Introduction to Redis
Introduction to RedisIntroduction to Redis
Introduction to Redis
Arnab Mitra
 
Top 5 Mistakes When Writing Spark Applications
Top 5 Mistakes When Writing Spark ApplicationsTop 5 Mistakes When Writing Spark Applications
Top 5 Mistakes When Writing Spark Applications
Spark Summit
 
The Rise of ZStandard: Apache Spark/Parquet/ORC/Avro
The Rise of ZStandard: Apache Spark/Parquet/ORC/AvroThe Rise of ZStandard: Apache Spark/Parquet/ORC/Avro
The Rise of ZStandard: Apache Spark/Parquet/ORC/Avro
Databricks
 
Beyond SQL: Speeding up Spark with DataFrames
Beyond SQL: Speeding up Spark with DataFramesBeyond SQL: Speeding up Spark with DataFrames
Beyond SQL: Speeding up Spark with DataFrames
Databricks
 
Hive, Presto, and Spark on TPC-DS benchmark
Hive, Presto, and Spark on TPC-DS benchmarkHive, Presto, and Spark on TPC-DS benchmark
Hive, Presto, and Spark on TPC-DS benchmark
Dongwon Kim
 
Ceph Object Storage Performance Secrets and Ceph Data Lake Solution
Ceph Object Storage Performance Secrets and Ceph Data Lake SolutionCeph Object Storage Performance Secrets and Ceph Data Lake Solution
Ceph Object Storage Performance Secrets and Ceph Data Lake Solution
Karan Singh
 
Why your Spark Job is Failing
Why your Spark Job is FailingWhy your Spark Job is Failing
Why your Spark Job is Failing
DataWorks Summit
 
Debunking the Myths of HDFS Erasure Coding Performance
Debunking the Myths of HDFS Erasure Coding Performance Debunking the Myths of HDFS Erasure Coding Performance
Debunking the Myths of HDFS Erasure Coding Performance
DataWorks Summit/Hadoop Summit
 
Hadoop Overview & Architecture
Hadoop Overview & Architecture  Hadoop Overview & Architecture
Hadoop Overview & Architecture
EMC
 

What's hot (20)

Introduction to Hadoop
Introduction to HadoopIntroduction to Hadoop
Introduction to Hadoop
 
Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...
Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...
Apache Spark vs Apache Spark: An On-Prem Comparison of Databricks and Open-So...
 
Introduction to Apache Spark
Introduction to Apache SparkIntroduction to Apache Spark
Introduction to Apache Spark
 
Spark 의 핵심은 무엇인가? RDD! (RDD paper review)
Spark 의 핵심은 무엇인가? RDD! (RDD paper review)Spark 의 핵심은 무엇인가? RDD! (RDD paper review)
Spark 의 핵심은 무엇인가? RDD! (RDD paper review)
 
Optimizing Apache Spark SQL Joins
Optimizing Apache Spark SQL JoinsOptimizing Apache Spark SQL Joins
Optimizing Apache Spark SQL Joins
 
RAPIDS – Open GPU-accelerated Data Science
RAPIDS – Open GPU-accelerated Data ScienceRAPIDS – Open GPU-accelerated Data Science
RAPIDS – Open GPU-accelerated Data Science
 
PySpark Best Practices
PySpark Best PracticesPySpark Best Practices
PySpark Best Practices
 
Dynamic Partition Pruning in Apache Spark
Dynamic Partition Pruning in Apache SparkDynamic Partition Pruning in Apache Spark
Dynamic Partition Pruning in Apache Spark
 
Data Source API in Spark
Data Source API in SparkData Source API in Spark
Data Source API in Spark
 
Introduction to Apache Spark
Introduction to Apache SparkIntroduction to Apache Spark
Introduction to Apache Spark
 
Spark overview
Spark overviewSpark overview
Spark overview
 
Introduction to Redis
Introduction to RedisIntroduction to Redis
Introduction to Redis
 
Top 5 Mistakes When Writing Spark Applications
Top 5 Mistakes When Writing Spark ApplicationsTop 5 Mistakes When Writing Spark Applications
Top 5 Mistakes When Writing Spark Applications
 
The Rise of ZStandard: Apache Spark/Parquet/ORC/Avro
The Rise of ZStandard: Apache Spark/Parquet/ORC/AvroThe Rise of ZStandard: Apache Spark/Parquet/ORC/Avro
The Rise of ZStandard: Apache Spark/Parquet/ORC/Avro
 
Beyond SQL: Speeding up Spark with DataFrames
Beyond SQL: Speeding up Spark with DataFramesBeyond SQL: Speeding up Spark with DataFrames
Beyond SQL: Speeding up Spark with DataFrames
 
Hive, Presto, and Spark on TPC-DS benchmark
Hive, Presto, and Spark on TPC-DS benchmarkHive, Presto, and Spark on TPC-DS benchmark
Hive, Presto, and Spark on TPC-DS benchmark
 
Ceph Object Storage Performance Secrets and Ceph Data Lake Solution
Ceph Object Storage Performance Secrets and Ceph Data Lake SolutionCeph Object Storage Performance Secrets and Ceph Data Lake Solution
Ceph Object Storage Performance Secrets and Ceph Data Lake Solution
 
Why your Spark Job is Failing
Why your Spark Job is FailingWhy your Spark Job is Failing
Why your Spark Job is Failing
 
Debunking the Myths of HDFS Erasure Coding Performance
Debunking the Myths of HDFS Erasure Coding Performance Debunking the Myths of HDFS Erasure Coding Performance
Debunking the Myths of HDFS Erasure Coding Performance
 
Hadoop Overview & Architecture
Hadoop Overview & Architecture  Hadoop Overview & Architecture
Hadoop Overview & Architecture
 

Similar to Quantcast File System (QFS) - Alternative to HDFS

HDFS presented by VIJAY
HDFS presented by VIJAYHDFS presented by VIJAY
HDFS presented by VIJAYthevijayps
 
Apache hadoop basics
Apache hadoop basicsApache hadoop basics
Apache hadoop basicssaili mane
 
NYC Hadoop Meetup - MapR, Architecture, Philosophy and Applications
NYC Hadoop Meetup - MapR, Architecture, Philosophy and ApplicationsNYC Hadoop Meetup - MapR, Architecture, Philosophy and Applications
NYC Hadoop Meetup - MapR, Architecture, Philosophy and Applications
Jason Shao
 
Hadoop ecosystem framework n hadoop in live environment
Hadoop ecosystem framework  n hadoop in live environmentHadoop ecosystem framework  n hadoop in live environment
Hadoop ecosystem framework n hadoop in live environment
Delhi/NCR HUG
 
Improving Hadoop Resiliency and Operational Efficiency with EMC Isilon
Improving Hadoop Resiliency and Operational Efficiency with EMC IsilonImproving Hadoop Resiliency and Operational Efficiency with EMC Isilon
Improving Hadoop Resiliency and Operational Efficiency with EMC Isilon
DataWorks Summit/Hadoop Summit
 
Seattle Scalability Meetup - Ted Dunning - MapR
Seattle Scalability Meetup - Ted Dunning - MapRSeattle Scalability Meetup - Ted Dunning - MapR
Seattle Scalability Meetup - Ted Dunning - MapR
clive boulton
 
Apache hadoop, hdfs and map reduce Overview
Apache hadoop, hdfs and map reduce OverviewApache hadoop, hdfs and map reduce Overview
Apache hadoop, hdfs and map reduce OverviewNisanth Simon
 
Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...
Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...
Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...Yahoo Developer Network
 
getFamiliarWithHadoop
getFamiliarWithHadoopgetFamiliarWithHadoop
getFamiliarWithHadoop
AmirReza Mohammadi
 
What's new in Hadoop Common and HDFS
What's new in Hadoop Common and HDFS What's new in Hadoop Common and HDFS
What's new in Hadoop Common and HDFS
DataWorks Summit/Hadoop Summit
 
Hadoop technology
Hadoop technologyHadoop technology
Hadoop technology
tipanagiriharika
 
Pilot Hadoop Towards 2500 Nodes and Cluster Redundancy
Pilot Hadoop Towards 2500 Nodes and Cluster RedundancyPilot Hadoop Towards 2500 Nodes and Cluster Redundancy
Pilot Hadoop Towards 2500 Nodes and Cluster Redundancy
Stuart Pook
 
Hadoop Interview Questions and Answers by rohit kapa
Hadoop Interview Questions and Answers by rohit kapaHadoop Interview Questions and Answers by rohit kapa
Hadoop Interview Questions and Answers by rohit kapa
kapa rohit
 
Hadoop and BigData - July 2016
Hadoop and BigData - July 2016Hadoop and BigData - July 2016
Hadoop and BigData - July 2016
Ranjith Sekar
 
Core concepts and Key technologies - Big Data Analytics
Core concepts and Key technologies - Big Data AnalyticsCore concepts and Key technologies - Big Data Analytics
Core concepts and Key technologies - Big Data AnalyticsKaniska Mandal
 
02.28.13 WANdisco ApacheCon 2013
02.28.13 WANdisco ApacheCon 201302.28.13 WANdisco ApacheCon 2013
02.28.13 WANdisco ApacheCon 2013
WANdisco Plc
 
Hadoop introduction
Hadoop introductionHadoop introduction
Hadoop introduction
Chirag Ahuja
 
Hadoop - HDFS
Hadoop - HDFSHadoop - HDFS
Hadoop - HDFS
KavyaGo
 
Hadoop description
Hadoop descriptionHadoop description
Hadoop description
Hadoop online training
 

Similar to Quantcast File System (QFS) - Alternative to HDFS (20)

HDFS presented by VIJAY
HDFS presented by VIJAYHDFS presented by VIJAY
HDFS presented by VIJAY
 
Apache hadoop basics
Apache hadoop basicsApache hadoop basics
Apache hadoop basics
 
NYC Hadoop Meetup - MapR, Architecture, Philosophy and Applications
NYC Hadoop Meetup - MapR, Architecture, Philosophy and ApplicationsNYC Hadoop Meetup - MapR, Architecture, Philosophy and Applications
NYC Hadoop Meetup - MapR, Architecture, Philosophy and Applications
 
Hadoop ecosystem framework n hadoop in live environment
Hadoop ecosystem framework  n hadoop in live environmentHadoop ecosystem framework  n hadoop in live environment
Hadoop ecosystem framework n hadoop in live environment
 
Improving Hadoop Resiliency and Operational Efficiency with EMC Isilon
Improving Hadoop Resiliency and Operational Efficiency with EMC IsilonImproving Hadoop Resiliency and Operational Efficiency with EMC Isilon
Improving Hadoop Resiliency and Operational Efficiency with EMC Isilon
 
Seattle Scalability Meetup - Ted Dunning - MapR
Seattle Scalability Meetup - Ted Dunning - MapRSeattle Scalability Meetup - Ted Dunning - MapR
Seattle Scalability Meetup - Ted Dunning - MapR
 
Apache hadoop, hdfs and map reduce Overview
Apache hadoop, hdfs and map reduce OverviewApache hadoop, hdfs and map reduce Overview
Apache hadoop, hdfs and map reduce Overview
 
Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...
Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...
Apache Hadoop India Summit 2011 talk "Hadoop Map-Reduce Programming & Best Pr...
 
Hadoop Research
Hadoop Research Hadoop Research
Hadoop Research
 
getFamiliarWithHadoop
getFamiliarWithHadoopgetFamiliarWithHadoop
getFamiliarWithHadoop
 
What's new in Hadoop Common and HDFS
What's new in Hadoop Common and HDFS What's new in Hadoop Common and HDFS
What's new in Hadoop Common and HDFS
 
Hadoop technology
Hadoop technologyHadoop technology
Hadoop technology
 
Pilot Hadoop Towards 2500 Nodes and Cluster Redundancy
Pilot Hadoop Towards 2500 Nodes and Cluster RedundancyPilot Hadoop Towards 2500 Nodes and Cluster Redundancy
Pilot Hadoop Towards 2500 Nodes and Cluster Redundancy
 
Hadoop Interview Questions and Answers by rohit kapa
Hadoop Interview Questions and Answers by rohit kapaHadoop Interview Questions and Answers by rohit kapa
Hadoop Interview Questions and Answers by rohit kapa
 
Hadoop and BigData - July 2016
Hadoop and BigData - July 2016Hadoop and BigData - July 2016
Hadoop and BigData - July 2016
 
Core concepts and Key technologies - Big Data Analytics
Core concepts and Key technologies - Big Data AnalyticsCore concepts and Key technologies - Big Data Analytics
Core concepts and Key technologies - Big Data Analytics
 
02.28.13 WANdisco ApacheCon 2013
02.28.13 WANdisco ApacheCon 201302.28.13 WANdisco ApacheCon 2013
02.28.13 WANdisco ApacheCon 2013
 
Hadoop introduction
Hadoop introductionHadoop introduction
Hadoop introduction
 
Hadoop - HDFS
Hadoop - HDFSHadoop - HDFS
Hadoop - HDFS
 
Hadoop description
Hadoop descriptionHadoop description
Hadoop description
 

More from bigdatagurus_meetup

Apache Sentry for Hadoop security
Apache Sentry for Hadoop securityApache Sentry for Hadoop security
Apache Sentry for Hadoop security
bigdatagurus_meetup
 
Hypertable - massively scalable nosql database
Hypertable - massively scalable nosql databaseHypertable - massively scalable nosql database
Hypertable - massively scalable nosql database
bigdatagurus_meetup
 
Big data beyond the hype may 2014
Big data beyond the hype may 2014Big data beyond the hype may 2014
Big data beyond the hype may 2014
bigdatagurus_meetup
 
What enterprises can learn from Real Time Bidding (RTB)
What enterprises can learn from Real Time Bidding (RTB)What enterprises can learn from Real Time Bidding (RTB)
What enterprises can learn from Real Time Bidding (RTB)
bigdatagurus_meetup
 
Scaling HBase at Pinterest
Scaling HBase at PinterestScaling HBase at Pinterest
Scaling HBase at Pinterest
bigdatagurus_meetup
 
Continuuity Weave
Continuuity WeaveContinuuity Weave
Continuuity Weave
bigdatagurus_meetup
 
Cassandra 2.0 (Introduction)
Cassandra 2.0 (Introduction)Cassandra 2.0 (Introduction)
Cassandra 2.0 (Introduction)
bigdatagurus_meetup
 
Search On Hadoop
Search On HadoopSearch On Hadoop
Search On Hadoop
bigdatagurus_meetup
 
Apache Tez -- A modern processing engine
Apache Tez -- A modern processing engineApache Tez -- A modern processing engine
Apache Tez -- A modern processing engine
bigdatagurus_meetup
 
Cloudera Developer Kit (CDK)
Cloudera Developer Kit (CDK)Cloudera Developer Kit (CDK)
Cloudera Developer Kit (CDK)
bigdatagurus_meetup
 
Lipstick On Pig
Lipstick On Pig Lipstick On Pig
Lipstick On Pig
bigdatagurus_meetup
 

More from bigdatagurus_meetup (11)

Apache Sentry for Hadoop security
Apache Sentry for Hadoop securityApache Sentry for Hadoop security
Apache Sentry for Hadoop security
 
Hypertable - massively scalable nosql database
Hypertable - massively scalable nosql databaseHypertable - massively scalable nosql database
Hypertable - massively scalable nosql database
 
Big data beyond the hype may 2014
Big data beyond the hype may 2014Big data beyond the hype may 2014
Big data beyond the hype may 2014
 
What enterprises can learn from Real Time Bidding (RTB)
What enterprises can learn from Real Time Bidding (RTB)What enterprises can learn from Real Time Bidding (RTB)
What enterprises can learn from Real Time Bidding (RTB)
 
Scaling HBase at Pinterest
Scaling HBase at PinterestScaling HBase at Pinterest
Scaling HBase at Pinterest
 
Continuuity Weave
Continuuity WeaveContinuuity Weave
Continuuity Weave
 
Cassandra 2.0 (Introduction)
Cassandra 2.0 (Introduction)Cassandra 2.0 (Introduction)
Cassandra 2.0 (Introduction)
 
Search On Hadoop
Search On HadoopSearch On Hadoop
Search On Hadoop
 
Apache Tez -- A modern processing engine
Apache Tez -- A modern processing engineApache Tez -- A modern processing engine
Apache Tez -- A modern processing engine
 
Cloudera Developer Kit (CDK)
Cloudera Developer Kit (CDK)Cloudera Developer Kit (CDK)
Cloudera Developer Kit (CDK)
 
Lipstick On Pig
Lipstick On Pig Lipstick On Pig
Lipstick On Pig
 

Recently uploaded

Criminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdfCriminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdf
Criminal IP
 
一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理
一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理
一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理
mbawufebxi
 
Adjusting primitives for graph : SHORT REPORT / NOTES
Adjusting primitives for graph : SHORT REPORT / NOTESAdjusting primitives for graph : SHORT REPORT / NOTES
Adjusting primitives for graph : SHORT REPORT / NOTES
Subhajit Sahu
 
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
Timothy Spann
 
Predicting Product Ad Campaign Performance: A Data Analysis Project Presentation
Predicting Product Ad Campaign Performance: A Data Analysis Project PresentationPredicting Product Ad Campaign Performance: A Data Analysis Project Presentation
Predicting Product Ad Campaign Performance: A Data Analysis Project Presentation
Boston Institute of Analytics
 
一比一原版(UVic毕业证)维多利亚大学毕业证成绩单
一比一原版(UVic毕业证)维多利亚大学毕业证成绩单一比一原版(UVic毕业证)维多利亚大学毕业证成绩单
一比一原版(UVic毕业证)维多利亚大学毕业证成绩单
ukgaet
 
Sample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdf
Sample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdfSample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdf
Sample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdf
Linda486226
 
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...
pchutichetpong
 
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptxData_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
AnirbanRoy608946
 
一比一原版(YU毕业证)约克大学毕业证成绩单
一比一原版(YU毕业证)约克大学毕业证成绩单一比一原版(YU毕业证)约克大学毕业证成绩单
一比一原版(YU毕业证)约克大学毕业证成绩单
enxupq
 
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
74nqk8xf
 
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...
Subhajit Sahu
 
一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理
一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理
一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理
ahzuo
 
一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单
一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单
一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单
vcaxypu
 
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
oz8q3jxlp
 
standardisation of garbhpala offhgfffghh
standardisation of garbhpala offhgfffghhstandardisation of garbhpala offhgfffghh
standardisation of garbhpala offhgfffghh
ArpitMalhotra16
 
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
slg6lamcq
 
FP Growth Algorithm and its Applications
FP Growth Algorithm and its ApplicationsFP Growth Algorithm and its Applications
FP Growth Algorithm and its Applications
MaleehaSheikh2
 
The affect of service quality and online reviews on customer loyalty in the E...
The affect of service quality and online reviews on customer loyalty in the E...The affect of service quality and online reviews on customer loyalty in the E...
The affect of service quality and online reviews on customer loyalty in the E...
jerlynmaetalle
 
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...
2023240532
 

Recently uploaded (20)

Criminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdfCriminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdf
 
一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理
一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理
一比一原版(Bradford毕业证书)布拉德福德大学毕业证如何办理
 
Adjusting primitives for graph : SHORT REPORT / NOTES
Adjusting primitives for graph : SHORT REPORT / NOTESAdjusting primitives for graph : SHORT REPORT / NOTES
Adjusting primitives for graph : SHORT REPORT / NOTES
 
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
 
Predicting Product Ad Campaign Performance: A Data Analysis Project Presentation
Predicting Product Ad Campaign Performance: A Data Analysis Project PresentationPredicting Product Ad Campaign Performance: A Data Analysis Project Presentation
Predicting Product Ad Campaign Performance: A Data Analysis Project Presentation
 
一比一原版(UVic毕业证)维多利亚大学毕业证成绩单
一比一原版(UVic毕业证)维多利亚大学毕业证成绩单一比一原版(UVic毕业证)维多利亚大学毕业证成绩单
一比一原版(UVic毕业证)维多利亚大学毕业证成绩单
 
Sample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdf
Sample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdfSample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdf
Sample_Global Non-invasive Prenatal Testing (NIPT) Market, 2019-2030.pdf
 
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...
 
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptxData_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
 
一比一原版(YU毕业证)约克大学毕业证成绩单
一比一原版(YU毕业证)约克大学毕业证成绩单一比一原版(YU毕业证)约克大学毕业证成绩单
一比一原版(YU毕业证)约克大学毕业证成绩单
 
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
 
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...
 
一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理
一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理
一比一原版(UIUC毕业证)伊利诺伊大学|厄巴纳-香槟分校毕业证如何办理
 
一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单
一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单
一比一原版(ArtEZ毕业证)ArtEZ艺术学院毕业证成绩单
 
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
 
standardisation of garbhpala offhgfffghh
standardisation of garbhpala offhgfffghhstandardisation of garbhpala offhgfffghh
standardisation of garbhpala offhgfffghh
 
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
 
FP Growth Algorithm and its Applications
FP Growth Algorithm and its ApplicationsFP Growth Algorithm and its Applications
FP Growth Algorithm and its Applications
 
The affect of service quality and online reviews on customer loyalty in the E...
The affect of service quality and online reviews on customer loyalty in the E...The affect of service quality and online reviews on customer loyalty in the E...
The affect of service quality and online reviews on customer loyalty in the E...
 
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...
 

Quantcast File System (QFS) - Alternative to HDFS

  • 1. ! Presented by Silvius Rus, Director, Big Data Platforms! December 2013! ! ! ! Quantcast
 Petabyte Storage at Half Price with QFS! 12-131 Big Data Gurus Meetup!
  • 2. © Quantcast 2012! Quantcast File System (QFS)! A high performance alternative to the Hadoop Distributed File System (HDFS).! ! 12-13!Quantcast File System! 2! Manages multi-petabyte Hadoop workloads with significantly faster I/O than HDFS and uses only half the disk space.! Offers massive cost savings to large scale Hadoop users (fewer disks = fewer machines).! Production hardened at Quantcast under massive processing loads (multi exabyte).! Fully Compatible with Apache Hadoop.! 100% Open Source.! ! !
  • 3. © Quantcast 2012! Quantcast 
 Technology Innovation Timeline! 12-13!Quantcast File System! 3! Quantcast! Measurement! Launched! ! Quantcast! Advertising! Launched! ! ! Launch! QFS! Started! using! Hadoop! Using and! sponsoring! KFS! Turned! off! HDFS! Receiving! 1TB/day! Receiving! 10TB/day! Receiving! 20TB/day! Receiving! 40TB/day! Processing! 1PB/day! Processing! 10PB/day! Processing! 20PB/day! 2006! 2007! 2008! 2009! 2010! 2011! 2012! 2013!
  • 4. © Quantcast 2012! Architecture! 12-13!Quantcast File System! 4! Rack  1 Client Metaserver Rack  2 Chunk  servers Chunk  servers Chunk  Server ·∙  Handles  IO  to  locally  stored   64MB  chunks ·∙  Monitors  host  file  system  health ·∙  Replicates  and  recovers  chunks   as  metaserver  directsMetaserver ·∙  Maps  /file/paths  to  chunk  ids ·∙  Manages  chunk  locations ·∙  Directs  clients  to  chunk  servers Client ·∙  Implements  high  level  file   interface  (read/write/delete) ·∙  On  write,  RS  encodes  chunks   and  distributes  stripes  to  nine   chunk  servers. ·∙  On  read,  collects  RS  stripes   from  six  chunk  servers  and   recomposes  chunk. Locate  or   allocate  chunks Read/write   RS  encoded   data  from/to   chunk  servers Chunk  replication and  rebalancing instructions Copy/Recover chunks
  • 5. © Quantcast 2012! QFS vs. HDFS! Broadly comparable feature set, with significant storage efficiency advantages.! 12-13!Quantcast File System! 5! Feature! QFS! HDFS! Scalable, distributed storage designed for efficient batch processing! ü! ü! Open source! ü! ü! Hadoop compatible! ü! ü! Unix style file permissions! ü! ü! Error Recovery mechanism! Reed-Solomon encoding! Multiple data copies! Disk space required
 (as a multiple of raw data)! 1.5x! 3x!
  • 6. © Quantcast 2012! Reed-Solomon Error Correction
 Leveraging high-speed modern networks! HDFS optimizes toward data locality for older networks.! 10Gbps networks are now common, making disk I/O 
 a more critical bottleneck.! QFS leverages faster networks 
 to achieve better parallelism 
 and encoding efficiency.! Result: higher error tolerance, faster performance, with half 
 the disk space.! 12-13!Quantcast File System! 6! 1.  Break original data into 64K stripes.! 2.  Reed-Solomon generates three parity stripes for every six data strips! ! 3.  Write those to nine different drives.! 4.  Up to three stripes can become unreadable...! 5.  …yet the original data can still be recovered ! Every write parallelized across 9 drives, every read across 6! Reed-Solomon Parallel Data I/O!
  • 7. © Quantcast 2012! MapReduce on 6+3 Erasure Coded Files
 versus 3x Replicated Files! Positives! Writing is ½ off, both in terms of space and time! Any 3 broken or slow devices will be tolerated vs. any 2 with 3-way replication! Re-executed stragglers run faster due to reading from multiple devices (striping)! 12-13!Quantcast File System! 7! Negatives! There is no locality, reading will require the network! On read failure, recovery is needed – however it’s lightning fast on modern CPUs (2 GB/s per core)! Writes don’t achieve network line rate as original + parity data is written by a single client!
  • 8. © Quantcast 2012! Read/Write Benchmarks! 12-13!Quantcast File System! 8! 0 2 4 6 8 10 12 14 16 18 Write Read End-to-endtime(minutes) HDFS 64 MB HDFS 2.5 GB QFS 64 MB End-to-end 20 TB write test End-to-end 20 TB read test 8,000 workers * 2.5 GB each Tests ran as Hadoop MapReduce jobs
  • 9. © Quantcast 2012! Read/Write Benchmarks! 12-13!Quantcast File System! 9! 0 2 4 6 8 10 12 14 16 18 Write Read End-to-endtime(minutes) HDFS 64 MB HDFS 2.5 GB QFS 64 MB End-to-end 20 TB write test End-to-end 20 TB read test 8,000 workers * 2.5 GB each Tests ran as Hadoop MapReduce jobs Host network behavior during tests QFS write = ½ disk I/O of HDFS write QFS write à network/disk = 8/9 HDFS write à network/disk = 6/9 QFS read à network/disk = 1 HDFS read à network/disk = very small
  • 10. © Quantcast 2012! 0 50 100 150 200 250 300 ls mkdir rmdir stat Operations per second (thousands) QFS HDFS Metaserver Performance! 12-13!Quantcast File System! 10! Intel E5-2670 64 GB RAM 70 million directories
  • 11. © Quantcast 2012! Production Hardening for Petascale! Continuous I/O Balancing! •  Full feedback loop! •  Metaserver knows the I/O queue size of every device! •  Activity biased towards under-loaded chunkservers! •  Direct I/O = short loop! 12-13!Quantcast File System! 11! Optimization! •  Direct I/O and fixed buffer space = predictable RAM and storage device usage! •  C++, own memory allocation and layout! •  Vector instructions for Reed Solomon coding! Operations! •  Hibernation! •  Evacuation through recovery! •  Continuous space/integrity rebalancing! •  Monitoring and alerts!
  • 12. © Quantcast 2012! Fast and Efficient MapReduce
 Quantsort:All I/O over QFS! 12-13!Quantcast File System! 12! Concurrent append. 10,000 writers append to same file at once. http://qc.st/QCQuantsort Largest sort = 1 PB Daily = 1 to 2 PB, max = 3 PB
  • 13. © Quantcast 2012! How Well Does It Work! Reliable at Scale! Hundreds of days of metaserver uptime common! Quantcast MapReduce sorter uses QFS as distributed virtualized store instead of local disk! 8 petabytes of compressed data! Close to 1 billion chunks! 7,500 I/O devices! ! 12-13!Quantcast File System! 13!
  • 14. © Quantcast 2012! How Well Does It Work! Reliable at Scale! Hundreds of days of metaserver uptime common! Quantcast MapReduce sorter uses QFS as distributed virtualized store instead of local disk! 8 petabytes of compressed data! Close to 1 billion chunks! 7,500 I/O devices! ! 12-13!Quantcast File System! 14! Fast and Large! Ran petabyte sort last weekend.! Direct I/O not hurting fast scans: Sawzall query performance similar to Presto:! ! Presto/ HDFS Turbo/ QFS Seconds 16 16 Rows 920 M 970 M Bytes 31 G 294 G Rows/sec 57.5 M 60.6 M Bytes/sec 2.0 G 18.4 G
  • 15. © Quantcast 2012! How Well Does It Work! Reliable at Scale! Hundreds of days of metaserver uptime common! Quantcast MapReduce sorter uses QFS as distributed virtualized store instead of local disk! 8 petabytes of compressed data! Close to 1 billion chunks! 7,500 I/O devices! ! 12-13!Quantcast File System! 15! Fast and Large! Petabyte sort.! Direct I/O not hurting fast scans: Sawzall query performance similar to Presto:! ! Easy to Use! 1 Ops Engineer for QFS and MapReduce on 1,000+ node cluster! Neustar set up multi petabyte instance without help from Quantcast! Migrate from HDFS using hadoop distcp! Hadoop MapReduce “just works” on QFS! Presto/ HDFS Turbo/ QFS Seconds 16 16 Rows 920 M 970 M Bytes 31 G 294 G Rows/sec 57.5 M 60.6 M Bytes/sec 2.0 G 18.4 G
  • 16. © Quantcast 2012! Metaserver Statistics in Production! 12-13!Quantcast File System! 16! QFS metaserver statistics over Quantcast production file systems in July 2013. •  High Availability is nice to have but not a must-have for MapReduce. There are certainly other use cases where High Availability is a must. •  Federation may be needed to support file systems beyond 10 PB, depending on file size
  • 17. © Quantcast 2012! Chunkserver 12-13!Quantcast File System! 17! Other Features
 Tiered Storage! RAM 2 SSDs 10 Disks Chunkserver RAM 2 SSDs 10 Disks And 450 more just like them. Tier Range as File Attribute. Use tier across 450 machines. Used in production to accelerate MapReduce fanout
  • 18. © Quantcast 2012!12-13!Quantcast File System! 18! 94.5 16.7 8.5 4.8 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 HDFS Default HDFS Small Blocks QFS on Disk QFS in RAM BroadcastTime(s) Configuration Other Features
 Fast Broadcast through Wide Striping!
  • 19. © Quantcast 2012!12-13!Quantcast File System! 19! 700 7 0 100 200 300 400 500 600 700 800 HDFS QFS Time (msec) Time (msec) Refreshingly Fast Command Line Tool
 hadoop fs -ls / versus qfs –ls /!
  • 20. © Quantcast 2012! Who will find QFS valuable?! Likely to benefit from QFS! Existing Hadoop users with large-scale data clusters.! Data heavy, tech savvy organizations for whom performance and efficient use of hardware are high priorities.! 12-13!Quantcast File System! 20! May find HDFS a better fit! Small or new Hadoop deployments, as HDFS has been deployed in a broader variety of production environments.! Clusters with slow or unpredictable 
 network connectivity.! Environments needing specific HDFS features such as head node federation 
 or hot standby.! !
  • 21. © Quantcast 2012! Summary! Key Benefits of QFS! Delivers stable high performance alternative to 
 HDFS in a production-hardened 1.0 release! Offers high performance management of 
 multi-petabyte workloads! Faster I/O than HDFS with half the disk space.! Fully Compatible with Apache Hadoop! 100% Open Source!
  • 22. © Quantcast 2012! Future Work! What QFS Doesn’t Have Just Yet! Kerberos Security – under development! HA – No strong case at Quantcast, but nice to have! Federation – Not a strong case either at Quantcast! Contributions welcome!!
  • 23. New York
 432 ParkAvenue South
 New York, NY 10016! San Francisco
 201 Third Street
 San Francisco, CA94103! London
 48 Charlotte Street
 London, W1T 2NS! © Quantcast 2012! Thank You. 
 Questions? ! ! Download QFS for free at:
 github.com/quantcast/qfs! 12-13! 23!Quantcast File System!