green computing

PowerPoint Presentation

Towards energy efficient Internet Service Providers ECOnet Perspective

Constantinos Vassilakis

[email protected]

Greek Research and Technology Network

Utrecht, Netherlands,

5-6 March 2012

GN3 Green Networking: Advances in Environmental Policy and Practice

low Energy COnsumption NETworks

1

Outline

The ECONET Project

Energy consumption and energy efficiency demand

Decomposing the Energy Consumption in the Wired Network

A Taxonomy of Undertaken Approaches

ECONET approach

Potential Impact on the Wired Network


5-6 March 2012


2


Increasing the energy efficiency and the sustainable growth of our world is a global process where Telecommunications technologies (and the ICTs in general) play a key role.

But to obtain optimum results the process should involve the two faces of the same coin:

Green ICT reducing the carbon footprint of ICT

ICT for Green using ICT for reducing third party-wastes.

ECONET is dealing with the first aspect

Focused on short and medium time exploitation

The ECONET project


5-6 March 2012

3

Participant organisation nameShort nameCountryConsorzio Nazionale Interuniversitario per le Telecomunicazioni UdR at DIST University of Genoa (Coordinator) CNITItalyMellanox TechnologiesMLXIsraelAlcatel LucentALUItalyLantiqLQDEGermanyEricsson Telecomunicazioni S.p.A.TEIItalyTelecom ItaliaTELITItalyGreek Research & Technology NetworkGRNETGreeceResearch and Academic Computer NetworkNASKPolandDublin City UniversityDCUIrelandVTT Technical Research CentreVTTFinlandWarsaw University of TechnologyWUTPolandNetVisorNVRHungaryEthernityETYIsraelLightCommLGTItalyInfoComINFOItaly
Manufacturers

Operators

Academic /research centers

Small/Medium Enterprises (SMEs)

The Consortium

The ECONET project


5-6 March 2012

4


Goals: re-thinking and re-designing network equipment towards more energy-sustainable and eco-friendly technologies and perspectives.

The overall idea is to introduce novel green network-specific paradigms and concepts enabling the reduction of energy requirements of wired network equipment by 50% in the short/mid-term (and by 80% in the long run) with respect to the business-as-usual scenario.

To this end, the main challenge is to design, develop and test novel technologies, integrated control criteria and mechanisms for network equipment allowing energy saving by dynamically adapting the device capacities and consumptions to current traffic loads and user requirements.

The ECONET project


5-6 March 2012

5



There are two main motivations that drive the quest for green ICT:

the environmental one, which is related to the reduction of wastes, in order to impact on CO2 emission;

the economical one, which stems from the reduction of operating costs (OPEX) of ICT services.

Gartner Group, Inc. (2007)

The global information and communications technology (ICT) industry accounts for approximately 2% of global carbon dioxide (CO2) emissions, a figure equivalent to aviation.

Note that the ICT sector raises much faster than aviation

How much is 2% of CO2?


5-6 March 2012

6



The figures refer to the whole corporate consumption. As such, they account for numerous sources, other than the operational absorption of the networking equipment (e.g., offices heating and lights). Notwithstanding, they give an idea of the general trend.


5-6 March 2012

7



Electrical energy consumption evolution and future trends for TELITs fixed network. Source: Telecom Italia

Source: C. Bianco, F. Cucchietti, G. Griffa, Energy consumption trends in the Next Generation

Access Network - a Telco perspective, IEEE INTELEC 2007.


5-6 March 2012

8


Decomposing the Energy ConsumptionThe Wired Network

Typical access, metro and core device density and energy requirements in todays typical networks deployed by telcos, and ensuing overall energy requirements of access and metro/core networks.

Source: R. Bolla, R. Bruschi, F. Davoli, F. Cucchietti, Energy Efficiency in the Future Internet: A Survey of Existing Approaches and Trends in Energy-Aware Fixed Network Infrastructures, IEEE Communications Surveys & Tutorials, vol. 13, no. 2, pp. 223-244, 2nd Qr. 2011.


5-6 March 2012

9


Decomposing the Energy Consumption High-end Routers

Estimate of power consumption sources in a generic platform of high-end IP router.

Source: R. Tucker, Will optical replace electronic packet switching?, SPIE Newsroom, 2007.


5-6 March 2012

10


Decomposing the Energy Consumption Is the energy consumption currently load-dependent?

Network engineers only speak about the capacity of a device or of a link interface

as a matter of fact, device and link are specifically designed to work at the maximum speed

Source: The ECONET Consortium, End-user requirements, technology specifications and benchmarking methodologies, Deliverable 2.1.

Power consumption in GRNET core routers (24-hour period)

Daily traffic profile of core GRNET network router (peering with GEANT)


5-6 March 2012

11


Decomposing the Energy Consumption Is the energy consumption currently load-dependent?

Power Consumption of Cisco Catalyst 2970 Switch

Source: K. Christensen, P. Reviriego, B. Nordman, M. Bennett, M. Mostowfi, J.A. Maestro, "IEEE 802.3az: the road to energy efficient ethernet," IEEE Communications Magazine, vol.48, no.11, pp.50-56, November 2010.

There is no significant difference in power consumption whether a port is running at 10 Mbps or 100 Mbps.

The switch power consumption is increased by connecting a new link, even if there is no data being transmitted on this link.

The difference in power consumption is quite low when a 1 Gbps link is fully utilized compared to when it is zero utilized.


5-6 March 2012

12


Decomposing the Energy Consumption Day & Night Traffic Profiles

Percentage w.r.t. peak level. The profiles exhibit regular, daily cyclical traffic patterns with Internet traffic dropping at night and growing during the day.

Traffic load fluctuation at peering links for about 40 ISPs from USA and Europe

Source: http://asert.arbornetworks.com/2009/08/what-europeans-do-at-night/


5-6 March 2012

13


Decomposing the Energy Consumption Energy wastes

Networks and devices are lightly utilized.

Often peak loads during rush hours are generally much lower than capacities of links and devices.

It is well known that the overdimensioning is the best design strategy for assuring QoS levels

Moreover, traffic loads follow well-known day & night fluctuations.

On the other hand, the energy requirements of network devices remain substantially flat according to their workload.

Furthermore, networks are highly overprovisioned /redundant to assure service availability.


5-6 March 2012

14



Source: R. Bolla, R. Bruschi, F. Davoli, F. Cucchietti, Energy Efficiency in the Future Internet: A Survey of Existing Approaches and Trends in Energy-Aware Fixed Network Infrastructures, IEEE Communications Surveys & Tutorials, vol. 13, no. 2, pp. 223-244, 2nd Qr. 2011.

The largest part of undertaken approaches regarding engineered improvements is funded on few base concepts, which have been generally inspired by energy-saving mechanisms and power management criteria that are already partially available in computing systems.


5-6 March 2012

15


Re-engineering approaches aim at:

introducing and designing more energy-efficient elements for network device architectures

suitably dimensioning and optimizing the internal organization of devices

reducing their intrinsic complexity levels.


Re-engineering


5-6 March 2012

16


The dynamic adaptation of network/device resources is designed to modulate capacities of packet processing engines and of network interfaces, to meet actual traffic loads and requirements.

This can be performed by using two power-aware capabilities, namely, dynamic voltage scaling and idle logic, which both allow the dynamic trade-off between packet service performance and power consumption.


Dynamic Adaptation


5-6 March 2012

17


Standard operations

Idle logic

Power scaling

Idle + power scaling

Wakeup and sleeping times

Increased service times

Wakeup and sleeping + increased service times


Dynamic Adaptation


5-6 March 2012

First version: Adaptive Link Rate proposed by Christensen and Nordman

Final Version: based on the low power idle concept, proposed by Intel.

Idea: transmit data at the maximum speed, and put the link to sleep when it is idle.


19

0

5

10

15

Link speed (Mb/sec)

Power (W)

10

100

1000

10000


Dynamic Adaptation: Green Ethernet (IEEE 802.3 az)

Tw and Ts for 10 Gb/s in IEEE Std 802.3az-2010 are 4.48 s and 2.88 s, respectively

LPI can possibly be

asynchronous


5-6 March 2012

19

In PC-based devices, the Advanced Configuration and Power Interface (ACPI) provides a standardized interface between the hardware and the software layers.

ACPI introduces two power saving mechanisms, which can be individually employed and tuned for each core:

Power States (C-states)

C0 is the active power state

C1 through Cn are processor sleeping or idle states (where the processor consumes less power and dissipates less heat).

Performance States (P-states)

while in the C0 state, ACPI allows the performance of the core to be tuned through P-state transitions.P-states allow to modify the operating energy point of a processor/core by altering the working frequency and/or voltage, or throttling the clock.


Dynamic Adaptation: SW routers & ACPI



5-6 March 2012

20

21

Source: R. Bolla, R. Bruschi, A. Ranieri, Green Support for PC-based Software Router: Performance Evaluation and Modeling, Proc. IEEE ICC 2009, Dresden, Germany, June 2009. Best Paper Award.

[MHz]


Dynamic Adaptation: SW routers & ACPI



5-6 March 2012

21

Sleeping/standby approaches are used to smartly and selectively drive unused network/device portions to low standby modes, and to wake them up only if necessary.

However,

since todays networks and related services and applications are designed to be continuously and always available,

standby modes have to be explicitly supported with special techniques able to maintain the network presence of sleeping nodes/components.


Sleeping/Standby



5-6 March 2012

22

Scenario: networked hosts (PCs, consumer electronics, etc.);

Problem: when an end-host enters standby mode, it freezes all network services, and it is not able to maintain its network presence;

Idea: introduce a Network Connection Proxy (NCP), which is devoted to maintain the network presence of sleeping hosts.

Sleeping host

NCP

Internet

Continuous and full connectivity

Wakeup/sleep

messages

Application-specific messages

Zzzzz

I want to sleep

Source: M. Allman, K. Christensen, B. Nordman, V. Paxson, Enabling an Energy-Efficient Future Internet Through Selectively Connected End Systems, Proc. ACM SIGCOMM HotNets, Atlanta, GA, Nov. 2007.


Sleeping/Standby: Proxying the Network Presence



5-6 March 2012

23

Scenario: Core Networks

Idea: put links, interfaces and part of nodes (e.g., line-cards) to sleep

Problem: Network stability, convergence times at multiple levels (e.g., MPLS traffic engineering + IP routing)

Source: R. Bolla, R. Bruschi, A. Cianfrani, M. Listanti, Putting Backbone Networks to Sleep, IEEE Network Magazine, Special Issue on Green Networking, vol. 25, no. 2, pp. 26-31, March/April 2011.





5-6 March 2012

24

Solution: they exploited two features already present in todays networks and devices:

network resource virtualization

modular architecture of network nodes.

This approach allows to:

Put physical resources to sleep (e.g., links, linecards, etc.);

Move the logical entities working on physical elements going to sleep, to other physical elements on the device.

If suitable L2 protocols are used, the complexity of standby management can be hidden from the IP layer, and totally managed inside traffic engineering procedures.





5-6 March 2012

25

Standby states have usually much lower energy requirements than active states.

Network-wide control strategies (i.e., routing and traffic engineering) give the possibility of moving traffic load among network nodes.

When a network is under-utilized, we can move network load on few active nodes, and put all the other ones in standby.

Different network nodes can have heterogeneous energy capabilities and profiles.

Recent studies, obtained with real data from Telcos (topologies and traffic volumes) suggested that network-wide control strategies could cut the overall energy consumption by more than 23%.

Standby state

Performance scaling

Power Consumption

Energy-aware state



Green network-wide control: Traffic engineering & routing


5-6 March 2012

26

Only local control policies

Local + network-wide control policies

Once network devices will include energy management primitives, further energy reduction will be possible by moving traffic flows among the network nodes, in order to minimize the energy consumption of the entire infrastructure.



Green network-wide control: Traffic engineering & routing


5-6 March 2012

27


The ECONET approach


5-6 March 2012

28


The ECONET approach


5-6 March 2012

29


The ECONET approach


5-6 March 2012

30


Green Abstraction Layer

The ECONET approach


5-6 March 2012

31


The ECONET approach

ECONET Test Bench

@ TELIT Test Plant


5-6 March 2012

32


Potential Impact on the Wired Network

The previously mentioned green technologies allow designing new-generation network devices characterized by energy profiles

Reference: R. Bolla, R. Bruschi, A. Carrega, F. Davoli, D. Suino, C. Vassilakis, A. Zafeiropoulos, Cutting the Energy Bills of Internet Service Providers and Telecoms through Power Management: an Impact Analysis, Elsevier Computer Networks, Special Issue on Green Communication Networks, to appear


5-6 March 2012

33


Potential Impact on the Wired NetworkTELIT reference scenario

Network load statistics and topology data

2015-2020 network forecast: device density and energy requirements

(example based on Italian network)
customers per DSLAM640average usage of a network access30%average traffic when a user is connected10%redundancy degree for metro/transport devices13%redundancy degree for core devices100%redundancy degree of metro/transport device links100%redundancy degree of core device links50%average traffic load in metro networks40%average traffic load in core networks40%standby efficiency85% performance scaling efficiency50% network-wide control efficiency20% air cooling/power supply efficiency15%
Home/Access

Metro/Transport/Core

target

Source: forecast based on: carrier grade topologies; traffic analysis and indicators (ETSI TR 102530, ODYSSEE) and projected traffic load.
power consumption (Wh)number of devicesoverall consumption (GWh/year)Home1017,500,0001,533Access1,28027,344307Metro/transport6,0001,75092Core10,00017515
Sources: 1) BroadBand Code of Conduct V.3 (EC-JRC) and inertial technology improvements to 2015-2020 (home and access cons.)

2) Telecom Italia measurements and evaluations (power consumption of metro/core network and number of devices)
Data PlaneControl PlaneCooling/Power SupplyHome79%3%18%Access84%3%13%Metro/transport73%13%14%Core54%11%35%
Device internal sources of energy consumption

Sources: Information from vendors.

Sources: BroadBand Code of Conduct V.3 (EC-JRC) and technology improvements to 2015-2020.


5-6 March 2012

34


Potential Impact on the Wired NetworkTELIT network topology and traffic profiles


5-6 March 2012

35


Yearly Energy consumption estimation for TELIT

Potential Impact on the Wired NetworkIs There Room for Energy Saving Optimization?

Room for Energy Saving Optimization

Home/access

Metro/Transport

Core


5-6 March 2012

36


Potential Impact on the Wired NetworkEnergy consumption model outline

Source: R. Bolla, R. Bruschi, A. Carrega, F. Davoli, D. Suino, C. Vassilakis, A. Zafeiropoulos, Cutting the Energy Bills of Internet Service Providers and Telecoms through Power Management: an Impact Analysis, Elsevier Computer Networks, Special Issue on Green Communication Networks,

to appear


5-6 March 2012

37


DPS primitives only

Standby primitives only

DPS & Standby primitives

Potential Impact on the Wired NetworkEstimated energy saving for the TELIT network

We suppose standby capabilities to be applied only where alternative paths are present.


5-6 March 2012

38


Potential Impact on the Wired NetworkThe GRNET network case

Yearly Energy consumption estimation for GRNET

DPS & Standby primitives

GRNET network does not have Access/Home parts


5-6 March 2012

39


5-6 March 2012


Thank you for your attention!

Questions?

[email protected]

http://econet-project.eu

http://green.grnet.gr

low Energy COnsumption NETworks

40


5-6 March 2012


Backup slides


5-6 March 2012

41


Decomposing the Energy Consumption Access Technologies

Power consumption of DSL, HFC, PON, FTTN, PtP, WiMAX, and UMTS as a function of access rate with an oversubscription rate of 20. The technology used is fixed at 2010 vintage for all access rates.

Source: Baliga, J.; Ayre, R.; Hinton, K.; Tucker, R.S.; , "Energy consumption in wired and wireless access networks," IEEE Communications Magazine, vol. 49, no. 6, pp. 70-77, June 2011.


5-6 March 2012

42


Adoption of pure optical switching architectures:

They can potentially provide terabits of bandwidth at much lower power dissipation than current network devices.

But their widespread adoption is still hindered by technological challenges: problems mainly regard the limited number of ports and the feasibility of suitable buffering schemes.

Decreasing feature sizes in semiconductor technology have contributed to performance gains:

allowing higher clock frequencies

designing improvements such as increased parallelism.

the same technology trends have also allowed for a decrease in voltage that has reduced the power per byte transmitted by half every two years, as suggested by Dennards scaling law.


Re-engineering


5-6 March 2012

43

Source: R. Bolla, R. Bruschi, A. Carrega, F. Davoli, Green Network Technologies and the Art of Trading-off, Proc. IEEE INFOCOM 2011 Workshop on Green Communications and Networking, Shanghai, China, April 2001, pp. 301-306.

t

(t)

a(Py)

idle(Cx)

t(Cx)

on

off

conf

TR

TI

TB


Dynamic Adaptation: Understanding the Power-Performance Tradeoff

Modeling and control

Recently a simple model has been proposed by Bolla et al, which is based on classical queueing theory and allows representing the trade-off between energy and network performance in the presence of both AR and LPI capabilities.

The model is aimed at describing the behaviour of packet processing engines.

It is based on a Mx/D/1/SET queueing system.



5-6 March 2012

44


Dynamic Adaptation: Understanding the Power-Performance Tradeoff

Modeling and control



5-6 March 2012

45



Re-engineering: Optical Backbone Networks

The creation of optical paths (via DWDM) within optical backbone networks has been utilized for the dynamic establishment of high capacity circuits with reduced energy demands


5-6 March 2012

46


Standardization efforts

The European Union already published a number of Codes of Conduct

covering different categories of equipment, including broadband equipment, data centres, power supplies, UPS. The Code of Conduct on Energy Consumption of Broadband Equipment has been defined by the EU, which sets targets in reducing energy consumption in the access network

IEEE has also ratified the Energy Efficient Ethernet (EEE) standard in October 2010, also known as IEEE 802.3az,

which is a set of enhancements to the twisted-pair and backplane Ethernet networking standards that will allow for more than 50% less power consumption during periods of low data activity, while retaining full compatibility with existing equipment.

ENERGY STAR is a joint program of the U.S. Environmental Protection Agency and the U.S. Department of Energy that has defined the ENERGY STAR Product Specifications.

IETF has recently established the Energy Management (EMAN) Working Group.

Different interesting issues are under consideration by the Environmental Engineering Technical Body in ETSI

The Home Gateway Initiative (HGI) launched an internal task force called Energy Saving with the objective of setting up requirements and specifications for energy efficiency in the home gateways

ITU-T Study Group 15 (Optical transport networks and access network infrastructures)

ITU-T created in September 2008 a new Focus Group, namely, FG ICT & Climate Change


5-6 March 2012

47

ECONET approach



5-6 March 2012

84889296GWhYears981008

Start of network digitalizationEnd of network digitalization

E TOTE TLC

Fixed network domain

E TOT: total energy consumption from mains (TLC equipment, cooling, ausiliary systems)

E TLC: energy consumption of TLC equipment

End user appliancesPower Consumption

New challenge on energy saving

Need of further actions

on TLC equipments

Energy consumption became a Key Issue

Start ADSL deployment

Page 43 of 76

FP7-ICT-258454 D2.2

5.2.1.1 TELIT Testbed The chain that will be created in the TELIT test plant will reproduce a complete

telecommunication network from the backbone transport to the final customer devices.

In order to guarantee the correct devices configuration and the appropriate interconnection among them, the chain will be realized with the presence and the contribution of all the partners who have provided the nodes of the network.

The scheme of the test network is shown in Figure 32.

Figure 32: Scheme of the chain that will be used in the testbed.

The test will be performed by measuring the consumption of the entire network while the appropriate traffic (as defined in section 4.2) is generated and sent into the network. All the nodes will be fed by the same electrical source in order to collect all the absorbed power. At the same time the power consumption of single devices will be collected from the on-board monitoring system of each network component thus is will be feasible to make a fine analysis of the power absorbed by each part of the chain.

All consumption data will be collected and synchronized with the traffic analysis performed by the traffic generator and analyser.

The tests will be conducted in different ways in order to collect the consumption characteristics in several conditions. In particular, an initial measurement will be performed with all energy saving functionality (where possible) disabled in order to obtain a base comparison situation.

Then, the measurement will be repeated in different device configurations and by applying the different traffic engineering policies that will be developed in the project.

LQDE

VDSL + vectoring

LQDE + INFO

GbE

INFO

ETYETY

GbE

ADSL

ETY

24xGbE 1x10GbE?

ETYLQDE

GbE

VDSL

1x1GbE

ALU*

ALU*

CNIT/MLX/DCU

CNIT/MLX/DCU

MLX

MLX

ALU*

home access metro Core/transport

TEI

Optional Data-center emulation

1x10GbE

1x10GbE

1x10GbE

1x10GbE

1x10

GbE

1x10

GbE

1x1GbE

1x10GbE

1x1GbE

1x1GbE

1x1GbE

1x1GbE

Multiple nodes will be realized by virtualizing the

data-plane of two physical nodes

Page 43 of 76

FP7-ICT-258454 D2.2

5.2.1.1 TELIT Testbed

The chain that will be created in the TELIT test plant will reproduce a complete

telecommunication network from the backbone transport to the final customer devices.

In order to guarantee the correct devices configuration and the appropriate interconnection

among them, the chain will be realized with the presence and the contribution of all the partners

who have provided the nodes of the network.

The scheme of the test network is shown in Figure 32.

Figure 32: Scheme of the chain that will be used in the testbed.

The test will be performed by measuring the consumption of the entire network while the

appropriate traffic (as defined in section 4.2) is generated and sent into the network. All the nodes

will be fed by the same electrical source in order to collect all the absorbed power. At the same time

the power consumption of single devices will be collected from the on-board monitoring system of

each network component thus is will be feasible to make a fine analysis of the power absorbed by

each part of the chain.

All consumption data will be collected and synchronized with the traffic analysis performed by

the traffic generator and analyser.

The tests will be conducted in different ways in order to collect the consumption characteristics in

several conditions. In particular, an initial measurement will be performed with all energy saving

functionality (where possible) disabled in order to obtain a base comparison situation.

Then, the measurement will be repeated in different device configurations and by applying the

different traffic engineering policies that will be developed in the project.

LQDE

VDSL +

vectoring

LQDE + INFO

GbE

INFO

ETY

ETY

GbE

ADSL

ETY

24xGbE1x10GbE?

ETY

LQDE

GbE

VDSL

1x1GbE

ALU*

ALU*

CNIT/MLX/DCU

CNIT/MLX/DCU

MLX

MLX

ALU*

homeaccess

metro Core/transport

TEI

Optional Data-center emulation

1x10GbE

1x10GbE

1x10GbE

1x10GbE

1

x

1

0

G

b

E

1

x

1

0

G

b

E

1x1GbE

1x10GbE

1x1GbE

1x1GbE

1x1GbE

1x1GbE

Multiple nodes will

be realized by

virtualizing the

data-plane of two

physical nodes

0%

20%

40%

60%

80%

100%

P

e

r

c

e

n

t

a

g

e

o

f

l

i

n

k

o

c

c

u

p

a

t

i

o

n

i

n

r

e

s

p

e

c

t

t

o

t

h

e

m

a

x

i

m

u

m

[

%

]

Time [h]

Typical traffic Profile for a business link

Working day

Holiday

0%

20%

40%

60%

80%

100%

P

e

r

c

e

n

t

a

g

e

o

f

l

i

n

k

o

c

c

u

p

a

t

i

o

n

i

n

r

e

s

p

e

c

t

t

o

t

h

e

m

a

x

i

m

u

m

[

%

]

Time [h]

Typical traffic Profile for a residential link

Working day

Holyday

M I

PD

TS

BS

BO

TO

GE

FI

PA

RM

NA

BA

SV

AL

BG

CO

VR

VE

BZ

MO

RI

PI

AN

PG

PE

CA

TA

CZ

CT

NL

0%1%2%3%4%5%6%7%8%9%10%0,E+002,E-064,E-066,E-068,E-061,E-051,E-050.005.0010.0015.0020.001.006.0011.0016.0021.002.007.0012.0017.0022.003.008.0013.0018.0023.00Error (%)Average Latency Time [s]Time [hh:mm]Max Error (%)AM{P0, C1}AM{P0, C2}AM{P1, C1}AM{P1, C2}AM{P2, C1}AM{P2, C2}AM{P3, C1}AM{P3, C2}SR{P0, C1}SR{P0, C2}SR{P1, C1}SR{P1, C2}SR{P2, C1}SR{P2, C2}SR{P3, C1}SR{P3, C2}

0,0%0,2%0,4%0,6%0,8%1,0%-5,0E-06-1,0E-205,0E-061,0E-051,5E-052,0E-052,5E-053,0E-050.005.0010.0015.0020.001.006.0011.0016.0021.002.007.0012.0017.0022.003.008.0013.0018.0023.00Error (%)Loss ProbabilityTime [HH:mm]Max Error (%)AM{P0, C1}AM{P0, C2}AM{P1, C1}AM{P1, C2}AM{P2, C1}AM{P2, C2}AM{P3, C1}AM{P3, C2}SR{P0, C1}SR{P0, C2}SR{P1, C1}SR{P1, C2}SR{P2, C1}SR{P2, C2}SR{P3, C1}SR{P3, C2}

0,0%0,5%1,0%1,5%2,0%2,5%3,0%3,5%4,0%4,5%5,0%7891011121300.0005.0010.0015.0020.0001.0006.0011.0016.0021.0002.0007.0012.0017.0022.0003.0008.0013.0018.0023.00Maximum Error (%)Power Consumption (W)Time [hh:mm]Error (%)AM{P0, C1}AM{P0, C2}AM{P1, C1}AM{P1, C2}AM{P2, C1}AM{P2, C2}AM{P3, C1}AM{P3, C2}SR{P0, C1}SR{P0, C2}SR{P1, C1}SR{P1, C2}SR{P2, C1}SR{P2, C2}SR{P3, C1}SR{P3, C2}

green computing

Documents

green ict

environmental policy

2012gn3 green networking

energy efficiency demandthere

practiceenergy consumption

redesigning network

wired networkutrecht

reduction of wastes