The Implications from Benchmarking Three Big

Performance and Energy Efficiency
Evaluation of Big Data Systems
Presented by Yingjie Shi
Institute of Computing Technology, CAS
2013-10-31
Goals of Big Data Systems
Larger
Faster
Greener
BPOE 2013 | HPCChina 2013
Performance V.S. Energy Efficiency
Performance
Energy Efficiency
Faster & More
Powerful




Tradeoff
More servers
Bigger clusters
Powerful processors
Sophisticated
processing
algorithms
Evaluation
…
BPOE 2013 | HPCChina 2013
Greener &
Cheaper
 Lightweight servers
 Efficient processors
 Simpler processing
algorithms
…
Evaluation of Performance & Energy
Efficiency Tradeoff
How to measure?
AxPUE: Application Level Metrics for Power Usage Effectiveness
in Big Data Systems
How to get balance?
The Implications from Benchmarking Three Big Data Systems
BPOE 2013 | HPCChina 2013
Motivation
If you can not measure it, you can not improve it.
– Lord Kelvin
PUE(Power usage effectiveness): a measure of how
efficiently a computer data center uses its power;
specifically, how much of the power is actually used by
the information technology equipment.
BPOE 2013 | HPCChina 2013
PUE & Its Variants
Metric
Time
Organization
PUE
2007
GreenGrid
DCiE
DCeP
2008
2008
Computing Formulas
Total Facility Energy
IT Equipment Energy
IT Equipment Energy
*100%
Total Facility Energy
GreenGrid
GreenGrid
pPUE
2012
GreenGrid
PUE
Scalability
2013
GreenGrid
UsefulWork Pr oduced
Total Quantity of ResourceConsumed Producing this Work
BPOE 2013 | HPCChina 2013
Total Facility Energy insidethe Boundary
IT Equipment Energy insidethe Boundary
mActual
*100%
mPUE
Motivation
• Scenario1
An Improved Data Classification Algorithm
Does it contribute to greening the data centers?
Run the Algorithms on Data Center
Compare the PUEs
Data Management
PUE can Researcher
not measure
the effectiveness of any
changes made upon the data center infrastructure!
No Obvious Variations!
BPOE 2013 | HPCChina 2013
Motivation
• Scenario2
Give a budget plan of the data center energy
consumption in the next year
Estimate the data volume based on
the business development
Data Center Administrators
How to estimate the energy
increasement?
PUE provides little reference information for data
center planning according to data scale
and application complexity
BPOE 2013 | HPCChina 2013
Calculation Framework
AxPUE
PUE
BPOE 2013 | HPCChina 2013
Definition - ApPUE
• ApPUE (Application Performance Power Usage Effectiveness): a
metric that measures the power usage effectiveness of IT equipments,
specifically, how much of the power entering IT equipments is used to
improve the application performance.
• Computation Formulas:
Data processing performance of applications
ApPUE 
Application Performance
IT Equipment Power
The average rate of IT Equipment Energy consumed
BPOE 2013 | HPCChina 2013
Definition - AoPUE
• AoPUE (Application Overall Power Usage Effectiveness ): a metric
that measures the power usage effectiveness of the overall data
center system, specifically, how much of the total facility power is
used to improve the application performance.
• Computation Formulas:
AoPUE 
Application Performance
Total Facility Power
ApPUE
AoPUE 
PUE
The average rate of Total Facility Energy Used
BPOE 2013 | HPCChina 2013
Acquisition – Application Performance
Application
Category
Examples
Metric
Service Application
Search engine, Ad-hoc
queries
Number of requests
answered in unit time
Data Analysis
Application
Data mining, Reporting,
Decision support, Log analysis
Volume of data
processed in unit time
Interactive Real-time
Application
E-commerce, Profile data
management
Number of transactions
completed in unit time
High Performance
Computing
Scientific Computing
Number of floating-point
operations in unit time
BPOE 2013 | HPCChina 2013
Acquisition – Benchmark
• Requirements of Benchmarks
– Provide representative workloads for big data
applications
– Provide a scalable data generation tool
• BigDataBench
– A big data benchmark suite open-sourced recently
and publicly available
– All the requirements are well fullfilled
BPOE 2013 | HPCChina 2013
Experiment Overview
• Testbed
– Data center of 18 racks,362 servers
– Sample 8 servers
• Workloads
• Two experiments
– Different Applications
– Different Implementation Algorithms
BPOE 2013 | HPCChina 2013
Experiments on Different Applications
17.2 11.5
269.9 179.7
PUE
ApPUE
AoPUE
BigDataBench
SVM
Sort
Grep
BPOE 2013 | HPCChina 2013
Linpack
Experiments on Different Algorithms
• Two Implementations for Sort
– Several reducers with random sampling partitioning
– One reducer without partitioning
PUE(Sort1)
ApPUE(Sort1)
PUE(Sort2)
ApPUE(Sort2)
Data Size
BPOE 2013 | HPCChina 2013
Conclusions
• We analyze the requirements of application-level energy
effectiveness metrics AxPUE in data centers.
• We propose two novel application-level metrics ApPUE
and AoPUE to measure the energy consumed to improve
the application performance.
• The experiment results show that AxPUE could provide
meaningful guidance to data center design and
optimization.
BPOE 2013 | HPCChina 2013
Evaluation of Performance & Energy
Efficiency Tradeoff
How to measure?
AxPUE: Application Level Metrics for Power Usage Effectiveness
in Data Centers
How to get balance?
The Implications from Benchmarking Three Big Data Systems
BPOE 2013 | HPCChina 2013
New Solutions
……
BPOE 2013 | HPCChina 2013
Experimental Platforms
Xeon (Common processor)
Atom ( Low power processor)
Tilera
Brief Comparison
(ManyBasic
coreInformation
processor)
CPU Type
Intel Xeon
E5310
Intel Atom D510
Tilera TilePro36
CPU Core
4 cores @
1.6GHz
2 cores @
1.66GHz
36 cores @
500MHz
L1 I/D
Cache
32KB
24KB
16KB/8KB
L2 Cache
4096KB
512KB
64KB
BPOE 2013 | HPCChina 2013
Benchmark Selection
BigDataBench
A big data benchmark suite from big data applications
Respective applications
An innovative data generation tool
Application
Time
Complexity
Characteristics
Sort
O(n*log2n)
Integer comparison
WordCount
O(n)
Integer comparison and
calculation
Grep
O(n)
String comparison
Naïve Bayes
O(m*n)
Floating-point computation
SVM
O(n3)
Floating-point computation
BPOE 2013 | HPCChina 2013
Metrics
Performance: Data processed per second (DPS)
Energy Efficiency: Application Performance
Power Usage Effectiveness(DPJ)
BPOE 2013 | HPCChina 2013
General Observations
Xeon
Atom
DPS
DPJ
BPOE 2013 | HPCChina 2013
Tilera
General Observations
Xeon
Atom
Tilera
Data scale has a significant impact on the performance and energy efficiency of big data
systems.
The performance and energy efficiency trends of different applications are diverse.
BPOE 2013 | HPCChina 2013
Xeon VS Atom – DPS
BPOE 2013 | HPCChina 2013
Xeon VS Atom – DPJ
BPOE 2013 | HPCChina 2013
Xeon VS Atom – DPS & DPJ
500MB
1GB
10GB
25GB
50GB
100G
B
Sort
DPS
DPJ
3.67
0.87
4.51
1.08
1.89
0.45
1.54
0.36
1.36
0.32
1.40
0.33
Wordcount
DPS
DPJ
2.27
0.55
2.38
0.58
2.74
0.61
2.84
0.61
2.82
0.62
2.79
0.60
Grep
DPS
DPJ
1.83
0.48
1.82
0.46
2.30
0.54
2.79
0.62
2.87
0.63
2.89
0.64
Naïve
Bayes
DPS
DPJ
3.83
0.89
3.89
0.87
4.52
1.01
4.64
0.99
4.54
0.97
4.58
0.90
SVM
DPS
DPJ
3.19
0.69
3.06
0.64
3.17
0.66
3.14
0.67
Xeon is more powerful than Atom on processing capacity.
Atom is more energy –saving than Xeon when dealing with
simple computation logic applications.
BPOE 2013 | HPCChina 2013
Xeon VS Atom -- Summary
Xeon is more powerful than Atom on
processing capacity.
Atom is energy conservation than Xeon when
dealing with applications with simple
computation logic.
Atom doesn’t show energy advantage when
dealing with complex applications.
BPOE 2013 | HPCChina 2013
Xeon VS Tilera – DPS
BPOE 2013 | HPCChina 2013
Xeon VS Tilera – DPJ
BPOE 2013 | HPCChina 2013
Xeon VS Tilera – DPS & DPJ
500MB
1GB
10GB
25GB
Sort
DPS
DPJ
3.67
0.48
3.39
0.45
2.41
0.31
2.60
0.34
Wordcount
DPS
DPJ
5.19
0.67
5.04
0.65
7.35
0.87
7.78
0.92
Grep
DPS
DPJ
3.60
0.51
3.52
0.48
7.45
0.94
9.93
1.21
Naïve Bayes
DPS
DPJ
5.91
0.75
5.78
0.70
7.59
0.89
7.94
0.92
Xeon is more powerful than Tilera on processing capacity
Tilera is more energy-saving than Xeon when dealing with the
simple computation logic and I/O intensive applications
Tilera don’t show energy advantage when dealing with complex
applications
BPOE 2013 | HPCChina 2013
Xeon VS Tilera
DPSThe
of Atom
DPS of Tilera
The DPS of The
Xeon
BPOE 2013 | HPCChina 2013
Xeon VS Tilera
Tilera is more
suitable to process
I/O intensive
applications
The DPS of Tilera
BPOE 2013 | HPCChina 2013
Xeon VS Tilera -- Summary
Xeon is more powerful than Tilera on
processing capacity.
Tilera is more energy conservation than Xeon
when dealing with simple computation logic and
I/O intensive applications.
Tilera don’t show energy advantage when
dealing with complex applications.
Tilera is more suitable to process I/O intensive
applications.
BPOE 2013 | HPCChina 2013
36
Implications
The performance of a big data system is not only
related to the hardware itself, but also the application
type and data volume of workloads.
The weak processors aren’t suitable to deal with
complex applications. Even they have lower TDP,
they don’t show energy cost advantage.
BPOE 2013 | HPCChina 2013
Implications Cont.
Xeon generally has better processing capacity
accompanied with high energy consumption, especially
to some light scale-out applications.
Atom and Tilera show energy consumption advantage
when dealing with light scale-out applications.
Tilera exerts energy advantage on processing I/O
intensive application.
BPOE 2013 | HPCChina 2013
BPOE 2013 | HPCChina 2013