57
Smart Devices Collaboration for Energy Saving in Home Networks Han YAN Ph.D defense 19 December, 2014 Orange Labs, France IRISA, France Pr. Ye-Qiong SONG Pr. Mohamed Yacine GHAMRI DOUDANE Dr. Stéphane LOHIER Pr. Bernard COUSIN Dr. Cédric GUEGUEN Dr. Jean-Paul VUICHARD Jury

Soutenance han2014v3

  • Upload
    hanhy

  • View
    119

  • Download
    2

Embed Size (px)

Citation preview

Page 1: Soutenance han2014v3

Smart Devices Collaboration for Energy Saving in Home Networks

Han YAN Ph.D defense 19 December, 2014

Orange Labs, France IRISA, France

Pr. Ye-Qiong SONG

Pr. Mohamed Yacine GHAMRI DOUDANE

Dr. Stéphane LOHIER

Pr. Bernard COUSIN

Dr. Cédric GUEGUEN

Dr. Jean-Paul VUICHARD

Jury

Page 2: Soutenance han2014v3

2

08/01/2015

Why Green Home ?

Economic Reason [2] [3]:

• In 2012, electricity price increases 2% faster than inflation

• In 2014, the total annual cost of electricity is estimated at 1,775€. In 2020, it is

expected to reach 2,486 €

[1] IPCC. Climate change 2013: The physical science basis, 2013

[2] Christophe Dromacque and Anna Bogacka. European residential energy price report, 2013

[3] Vers electricite plus chere, ecosocioconso, 2014

Environmental Reason [1]:

• Information and Communication technology sector represents 2% CO2emssion

Page 3: Soutenance han2014v3

3

08/01/2015

Global energy consumption is increasing with a rhythm at 3% every year

Residential energy consumption has multiplied 5 times [4]

Twh

Energy Consumption Evolution in France

[4] Le bilan énergétique de la France pour 2010. Technical report, SOeS, 2010

Page 4: Soutenance han2014v3

4

08/01/2015

Lighting13%

Cooling23%

Audio video 20%

Washing 15%

Personal computer

15%

Others14%

Other Specific DevicesEnergy Consumption [5]

Power Consumption in Digital Home

More and more energy is consumed for the entertainment usage

Devices work no more alone, the peripheral devices spread rapidly[5] Statistics Explained Eurostat. Consumption of energy. Technical report, 2012

Page 5: Soutenance han2014v3

5

08/01/2015

Energy Saving Challenges in the Home Network

NAS

Laptop

HGW

STB

Pad

controller

PLC

plug

WANPLC plug

WiFi Ethernet

ADSL PLC

Devices should be used efficiently:

• In collaborative services

• Energy consumption efficiency

Turned on / off efficiently

User is important in the home network

• Increasing satisfaction

• Behavior learning

User satisfaction

Energy efficiency

WAN

Page 6: Soutenance han2014v3

6

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management

• Ethernet, memory etc.

Network level

• Low power network technologies

• Home automation energy control

• Network connection power control

• Power control elements

State of the Art in Power Management

Page 7: Soutenance han2014v3

7

08/01/2015

Device level

• Advanced configuration and power interface specification [6]

• Dynamic power management

• Ethernet, memory etc.

Network level

• Low power network technologies

• Home automation energy control

• Network connection power control

• Power control elements

State of the Art in Power Management

[6] ACPI Advanced configuration and power interface specification

Page 8: Soutenance han2014v3

8

08/01/2015

Advanced Configuration and Power Interface

Page 9: Soutenance han2014v3

9

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management [7]-[9]

• Ethernet, memory etc.

Network level

• Low power network technologies

• Home automation energy control

• Network connection power control

• Power control elements

State of the Art in Power Management

[7] Trevor Pering, Tom Burd, and Robert Brodersen, The simulation and evaluation of dynamic voltage scaling algorithms,1998 international

symposium on Low power electronics and design, pages 76--81. ACM, 1998

[8] Padmanabhan Pillai and Kang G Shin. Real-time dynamic voltage scaling for low-power embedded operating systems. In ACM SIGOPS

Operating Systems Review, 35, pages 89–102. ACM, 2001

[9] Dirk Grunwald, Charles B Morrey III, Philip Levis, Michael Neufeld, and Keith I Farkas. Policies for dynamic clock scheduling. In

Proceedings of the 4th conference on Symposium on Operating System Design & Implementation-Volume 4, pages 6–6. USENIX

Association, 2000

Page 10: Soutenance han2014v3

10

08/01/2015

Dynamic Power Management

𝑃 ∝ 𝐶𝑉2𝑓 [10]

Capacitance

Voltage Frequency

Dissipated power

The voltage or the frequency or both parameters could reduce the power

consumption of the system:

• Dynamic voltage scaling power management [7] [8]

• Dynamic frequency scaling power management [9]

[7] Trevor Pering, Tom Burd, and Robert Brodersen, The simulation and evaluation of dynamic voltage scaling algorithms,1998 international

symposium on Low power electronics and design, pages 76--81. ACM, 1998

[8] Padmanabhan Pillai and Kang G Shin. Real-time dynamic voltage scaling for low-power embedded operating systems. In ACM SIGOPS

Operating Systems Review, 35, pages 89–102. ACM, 2001

[9] Dirk Grunwald, Charles B Morrey III, Philip Levis, Michael Neufeld, and Keith I Farkas. Policies for dynamic clock scheduling. In

Proceedings of the 4th conference on Symposium on Operating System Design & Implementation-Volume 4, pages 6–6. USENIX

Association, 2000

[10] Neil Weste and David Harris. Cmos vlsi design. A Circuits and Systems perspective, Pearson Addison Wesley, 2005

Page 11: Soutenance han2014v3

11

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management

• Ethernet, memory etc. [11]-[13]

Network level

• Low power network technologies

• Home automation energy control

• Network connection power control

• Power control elements

State of the Art in Power Management

[11] Chamara Gunaratne and Ken Christensen. Ethernet adaptive link rate: System design and performance evaluation. In Local

Computer Networks, Proceedings 2006 31st IEEE Conference on, pages 28–35. IEEE, 2006

[12] Maruti Gupta and Suresh Singh. Dynamic ethernet link shutdown for energy conservation on ethernet links. In Communications

ICC’07. IEEE International Conference on, pages 6156–6161. IEEE, 2007

[13] Kiran Puttaswamy, Kyu-Won Choi, Jun Cheol Park, Vincent J Mooney III, Abhijit Chatterjee, and Peeter Ellervee. System level power-

performance trade-offs in embedded systems using voltage and frequency scaling of off-chip buses and memory. In Proceedings of the

15th international symposium on System Synthesis, pages 225–230. ACM, 2002.

Page 12: Soutenance han2014v3

12

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management

• Ethernet, memory etc.

Network level

• Low power network technologies [14]-[16]

• Home automation energy control

• Network connection power control

• Power control elements

State of the Art in Power Management

[14] Carles Gomez, Joaquim Oller, and Josep Paradells. Overview and evaluation of bluetooth low energy: An emerging low-power

wireless technology. Sensors, 12[9]:11734–11753, 2012

[15] ZigBee Alliance. Zigbee specifications, 2008

[16] Zach Shelby and Carsten Bormann. 6LoWPAN: The wireless embedded Internet, 43. John Wiley & Sons, 2011.

Page 13: Soutenance han2014v3

13

08/01/2015

Low Power TechnologiesTechnologies ZigBee/

6LoWPAN(over 802.15.4)

Bluetooth LowEnergy

WiFi

IEEE spec 802.15.4 802.15.1 802.11 a/b/g

FrequencyBand

868/915 MHz;2.4 GHz

2.4 GHz 2.4 GHz;5 GHz

NominalRange

10-100 m 10 m 100 m

Chipset cc2531 cc2540 cx53111

RX 25 mA 19.6 mA 219 mA

TX 34 mA 31.6 mA 215 mA

Page 14: Soutenance han2014v3

14

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management

• Ethernet, memory etc.

Network level

• Low Power network technologies

• Home automation energy control[17][18]

• Network connection power control

• Power control elements

State of the Art in Power Management

[17] Han and Jae-Hyun Lim. Design and implementation of smart home energy management systems based on zigbee. Consumer

Electronics, IEEE Transactions on, 56[3]:1417-1425, 2010

[18] Il-Kyu Hwang, Dae-sung Lee, and Jin-wook Baek. Home network conguring scheme for all electric appliances using zigbee-based

integrated remote controller. Consumer Electronics, IEEE Transactions on,55[3]:1300{1307, 2009

Page 15: Soutenance han2014v3

15

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management

• Ethernet, memory etc.

Network level

• Low Power network technologies

• Home automation energy control

• Network connection power control [19][20]

• Power control elements

State of the Art in Power Management

[19] Olivier Bouchet, Abdesselem Kortebi, and Mathieu Boucher. Inter-mac green path selection for heterogeneous networks. In

Globecom Workshops (GC Wkshps), 2012 IEEE, pages 487-491. IEEE, 2012

[20] Vincenzo Suraci, Alessio Cimmino, Roberto Colella, Guido Oddi, and Marco Castrucci. Convergence in home gigabit networks:

implementation of the inter-mac layer as a pluggable kernel module. In Personal Indoor and Mobile Radio Communications (PIMRC),

2010 IEEE 21st International Symposium on, pages 2569{2574. IEEE, 2010

Page 16: Soutenance han2014v3

16

08/01/2015

Device level

• Advanced configuration and power interface specification

• Dynamic power management

• Ethernet, memory etc.

Network level

• Low Power network technologies

• Home automation energy control

• Network connection power control

• Power control elements [21]

State of the Art in Power Management

[21] Youn-Kwae Jeong, Intark Han, and Kwang-Roh Park. A network level power management for home network devices. Consumer

Electronics, IEEE Transactions on, 54[2]:487-493, 2008

Page 17: Soutenance han2014v3

17

08/01/2015

NAS

HGW

STB

PLC

plug

PLC plug

WiFi Ethernet ADSL PLC

Audio Video Use Case

Laptop

Devices work no more alone in home network

Devices are not used efficiently

Film selectionVideo Push

Film selection Video Push

Local network connection

time

Page 18: Soutenance han2014v3

18

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 19: Soutenance han2014v3

19

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 20: Soutenance han2014v3

20

08/01/2015

Power Consumption of Home Network DevicesPower consumption (Watt) of the home network devices

in different power states [22]:

States

Device

Working Idling Sleeping Soft-off

HGW 12.6 11 3.9 2

STB 21 19.2 13.5 2.5

Workstation 205 123.5 4.9 3.2

Laptop 79 54 5 2.5

PLC plug 6 3 2.6 0.15

A great power consumption difference from state to state

How to go to the low power consumption states?

[22] Yan, H, Vuichard, J, Cousin, B, Gueguen, C, and Mardon, G "Green Home Network based on an Overlay Energy Control Network"

in the book "Green Networking and Communications" , CRC Press, USA, oct 2013

Page 21: Soutenance han2014v3

21

08/01/2015

Working

Sleeping

Idling

Soft off

21 watt

19.2 watt

13.5 watt2.5 watt

An Example of STB Power StatesService execution time

Less than 4 hours

After 4 hours idling time

User controls manually

to turn on/off

No s

erv

ice

User s

erv

ice

request

• Long waiting time from idling to sleeping

• Explicit user commands are required to go to soft off state

• User controls manually to

go to sleeping state

• User service request to

turn on

Page 22: Soutenance han2014v3

22

08/01/2015

Overlay Energy Control Network

Proposition:

Controlling the devices by a ultra-low power consumption overlay network

The Overlay Energy Control Network (OECN) is formed by:

• Control nodes associated to each device

• OECN manager

Overlay Energy Control Network (OECN) can switche devices:

• From working or idling to sleeping power state much more quickly

• From working to soft off power state automatically

Page 23: Soutenance han2014v3

23

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 24: Soutenance han2014v3

24

08/01/2015

ZigBee Mandatory energy saving Solution (ZMS)

The advantages of ZMS:

• A device can be turned off and can also be started up by the associated ZigBee

module which is always on

• This device can go to a real low power consumption state (soft off state)

• All the energy control nodes are ZigBee

• Energy control messages are sent via ZigBee network

workstation

HGWSTB WANPLC plug

WiFi Ethernet ADSL PLC

Laptop

PLC plug

ZigBee

Page 25: Soutenance han2014v3

25

08/01/2015

Working

Sleeping

Idling

Soft off

Service execution time

Device stays in soft off state

while there is no service

• ZMS turns on device for

service execution

• ZMS turns off device while

there is no service

Device Power States Control Model by ZMS

Page 26: Soutenance han2014v3

26

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 27: Soutenance han2014v3

27

08/01/2015

• Device itself can become the energy control node

• Energy control messages could be sent via the data home network

workstation

HGWSTB WAN

WiFi Ethernet ADSL PLC

Laptop

ZigBee Optional energy saving Solution (ZOS)

The advantages of ZOS:

• It might not be possible to connect a ZigBee module on device

• ZigBee transmission diameter is limited, we have another alternative solution

ZigBee

PLC

plug

PLC

plug

Page 28: Soutenance han2014v3

28

08/01/2015

Working

Sleeping

Idling

Soft off

Service execution time

• ZOS switches device to working state

for service execution

• ZOS switches device to sleeping state

while there is no service

Device stays in sleeping state

while there is no service

Device Power States Control Model by ZOS

Page 29: Soutenance han2014v3

29

08/01/2015

33.5 21.3 22.4

735.3527.1 575.2

244.8

188.3197.6

475.5

425.2441.2

0

200

400

600

800

1000

1200

1400

1600

Self-controlled ZMS ZOS

Annual Power ConsumptionKwh

13.9 160 1612.3

436

4411.6

457

46.2

7.4

86

8

0

200

400

600

800

1000

1200

Self-controlled ZMS ZOS

Daily DelaySecond

Energy Consumption and Delay Resultsin 4 Types of Days

ZMS is more energy efficient; but ZMS has a relatively high delay

ZOS is a good tradeoff between the energy gain and delay

ZMS ZOS

21.97% 16.96%

Energy gain of two solutions

Page 30: Soutenance han2014v3

30

08/01/2015

Expense results

• ZMS and ZOS are profitable after 1.2 year

• ZMS is more profitable than ZOS after 1.6 year

Total expense is a sum of total ZigBee modules expense and

total electricity expense

Page 31: Soutenance han2014v3

31

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 32: Soutenance han2014v3

32

08/01/2015

UPnP device ZigBee

NAS

LaptopSTB

HGW

Pad

controller

PLC

plug

PLC plug

WiFi Ethernet ADSL PLC

HOme Power Efficiency system

HOme Power Efficiency (HOPE) system has two controlling methods:

• ZigBee on laptop, STB, PLC plugs and HGW

• UPnP Low Power on Pad and NAS

Page 33: Soutenance han2014v3

33

08/01/2015

UPnP Low Power

UPnP Low Power is proposed to implement different power saving modes to save

energy for devices

There are three types of elements:

• UPnP Low Power device

• UPnP Low Power power management proxy

• UPnP Low Power aware control point

Announcement:

• Power states

• Methods of waking

• Entry & exit information

UPnP Low Power device

UPnP Low Power proxy

UPnP Low Power

aware control point

Discovery the UPnP Low

Power devices with their

waking methods

• Monitor

• Send a “wake up”

or “go to sleep”

Page 34: Soutenance han2014v3

34

08/01/2015

Software Architecture of HOPE System

Page 35: Soutenance han2014v3

35

08/01/2015

Testbed for Home Power Efficiency system

Page 36: Soutenance han2014v3

36

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 37: Soutenance han2014v3

37

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 38: Soutenance han2014v3

38

08/01/2015

NAS

LaptopSTB

HGW

Pad controller

WiFi activation of

the HGW through

ZigBee

PLC plugs

Three Use Cases for HOPE system

Energy Gain :

2.9 Watt

NAS is woke up

through ZigBee and

Wake-On-Lan

NAS is switched off

through ZigBee and

UPnP Low Power

+ 23.4 Watt

Wake up

PLC plugs

through ZigBee

PLC plugs are

switched off

through ZigBee

+ 12 Watt

Page 39: Soutenance han2014v3

39

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 40: Soutenance han2014v3

40

08/01/2015

Definition of collaborative power management

Page 41: Soutenance han2014v3

41

08/01/2015

Audio Video Use Case

NAS

HGW

STB

PLC

plug

PLC plug

Laptop

Film selectionVideo Push

Film selection Video Push

Local network connection

Film selection Content directory function block

Content directory function block

Transfer server, content directory function blocks

Video stream decoder; display interface;

authentication; transfer client function blocks

Connection function block

Video Push

Local network connection

time

Page 42: Soutenance han2014v3

42

08/01/2015

Energy model

Operating

Idling

StartingDifferent phases while device is on

Page 43: Soutenance han2014v3

43

08/01/2015

Delay model

time

time

𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑡_𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒

𝑡_𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒

Operating

Idling

Starting

Page 44: Soutenance han2014v3

44

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 45: Soutenance han2014v3

45

08/01/2015

Refined overlay algorithmsTwo propositions based on the refined overlay algorithms:

• Refined Overlay Power Management (ROPM):

We assume that ROPM registered user habits of using collaborative

services

• Refined Overlay Auto Learning (ROAL):

ROAL learns the habits how user uses their collaborative services

The control decisions depend on:

• ROPM Pre-saved user habit information

• ROAL Learns user habit when they request services

time

time

𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑡_𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒

𝑡_𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒

Operating

Idling

Starting

Page 46: Soutenance han2014v3

46

08/01/2015

Video Push

Power Control Elements

Reference model:

• Devices are configured as power control elements

• All necessary power control elements are on at the beginning

of the service

NAS

HGW

STB

PLC

plug

PLC plug

Laptop

Film selectionVideo Push

Film selection

Local network connection

time

Page 47: Soutenance han2014v3

47

08/01/2015

Simulation Results on the Auto Learning Period

Simulation time varies from 10 hours to 500 hours

ROAL needs about 200 hours to learn an accurate value of the request time

Simulation time (h)

Re

qu

es

tti

me

(s

)

Page 48: Soutenance han2014v3

48

08/01/2015

Simulation Result on the Energy Gain

In a collaborative service, if several functional blocks are

needed lately, we gain more energy in these use cases

ROPM ROAL

41.85% 35.25%

Energy gain of two solutions

Page 49: Soutenance han2014v3

49

08/01/2015

Simulation Result on the Delay

• ROPM and ROAL have greater delay than the PCE, but

much more efficient in the term of energy saving

• ROAL has less delay than ROPM

How to be more accurate in the user habits learning?

Page 50: Soutenance han2014v3

50

08/01/2015

Contributions of the Thesis

• ZigBee mandatory energy saving solution

• ZigBee optional energy saving solution

An overlay network for energy control

• Architecture of HOPE system

• Three use cases

HOme Power Efficiency (HOPE) system

• Collaborative services analysis and definitions

• Refined overlay and auto learning power management

• Power delay tradeoffs

Collaborative power control management

Page 51: Soutenance han2014v3

51

08/01/2015

𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡_𝑙𝑒𝑎𝑟𝑛𝑒𝑑i,j = 𝑘=1𝑁𝑏_𝑠𝑒𝑟 𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑖,𝑗

𝑘

𝑁𝑏_𝑠𝑒𝑟

𝑡_𝑡𝑟𝑎𝑑𝑒𝑜𝑓𝑓 = 𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡_𝑙𝑒𝑎𝑟𝑛𝑒𝑑_𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒i,j × 𝛼

𝑡_𝑑𝑒𝑐_𝑜𝑛𝑖,𝑗𝑘 = 𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡_𝑙𝑒𝑎𝑟𝑛𝑒𝑑i,j − 𝑡_𝑡𝑟𝑎𝑑𝑒𝑜𝑓𝑓

Collaborative Overlay Power management Power-Delay Tradeoff Algorithm

The t_tradeoff will be configured by the user satisfaction requirement α

and the variance of the request (𝑡_𝑟𝑒𝑞𝑢𝑒𝑠𝑡_𝑙𝑒𝑎𝑟𝑛𝑒𝑑_𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒) as shown

in formula

Our proposition Collaborative Overlay Power management Power-Delay

Tradeoff (COPM-PDT α) algorithm varies by tradeoff coefficient α.

Page 52: Soutenance han2014v3

52

08/01/2015

Energy Delay Results

COPM-PDT 99 COPM-PDT 75 COPM-PDT 50

Energy gain 23.62% 31.29% 34.15%

Increased delay

0.5% 20.06% 43.75%

404

78.9 79.3 98.7 140.30

100200300400500

Delay (s)

135.8

97.874.7 67.2 64.4

0

50

100

150

Energy (Watt-hour)

Energy gain and delay for different parameters of α

Page 53: Soutenance han2014v3

53

08/01/2015

Energy Delay Results by tuning the tradeoff coefficient

Tradeoff Coefficient Tradeoff Coefficient

Dela

y per

serv

ice (

s)

Energ

yconsum

ption

per

serv

ice (

watthour)

Energy consumption

• Control algorithms based on the learned information

• Good energy efficiency and low waiting delay obtained

Delay

Page 54: Soutenance han2014v3

54

08/01/2015

Conclusion: Responses for Energy Saving in Home Network

We proposed an always-on low power network over the traditional network

• ZMS is energy efficient, high delay

• ZOS is a good tradeoff between the energy gain and the delay

A testbed of home power efficiency system is implemented

• Devices are turned on by service requests

• Energy efficiency with a high ease of use

Collaborative power management

• Collaborative service analysis

• User behavior learning

• Control algorithms based on the learned information

• Good energy efficiency and low waiting delay obtained

Page 55: Soutenance han2014v3

55

08/01/2015

Perspectives for Future Works

Short Term Perspectives:

• Different types of function blocks

• Resource allocation

• Collaborative analysis

• Analysis the tradeoffs between energy and other metrics

• User habits learning could be more intelligent and adaptive

Long Term Perspectives:

• A more heterogeneous overlay control network

• The collected information could be explored for other usages

Page 56: Soutenance han2014v3

56

08/01/2015

Page 57: Soutenance han2014v3

57

08/01/2015

List of Publications International papers with peer review:• Yan, H, Cousin, B, Gueguen, C and Vuichard, J. “Refined Overlay Power Management in the Home Environment” IEEE

GREENCOM '2014, Taipei, Taiwan• Yan, H, Gueguen, C, Cousin, B, and Vuichard, J. "Collaborative Overlay Power Management based on the Delay-Power

Tradeoffs" IEEE FGCT '2014, London (Best paper award)• Yan, H, Fontaine, F, Bouchet, O, Vuichard, J, Javaudin, Lebouc, M, Hamon, M, Cousin, B, and Gueguen, C. "HOPE: HOme

Power Efficiency System for a Green Network" IEEE INFOCOM'2013 Demo/Poster Session, Turin, Italy

Book Chapter:• Yan, H, Vuichard, J, Cousin, B, Gueguen, C, and Mardon, G "Green Home Network based on an Overlay Energy Control

Network" in the book "Green Networking and Communications" (Published by CRC Press, USA, oct 2013)

Poster:• Yan, H, "Smart devices collaboration for a greener home network" Presented at "Journée des doctorants" Orange Labs,

France , Sep 2014

French patents:• Yan, H, Mardon, G, and Gueguen, C (Dec 2012). "Economiser l’énergie du réseau domestique tout en maintenant la QoE de

l’utilisateur" Patent 1261565 • Yan, H, Fontaine, F, and Vuichard, J (Avr 2013) "Procédé de contrôle de la consommation énergétique d’équipements d’un

réseau de communication local" Patent 1352881• Yan, H, Fontaine, F, (Oct 2013) "Gestion améliorée des connexions réseau", Patent 1359446 • Fontaine, F, Yan, H, (Dec 2013) "Technique de communication dans un réseau local", Patent 1362832 • Fontaine, F, Yan, H, (Feb 2014) "Mécanisme de lissage de la consommation électrique des équipements UPnP", Patent

1452655• Yan, H, Fontaine, F, (Feb 2014) "An adaptive proxy for compliance of the equipments to IEEE 1905" Patent 1454940 • Yan, H, Fontaine, F, (Sep 2014) "A repetition proxy for long distance communication between Bluetooth devices" Patent

1459280 • Fontaine, F, Yan, H,(Sep 2014) "Detection mechanism of Bluetooth Low Energy devices (BLE) on the IP network using UPnP"

Patent 1459283