Performance Optimization of Hadoop Using InfiniBand RDMA

Performance Optimization of Hadoop Using InfiniBand RDMA

Dhabaleswar K. Panda

The Ohio State University

E-mail: [email protected]

http://www.cse.ohio-state.edu/~panda

Talk at ISC Big Data (October 2014)

by

http://www.cse.ohio-state.edu/~panda

• Substantial impact on designing and utilizing modern data management and

processing systems in multiple tiers

– Front-end data accessing and serving (Online)

• Memcached + DB (e.g. MySQL), HBase

– Back-end data analytics (Offline)

• Hadoop: HDFS, MapReduce, Spark (This talk)

ISC Big Data '14

Data Management and Processing in Modern Datacenters

2

• Open-source implementation of Google MapReduce, GFS, and BigTable

for Big Data Analytics

• Hadoop Common Utilities (RPC, etc.), HDFS, MapReduce, YARN

• http://hadoop.apache.org

3

Overview of Apache Hadoop Architecture

ISC Big Data '14

Hadoop Distributed File System (HDFS)

MapReduce (Cluster Resource Management & Data Processing)

Hadoop Common/Core (RPC, ..)

Hadoop Distributed File System (HDFS)

YARN (Cluster Resource Management & Job Scheduling)

Hadoop Common/Core (RPC, ..)

MapReduce (Data Processing)

Other Models (Data Processing)

Hadoop 1.x Hadoop 2.x

HDFS and MapReduce in Apache Hadoop

• HDFS: Primary storage of Hadoop; highly reliable and fault-tolerant – NameNode stores the file system namespace

– DataNodes store data blocks

• MapReduce: Computing framework of Hadoop; highly scalable

– Map tasks read data from HDFS, operate on it, and write the intermediate

data to local disk

– Reduce tasks get data by shuffle, operate on it and write output to HDFS

• Adopted by many reputed organizations

– eg: Facebook, Yahoo!

• Developed in Java for platform-independence and portability

• Uses Sockets/HTTP for communication!

4 ISC Big Data '14

Spark Architecture Overview

• An in-memory data-processing framework – Iterative machine learning jobs

– Interactive data analytics

– Scala based Implementation

– Standalone, YARN, Mesos

• Scalable and communication intensive – Wide dependencies between

Resilient Distributed Datasets (RDDs)

– MapReduce-like shuffle operations to repartition RDDs

– Sockets based communication

ISC Big Data '14

http://spark.apache.org

5

http://spark.apache.org

• High End Computing (HEC) is growing dramatically

– High Performance Computing

– Big Data Computing

• Technology Advancement

– Multi-core/many-core technologies and accelerators

– Remote Direct Memory Access (RDMA)-enabled

networking (InfiniBand and RoCE)

– Solid State Drives (SSDs) and Non-Volatile Random-

Access Memory (NVRAM)

– Accelerators (NVIDIA GPGPUs and Intel Xeon Phi)

Growth in High End Computing (HEC)

Tianhe – 2 (1) Titan (2) Stampede (6) Tianhe – 1A (10)

ISC Big Data '14 6

• 223 IB Clusters (44.3%) in the June 2014 Top500 list

(http://www.top500.org)

• Installations in the Top 50 (25 systems):

Large-scale InfiniBand Installations

519,640 cores (Stampede) at TACC (7th) 120, 640 cores (Nebulae) at China/NSCS (28th)

62,640 cores (HPC2) in Italy (11th) 72,288 cores (Yellowstone) at NCAR (29th)

147, 456 cores (Super MUC) in Germany (12th) 70,560 cores (Helios) at Japan/IFERC (30th)

76,032 cores (Tsubame 2.5) at Japan/GSIC (13th) 138,368 cores (Tera-100) at France/CEA (35th)

194,616 cores (Cascade) at PNNL (15th) 222,072 cores (QUARTETTO) in Japan (37th)

110,400 cores (Pangea) at France/Total (16th) 53,504 cores (PRIMERGY) in Australia (38th)

96,192 cores (Pleiades) at NASA/Ames (21st) 77,520 cores (Conte) at Purdue University (39th)

73,584 cores (Spirit) at USA/Air Force (24th) 44,520 cores (Spruce A) at AWE in UK (40th)

77,184 cores (Curie thin nodes) at France/CEA (26h) 48,896 cores (MareNostrum) at Spain/BSC (41st)

65,320-cores, iDataPlex DX360M4 at Germany/Max-Planck (27th)

and many more!

ISC Big Data '14 7

http://www.top500.org/

ISC Big Data '14

Kernel Space

Interconnects and Protocols in the Open Fabrics Stack Application / Middleware

Verbs

Ethernet Adapter

Ethernet Switch

Ethernet Driver

TCP/IP

1/10/40 GigE

InfiniBand

Adapter

InfiniBand

Switch

IPoIB

IPoIB

Ethernet Adapter

Ethernet Switch

Hardware Offload

TCP/IP

10/40 GigE-TOE

InfiniBand

Adapter

InfiniBand

Switch

User Space

RSockets

RSockets

iWARP Adapter

Ethernet Switch

TCP/IP

User Space

iWARP

RoCE Adapter

Ethernet Switch

RDMA

User Space

RoCE

InfiniBand

Switch

InfiniBand

Adapter

RDMA

User Space

IB Native

Sockets

Application / Middleware Interface

Protocol

Adapter

Switch

InfiniBand

Adapter

InfiniBand

Switch

RDMA

SDP

SDP

8

• Challenges for Accelerating Big Data Processing

• The High-Performance Big Data (HiBD) Project

• RDMA-based designs for Apache Hadoop and Spark

– Case studies with HDFS, MapReduce, and Spark

– RDMA-based MapReduce on HPC Clusters with Lustre

• Ongoing and Future Activities

• Conclusion and Q&A

Presentation Outline

ISC Big Data '14 9

Wide Adoption of RDMA Technology

• Message Passing Interface (MPI) for HPC

• Parallel File Systems

– Lustre

– GPFS

• Delivering excellent performance:

– < 1.0 microsec latency

– 100 Gbps bandwidth

– 5-10% CPU utilization

• Delivering excellent scalability

ISC Big Data '14 10

MVAPICH2/MVAPICH2-X Software

11

• High Performance open-source MPI Library for InfiniBand, 10Gig/iWARP, and RDMA

over Converged Enhanced Ethernet (RoCE)

– MVAPICH (MPI-1), MVAPICH2 (MPI-2.2 and MPI-3.0), Available since 2002

– MVAPICH2-X (MPI + PGAS), Available since 2012

– Support for GPGPUs and MIC

– Used by more than 2,225 organizations in 73 countries

– More than 223,000 downloads from OSU site directly

– Empowering many TOP500 clusters

• 7th, 519,640-core (Stampede) at TACC

• 13th, 74,358-core (Tsubame 2.5) at Tokyo Institute of Technology

• 23rd, 96,192-core (Pleiades) at NASA and many others . . .

– Available with software stacks of many IB, HSE, and server vendors including

Linux Distros (RedHat and SuSE)

– http://mvapich.cse.ohio-state.edu

• Partner in the U.S. NSF-TACC Stampede System

ISC Big Data '14

http://mvapich.cse.ohio-state.edu/

ISC Big Data '14

Latency & Bandwidth: MPI over IB with MVAPICH2

0,00

1,00

2,00

3,00

4,00

5,00

6,00 Qlogic-DDR Qlogic-QDR ConnectX-DDR ConnectX2-PCIe2-QDR ConnectX3-PCIe3-FDR Sandy-ConnectIB-DualFDR Ivy-ConnectIB--DualFDR

Small Message Latency

Message Size (bytes)

Late

ncy

(us)

1.66

1.56

1.64

1.82

0.99 1.09

1.12

DDR, QDR - 2.4 GHz Quad-core (Westmere) Intel PCI Gen2 with IB switch FDR - 2.6 GHz Octa-core (SandyBridge) Intel PCI Gen3 with IB switch

ConnectIB-Dual FDR - 2.6 GHz Octa-core (SandyBridge) Intel PCI Gen3 with IB switch ConnectIB-Dual FDR - 2.8 GHz Deca-core (IvyBridge) Intel PCI Gen3 with IB switch

12

0

2000

4000

6000

8000

10000

12000

14000 Unidirectional Bandwidth

Band

wid

th (M

Byte

s/se

c)

Message Size (bytes)

3280

3385

1917 1706

6343

12485

12810

Designing Communication and I/O Libraries for Big Data Systems: Challenges

ISC Big Data '14

Big Data Middleware (HDFS, MapReduce, HBase, Spark and Memcached)

Networking Technologies (InfiniBand, 1/10/40GigE

and Intelligent NICs)

Storage Technologies

(HDD and SSD)

Programming Models (Sockets)

Applications

Commodity Computing System Architectures

(Multi- and Many-core architectures and accelerators)

Other Protocols?

Communication and I/O Library

Point-to-Point Communication

QoS

Threaded Models and Synchronization

Fault-Tolerance I/O and File Systems

Virtualization

Benchmarks

13

Can Big Data Processing Systems be Designed with High-Performance Networks and Protocols?

Current Design

Application

Sockets

1/10 GigE Network

• Sockets not designed for high-performance – Stream semantics often mismatch for upper layers

– Zero-copy not available for non-blocking sockets

Our Approach

Application

OSU Design

10 GigE or InfiniBand

Verbs Interface

ISC Big Data '14 14

• Two strategies – RDMA-based MapReduce/Spark over RDMA-based HDFS

– RDMA-based MapReduce over Lustre









ISC Big Data '14 15

• RDMA for Apache Hadoop 2.x (RDMA-Hadoop-2.x)

• RDMA for Apache Hadoop 1.x (RDMA-Hadoop)

• RDMA for Memcached (RDMA-Memcached)

• OSU HiBD-Benchmarks (OHB)

• http://hibd.cse.ohio-state.edu

• Users Base: 76 organizations from 13 countries

• RDMA for Apache HBase and Spark

The High-Performance Big Data (HiBD) Project

ISC Big Data '14 16

http://hibd.cse.ohio-state.edu

• High-Performance Design of Hadoop over RDMA-enabled Interconnects

– High performance design with native InfiniBand and RoCE support at the verbs-

level for HDFS, MapReduce, and RPC components

– Easily configurable for native InfiniBand, RoCE and the traditional sockets-

based support (Ethernet and InfiniBand with IPoIB)

• Current release: 0.9.9/0.9.1

– Based on Apache Hadoop 1.2.1/2.4.1

– Compliant with Apache Hadoop 1.2.1/2.4.1 APIs and applications

– Tested with

• Mellanox InfiniBand adapters (DDR, QDR and FDR)

• RoCE support with Mellanox adapters

• Various multi-core platforms

• Different file systems with disks and SSDs

– http://hibd.cse.ohio-state.edu

RDMA for Apache Hadoop 1.x/2.x Distributions

ISC Big Data '14 17

http://hadoop-rdma.cse.ohio-state.edu/









ISC Big Data '14 18

• RDMA-based Designs and Performance Evaluation – HDFS

– MapReduce

– Spark

Acceleration Case Studies and In-Depth Performance Evaluation

ISC Big Data '14 19

Design Overview of HDFS with RDMA

• Enables high performance RDMA communication, while supporting traditional socket interface

• JNI Layer bridges Java based HDFS with communication library written in native code

HDFS

Verbs

RDMA Capable Networks (IB, 10GE/ iWARP, RoCE ..)

Applications

1/10 GigE, IPoIB Network

Java Socket Interface

Java Native Interface (JNI)

Write Others

OSU Design

• Design Features – RDMA-based HDFS

write

– RDMA-based HDFS replication

– Parallel replication support

– On-demand connection setup

– InfiniBand/RoCE support

ISC Big Data '14 20

Communication Times in HDFS

• Cluster with HDD DataNodes

– 30% improvement in communication time over IPoIB (QDR)

– 56% improvement in communication time over 10GigE

• Similar improvements are obtained for SSD DataNodes

Reduced by 30%

ISC Big Data '14

0

5

10

15

20

25

2GB 4GB 6GB 8GB 10GB

Com

mun

icat

ion

Tim

e (s

)

File Size (GB)

10GigE IPoIB (QDR) OSU-IB (QDR)

21

N. S. Islam, M. W. Rahman, J. Jose, R. Rajachandrasekar, H. Wang, H. Subramoni, C. Murthy and D. K. Panda , High Performance RDMA-Based Design of HDFS over InfiniBand , Supercomputing (SC), Nov 2012

N. Islam, X. Lu, W. Rahman, and D. K. Panda, SOR-HDFS: A SEDA-based Approach to Maximize Overlapping in RDMA-Enhanced HDFS, HPDC '14, June 2014

ISC Big Data '14

Evaluations using Enhanced DFSIO of Intel HiBench on TACC-Stampede

• Cluster with 64 DataNodes (1K cores), single HDD per node

– 64% improvement in throughput over IPoIB (FDR) for 256GB file size

– 37% improvement in latency over IPoIB (FDR) for 256GB file size

0

200

400

600

800

1000

1200

1400

64 128 256

Agg

rega

ted

Thro

ughp

ut (M

Bps)

File Size (GB)

IPoIB (FDR) OSU-IB (FDR) Increased by 64%

Reduced by 37%

0

100

200

300

400

500

600

64 128 256

Exec

utio

n Ti

me

(s)

File Size (GB)

IPoIB (FDR)

OSU-IB (FDR)

22


– MapReduce

– Spark


ISC Big Data '14 23

Design Overview of MapReduce with RDMA

MapReduce

Verbs

RDMA Capable Networks

(IB, 10GE/ iWARP, RoCE ..)

OSU Design

Applications

1/10 GigE, IPoIB Network

Java Socket Interface

Java Native Interface (JNI)

Job

Tracker

Task

Tracker

Map

Reduce

ISC Big Data '14


• JNI Layer bridges Java based MapReduce with communication library written in native code

• Design Features – RDMA-based shuffle

– Prefetching and caching map output

– Efficient Shuffle Algorithms

– In-memory merge

– On-demand Shuffle Adjustment

– Advanced overlapping • map, shuffle, and merge

• shuffle, merge, and reduce

– On-demand connection setup


24

Advanced Overlapping among different phases

• A hybrid approach to achieve

maximum possible overlapping

in MapReduce across all phases

compared to other approaches – Efficient Shuffle Algorithms

– Dynamic and Efficient Switching

– On-demand Shuffle Adjustment

Default Architecture

Enhanced Overlapping

Advanced Overlapping

M. W. Rahman, X. Lu, N. S. Islam, and D. K. Panda, HOMR: A Hybrid Approach to Exploit Maximum Overlapping in MapReduce over High Performance Interconnects, ICS, June 2014.

ISC Big Data '14 25

• For 240GB Sort in 64 nodes (512

cores)

– 40% improvement over IPoIB (QDR)

with HDD used for HDFS

Performance Evaluation of Sort and TeraSort

ISC Big Data '14

0

200

400

600

800

1000

1200

Data Size: 60 GB Data Size: 120 GB Data Size: 240 GB

Cluster Size: 16 Cluster Size: 32 Cluster Size: 64

Job

Exec

utio

n Ti

me

(sec

)

IPoIB (QDR) UDA-IB (QDR) OSU-IB (QDR)

Sort in OSU Cluster

0

100

200

300

400

500

600

700

Data Size: 80 GB Data Size: 160 GB Data Size: 320 GB

Cluster Size: 16 Cluster Size: 32 Cluster Size: 64

Job

Exec

utio

n Ti

me

(sec

)

IPoIB (FDR) UDA-IB (FDR) OSU-IB (FDR)

TeraSort in TACC Stampede

• For 320GB TeraSort in 64 nodes

(1K cores)

– 38% improvement over IPoIB

(FDR) with HDD used for HDFS 26

• 50% improvement in Self Join over IPoIB (QDR) for 80 GB data size

• 49% improvement in Sequence Count over IPoIB (QDR) for 30 GB data size

Evaluations using PUMA Workload

ISC Big Data '14

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

AdjList (30GB) SelfJoin (80GB) SeqCount (30GB) WordCount (30GB) InvertIndex (30GB)

Nor

mal

ized

Exe

cuti

on T

ime

Benchmarks

10GigE IPoIB (QDR) OSU-IB (QDR)

27


– MapReduce

– Spark


ISC Big Data '14 28

Design Overview of Spark with RDMA

• Design Features – RDMA based shuffle

– SEDA-based plugins

– Dynamic connection management and sharing

– Non-blocking and out-of-order data transfer

– Off-JVM-heap buffer management


ISC Big Data '14


• JNI Layer bridges Scala based Spark with communication library written in native code

X. Lu, M. W. Rahman, N. Islam, D. Shankar, and D. K. Panda, Accelerating Spark with RDMA for Big Data Processing: Early Experiences, Int'l Symposium on High Performance Interconnects (HotI'14), August 2014

29

ISC Big Data '14

Preliminary Results of Spark-RDMA Design - GroupBy

0

1

2

3

4

5

6

7

8

9

4 6 8 10

Gro

upBy

Tim

e (s

ec)

Data Size (GB)

0

1

2

3

4

5

6

7

8

9

8 12 16 20

Gro

upBy

Tim

e (s

ec)

Data Size (GB)

10GigE

IPoIB

RDMA

Cluster with 4 HDD Nodes, GroupBy with 32 cores Cluster with 8 HDD Nodes, GroupBy with 64 cores

• Cluster with 4 HDD Nodes, single disk per node, 32 concurrent tasks

– 18% improvement over IPoIB (QDR) for 10GB data size

• Cluster with 8 HDD Nodes, single disk per node, 64 concurrent tasks

– 20% improvement over IPoIB (QDR) for 20GB data size

30









ISC Big Data '14 31

Optimized Apache Hadoop over Parallel File Systems

MetaData Servers

Object Storage Servers

Compute Nodes TaskTracker

Map Reduce

Lustre Client

Lustre Setup

• HPC Cluster Deployment – Hybrid topological solution of Beowulf

architecture with separate I/O nodes – Lean compute nodes with light OS; more

memory space; small local storage – Sub-cluster of dedicated I/O nodes with

parallel file systems, such as Lustre • MapReduce over Lustre

– Local disk is used as the intermediate data directory

– Lustre is used as the intermediate data directory

ISC Big Data '14 32

• For 500GB Sort in 64 nodes

– 44% improvement over IPoIB (FDR)

Case Study - Performance Improvement of RDMA-MapReduce over Lustre on TACC-Stampede

ISC Big Data '14



0

200

400

600

800

1000

1200

300 400 500

Job

Exec

utio

n Ti

me

(sec

)

Data Size (GB)

IPoIB (FDR)

OSU-IB (FDR)

0

50

100

150

200

250

300

350

400

450

500

20 GB 40 GB 80 GB 160 GB 320 GB 640 GB

Cluster: 4 Cluster: 8 Cluster: 16 Cluster: 32 Cluster: 64 Cluster: 128

Job

Exec

utio

n Ti

me

(sec

)

IPoIB (FDR) OSU-IB (FDR)

M. W. Rahman, X. Lu, N. S. Islam, R. Rajachandrasekar, and D. K. Panda, MapReduce over Lustre: Can RDMA-based Approach Benefit?, Euro-Par, August 2014.

• Local disk is used as the intermediate data directory

33



Case Study - Performance Improvement of RDMA-MapReduce over Lustre on TACC-Stampede

ISC Big Data '14



• Lustre is used as the intermediate data directory

0

20

40

60

80

100

120

140

160

180

200

80 120 160

Job

Exec

utio

n Ti

me

(sec

)

Data Size (GB)


0

50

100

150

200

250

300

80 GB 160 GB 320 GB

Cluster: 8 Cluster: 16 Cluster: 32 Jo

b Ex

ecut

ion

Tim

e (s

ec)


• Can more optimizations be achieved by leveraging more features of Lustre? 34









ISC Big Data '14 35

Big Data Middleware (HDFS, MapReduce, HBase, Spark and Memcached)

Networking Technologies (InfiniBand, 1/10/40GigE

and Intelligent NICs)

Storage Technologies

(HDD and SSD)

Programming Models (Sockets)

Applications

Commodity Computing System Architectures

(Multi- and Many-core architectures and accelerators)

Other Protocols?

Communication and I/O Library

Point-to-Point Communication

QoS

Threaded Models and Synchronization

Fault-Tolerance I/O and File Systems

Virtualization

Benchmarks

RDMA Protocols

ISC Big Data '14 36

Applications-Level

Benchmarks

Micro-Benchmarks

Upper level Changes?

Designing Communication and I/O Libraries for Big Data Systems: Solved a Few Initial Challenges

37

• TestDFSIO on TACC Stampede

– 2.6x improvement over Default-HDFS for 160GB file size on 32 nodes

• RandomWriter on TACC Stampede

– 38% improvement over Default-HDFS for 80GB file size on 32 nodes

TestDFSIO RandomWriter

Preliminary Results of MEM-HDFS Design

0

100

200

300

400

500

600

8:20GB 16:40GB 32:80GB

Exec

utio

n Ti

me

(s)

Cluster Size: Data Size

Default-HDFS MEM-HDFS

0

200

400

600

800

1000

1200

1400

8:40GB 16:80GB 32:160GB

Agg

rega

ted

Thro

ughp

ut (M

Bps)

Cluster Size: Data Size

Default-HDFS MEM-HDFS

Increased by 2.6x

Reduced by 38%

N. Islam, X. Lu, W. Rahman, R. Rajachandrasekar, and D. K. Panda, MEM-HDFS: In-Memory I/O and Replication for HDFS with Memcached: Early Experiences, IEEE Big Data '14, October 2014

ISC Big Data '14

• In-memory I/O and replication for HDFS with Memcached

• Upcoming Releases of RDMA-enhanced Packages will support

– Hadoop 2.x MapReduce & RPC

– Spark

– HBase

• Upcoming Releases of OSU HiBD Micro-Benchmarks (OHB) will

support

– HDFS

– MapReduce

– RPC

• Advanced designs with upper-level changes and optimizations

• E.g. MEM-HDFS

ISC Big Data '14

Future Plans of OSU High Performance Big Data (HiBD) Project

38

• Discussed challenges in accelerating Hadoop and Spark

• Presented initial designs to take advantage of InfiniBand/RDMA

for Hadoop and Spark

• Results are promising

• Many other open issues need to be solved

• Will enable Big Data management and processing community to

take advantage of modern HPC technologies to carry out their

analytics in a fast and scalable manner

ISC Big Data '14

Concluding Remarks

39

[email protected]

ISC Big Data '14

Thank You!

The High-Performance Big Data Project http://hibd.cse.ohio-state.edu/

40

Network-Based Computing Laboratory http://nowlab.cse.ohio-state.edu/

The MVAPICH2/MVAPICH2-X Project http://mvapich.cse.ohio-state.edu/

http://nowlab.cse.ohio-state.edu/

http://nowlab.cse.ohio-state.edu/

Performance Optimization of Hadoop Using InfiniBand RDMA

Technology

Transcript of Performance Optimization of Hadoop Using InfiniBand RDMA