1. The document examines how applying different design patterns impacts energy usage. It measured the energy usage of applications before and after applying 15 design patterns. 2. The results showed that applying a design pattern always impacted energy usage, though the impact was small for some patterns. The impact also varied inconsistently within the same design pattern category. 3. Due to this inconsistency, it remains unclear how to predict the energy impact of a design pattern at the design level alone. Further work is needed on power monitoring tools and infrastructure to better understand these relationships.

1. INITIAL EXPLORATIONS
ON DESIGN PATTERN
ENERGY USAGE
Cagri Sahin
James Clause
Lori Pollock
Furkan Cayci
Irene Lizeth Manotas Gutiérrez
Fouad Kiamilev
Kristina Winbladh
Computer and Information
Sciences Department
Electrical and Computer
Engineering Department
University of Delaware
This work was supported in part by an award from the University of Delaware Research Foundation (UDRF).

2. MOTIVATION
Historically, software developers have left concerns
about power consumption to lower-level engineers.

3. MOTIVATION
Historically, software developers have left concerns
about power consumption to lower-level engineers.
Hardware
cpu, disk, etc.

4. MOTIVATION
Historically, software developers have left concerns
about power consumption to lower-level engineers.
Hardware
cpu, disk, etc.
Operating
systems

5. MOTIVATION
Historically, software developers have left concerns
about power consumption to lower-level engineers.
Hardware
cpu, disk, etc.
Operating
systems
Compilers

6. MOTIVATION
While this strategy has been successful, it is likely that
encouraging software developers to participate in the
process can result in more efficient applications.
Historically, software developers have left concerns
about power consumption to lower-level engineers.
Hardware
cpu, disk, etc.
Operating
systems
Compilers

7. Power
profile(w)
Design
properties
Design
artifacts
foo:25
foo:46
Robustness ReliabilityMaintainabilityFault toleranceAvailabilityPerformance
Actor
System
Use Case
Use cases
Operation
Attribute
Class Name
*1
Operation
Attribute
Class Name
<<extends>>
Operation
Attribute
Class Name
Class diagrams
Object Name Object Name
Synchronous
Simple Message
Sequence diagrams
OVERALL APPROACH

8. Power
profile(w)
Design
properties
Design
artifacts
foo:25
foo:46
Robustness ReliabilityMaintainabilityFault toleranceAvailabilityPerformance
Actor
System
Use Case
Use cases
Operation
Attribute
Class Name
*1
Operation
Attribute
Class Name
<<extends>>
Operation
Attribute
Class Name
Class diagrams
Object Name Object Name
Synchronous
Simple Message
Sequence diagrams
OVERALL APPROACH

9. EXPERIMENTAL DATATable I
ENERGY USAGE MEASUREMENT OBTAINED BY RUNNING THE BEFORE AND AFTER VERSIONS OF THE DESIGN PATTERNS.
Design pattern # Iterations Energy usage # Objects # Messages
Before (J) After (J) Difference (J) Difference per
iteration (J)
% Change Before After Before After
Creational
Abstract factory 500 87.78 106.69 18.91 0.0378 21.55 11 13 7 12
Builder 750 111.75 113.08 1.33 0.0018 1.19 3 6 18 32
Factory method 500 118.13 118.06 0.08 0.0002 0.07 3 3 1 2
Prototype 750 99.60 98.68 0.93 0.0012 0.93 7 7 3 9
Singleton 250 98.70 99.11 0.42 0.0017 0.42 2 2 7 12
Structural
Bridge 35 99.78 99.54 0.24 0.0070 0.24 6 11 3 6
Composite 175 97.15 102.14 4.99 0.0285 5.14 17 19 10 11
Decorator 115 13.92 113.13 99.21 0.8627 712.89 14 24 14 15
Flyweight 500 92.89 38.94 53.95 0.1079 58.08 60 6 60 60
Proxy 500 104.33 66.28 38.05 0.0761 36.47 5 2 3 6
Behavioral
Command 750 98.32 96.53 1.79 0.0024 1.82 7 7 29 29
Mediator 250 120.55 109.02 11.53 0.0461 9.56 4 5 23 26
Observer 400 61.62 99.95 38.33 0.0958 62.20 3 7 4 8
Strategy 500 115.73 115.52 0.21 0.0004 0.18 4 3 9 12
Visitor 90 104.89 97.04 7.86 0.0873 7.49 10 14 16 26
Joules of the application before and after applying each
design pattern. The fifth column, Difference, shows the
difference in total energy usage between the before and after
versions of applying the design patterns. Positive numbers
indicate that applying the design pattern increased energy
usage, and negative numbers indicate that applying the
design pattern reduced energy usage. The next column in
the table, Difference per iteration, shows the difference in
total energy usage per iteration (i.e., the difference in the
total energy usage divided by the number of iterations).
Again, positive numbers indicate that the applying the design
relatively small differences in the amounts of energy used
(e.g., the energy usage of the before and after versions of
the application for factory method differ only by 0.08 J).
The results shown in the table also support our assumption
that applying a design pattern impacts energy usage. For all
15 design patterns, we detected a difference in the amount
of energy used between the before and after versions of the
application.
E. RQ2: Impact
Our motivation for investigating the impact on energy

10. EXPERIMENTAL DATATable I
ENERGY USAGE MEASUREMENT OBTAINED BY RUNNING THE BEFORE AND AFTER VERSIONS OF THE DESIGN PATTERNS.
Design pattern # Iterations Energy usage # Objects # Messages
Before (J) After (J) Difference (J) Difference per
iteration (J)
% Change Before After Before After
Creational
Abstract factory 500 87.78 106.69 18.91 0.0378 21.55 11 13 7 12
Builder 750 111.75 113.08 1.33 0.0018 1.19 3 6 18 32
Factory method 500 118.13 118.06 0.08 0.0002 0.07 3 3 1 2
Prototype 750 99.60 98.68 0.93 0.0012 0.93 7 7 3 9
Singleton 250 98.70 99.11 0.42 0.0017 0.42 2 2 7 12
Structural
Bridge 35 99.78 99.54 0.24 0.0070 0.24 6 11 3 6
Composite 175 97.15 102.14 4.99 0.0285 5.14 17 19 10 11
Decorator 115 13.92 113.13 99.21 0.8627 712.89 14 24 14 15
Flyweight 500 92.89 38.94 53.95 0.1079 58.08 60 6 60 60
Proxy 500 104.33 66.28 38.05 0.0761 36.47 5 2 3 6
Behavioral
Command 750 98.32 96.53 1.79 0.0024 1.82 7 7 29 29
Mediator 250 120.55 109.02 11.53 0.0461 9.56 4 5 23 26
Observer 400 61.62 99.95 38.33 0.0958 62.20 3 7 4 8
Strategy 500 115.73 115.52 0.21 0.0004 0.18 4 3 9 12
Visitor 90 104.89 97.04 7.86 0.0873 7.49 10 14 16 26
Joules of the application before and after applying each
design pattern. The fifth column, Difference, shows the
difference in total energy usage between the before and after
versions of applying the design patterns. Positive numbers
indicate that applying the design pattern increased energy
usage, and negative numbers indicate that applying the
design pattern reduced energy usage. The next column in
the table, Difference per iteration, shows the difference in
total energy usage per iteration (i.e., the difference in the
total energy usage divided by the number of iterations).
Again, positive numbers indicate that the applying the design
relatively small differences in the amounts of energy used
(e.g., the energy usage of the before and after versions of
the application for factory method differ only by 0.08 J).
The results shown in the table also support our assumption
that applying a design pattern impacts energy usage. For all
15 design patterns, we detected a difference in the amount
of energy used between the before and after versions of the
application.
E. RQ2: Impact
Our motivation for investigating the impact on energy
1. Design patterns do impact energy consumption.

11. EXPERIMENTAL DATATable I
ENERGY USAGE MEASUREMENT OBTAINED BY RUNNING THE BEFORE AND AFTER VERSIONS OF THE DESIGN PATTERNS.
Design pattern # Iterations Energy usage # Objects # Messages
Before (J) After (J) Difference (J) Difference per
iteration (J)
% Change Before After Before After
Creational
Abstract factory 500 87.78 106.69 18.91 0.0378 21.55 11 13 7 12
Builder 750 111.75 113.08 1.33 0.0018 1.19 3 6 18 32
Factory method 500 118.13 118.06 0.08 0.0002 0.07 3 3 1 2
Prototype 750 99.60 98.68 0.93 0.0012 0.93 7 7 3 9
Singleton 250 98.70 99.11 0.42 0.0017 0.42 2 2 7 12
Structural
Bridge 35 99.78 99.54 0.24 0.0070 0.24 6 11 3 6
Composite 175 97.15 102.14 4.99 0.0285 5.14 17 19 10 11
Decorator 115 13.92 113.13 99.21 0.8627 712.89 14 24 14 15
Flyweight 500 92.89 38.94 53.95 0.1079 58.08 60 6 60 60
Proxy 500 104.33 66.28 38.05 0.0761 36.47 5 2 3 6
Behavioral
Command 750 98.32 96.53 1.79 0.0024 1.82 7 7 29 29
Mediator 250 120.55 109.02 11.53 0.0461 9.56 4 5 23 26
Observer 400 61.62 99.95 38.33 0.0958 62.20 3 7 4 8
Strategy 500 115.73 115.52 0.21 0.0004 0.18 4 3 9 12
Visitor 90 104.89 97.04 7.86 0.0873 7.49 10 14 16 26
Joules of the application before and after applying each
design pattern. The fifth column, Difference, shows the
difference in total energy usage between the before and after
versions of applying the design patterns. Positive numbers
indicate that applying the design pattern increased energy
usage, and negative numbers indicate that applying the
design pattern reduced energy usage. The next column in
the table, Difference per iteration, shows the difference in
total energy usage per iteration (i.e., the difference in the
total energy usage divided by the number of iterations).
Again, positive numbers indicate that the applying the design
relatively small differences in the amounts of energy used
(e.g., the energy usage of the before and after versions of
the application for factory method differ only by 0.08 J).
The results shown in the table also support our assumption
that applying a design pattern impacts energy usage. For all
15 design patterns, we detected a difference in the amount
of energy used between the before and after versions of the
application.
E. RQ2: Impact
Our motivation for investigating the impact on energy
1. Design patterns do impact energy consumption.
2. Impact within a category is inconsistent.

12. EXPERIMENTAL DATATable I
ENERGY USAGE MEASUREMENT OBTAINED BY RUNNING THE BEFORE AND AFTER VERSIONS OF THE DESIGN PATTERNS.
Design pattern # Iterations Energy usage # Objects # Messages
Before (J) After (J) Difference (J) Difference per
iteration (J)
% Change Before After Before After
Creational
Abstract factory 500 87.78 106.69 18.91 0.0378 21.55 11 13 7 12
Builder 750 111.75 113.08 1.33 0.0018 1.19 3 6 18 32
Factory method 500 118.13 118.06 0.08 0.0002 0.07 3 3 1 2
Prototype 750 99.60 98.68 0.93 0.0012 0.93 7 7 3 9
Singleton 250 98.70 99.11 0.42 0.0017 0.42 2 2 7 12
Structural
Bridge 35 99.78 99.54 0.24 0.0070 0.24 6 11 3 6
Composite 175 97.15 102.14 4.99 0.0285 5.14 17 19 10 11
Decorator 115 13.92 113.13 99.21 0.8627 712.89 14 24 14 15
Flyweight 500 92.89 38.94 53.95 0.1079 58.08 60 6 60 60
Proxy 500 104.33 66.28 38.05 0.0761 36.47 5 2 3 6
Behavioral
Command 750 98.32 96.53 1.79 0.0024 1.82 7 7 29 29
Mediator 250 120.55 109.02 11.53 0.0461 9.56 4 5 23 26
Observer 400 61.62 99.95 38.33 0.0958 62.20 3 7 4 8
Strategy 500 115.73 115.52 0.21 0.0004 0.18 4 3 9 12
Visitor 90 104.89 97.04 7.86 0.0873 7.49 10 14 16 26
Joules of the application before and after applying each
design pattern. The fifth column, Difference, shows the
difference in total energy usage between the before and after
versions of applying the design patterns. Positive numbers
indicate that applying the design pattern increased energy
usage, and negative numbers indicate that applying the
design pattern reduced energy usage. The next column in
the table, Difference per iteration, shows the difference in
total energy usage per iteration (i.e., the difference in the
total energy usage divided by the number of iterations).
Again, positive numbers indicate that the applying the design
relatively small differences in the amounts of energy used
(e.g., the energy usage of the before and after versions of
the application for factory method differ only by 0.08 J).
The results shown in the table also support our assumption
that applying a design pattern impacts energy usage. For all
15 design patterns, we detected a difference in the amount
of energy used between the before and after versions of the
application.
E. RQ2: Impact
Our motivation for investigating the impact on energy
1. Design patterns do impact energy consumption.
2. Impact within a category is inconsistent.
3. At the design level, it’s unclear how to predict the impact.

13. ISSUES / DISCUSSION
CONCLUSION

14. ISSUES / DISCUSSION
1. Power monitoring tools / infrastructure
CONCLUSION

15. ISSUES / DISCUSSION
1. Power monitoring tools / infrastructure
2. Experimental subjects
CONCLUSION

16. ISSUES / DISCUSSION
1. Power monitoring tools / infrastructure
2. Experimental subjects
3. Prediction metrics
CONCLUSION

17. ISSUES / DISCUSSION
1. Power monitoring tools / infrastructure
2. Experimental subjects
3. Prediction metrics
CONCLUSION
Questions?