Oliver Holland - IEEE VTS UKRI - Energy efficiency challenges of data volume increases and the use of sleep modes facilitated by opportunistic cognitive radio networking as a solution

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Energy Efficiency Challenges of Data Volume Increases, and the use of Sleep Modes facilitated by Opportunistic Cognitive Radio Networking as a Solution

Oliver Holland

King’s College London, UK

2

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Overview • Energy consumption Implications of data volume

increases

• Opportunistic networking using cognitive radio to

facilitate sleep modes for radio network equipment

– Scenarios

– Example mechanism facilitating awareness

– Some example results

• Conclusion and future considerations

3

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Implications for energy consumption

• How do we maintain this same expectation?

illustration courtesy

of IEEE Spectrum

4

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Three ways to increase capacity (with fixed spectrum) – Achieve better link performance (closer to Shannon limit)

– Increase Tx power

– Increase density of frequency reuse

Capacity

SINR

Implications for energy consumption

5

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Implications for energy consumption

• Increase density of

frequency reuse

– Far smaller cells

– Lower power per cell

consumption and better

able to take advantage

of environment (e.g.,

propagation), BUT

– Latent energy

consumption an issue;

still very low Tx-to-input

power efficiency

ICT-EARTH D2.3

6

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Implications for energy consumption

• Increase density of frequency reuse

– Far smaller cells—embodied energy

smaller

cells

7

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Implications for energy consumption

• Embodied energy

8

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Opportunistic Networking Using Cognitive Radio to Save Energy

• So what can we do?

• Opportunistic cognitive radio connectivity/networking

– To minimise number of network elements that are active at any one

point in time through facilitating sleep modes

– To minimise the number that are deployed in first place

– Achieved by awareness through cognitive radio of what is deployed

and available (connectivity options)

– Awareness/prediction through cognitive radio of what has happened

and will happen in the future (user mobility affecting availability of

connectivity options, traffic variations, traffic requirements, etc.)

– Planning for connectivity options based on all this awareness

9

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Opportunistic peer-to-peer to reduce necessary transmission power and number of transmissions, given awareness of the end-node being in the vicinity and with a good channel

Opportunistic Networking Using Cognitive Radio to Save Energy

?

10

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Opportunistic usage of a more power efficient or better channel connectivity means given awareness of the connectivity means existing

Opportunistic Networking Using Cognitive Radio to Save Energy

?

11

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Transmission of delay-tolerant traffic at a more appropriate time based on mobility

Opportunistic Networking Using Cognitive Radio to Save Energy

?

12

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• “Store-carry-forward” for delay-tolerant traffic; facilitating the powering down of network elements (e.g., reducing necessary cell density) by transmitting at a more appropriate time.

Opportunistic Networking Using Cognitive Radio to Save Energy

13

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Network elements shutdown when p2p connectivity is sufficient

Opportunistic Networking Using Cognitive Radio to Save Energy

14

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Awareness of Opportunistic Networking Using IEEE 1900.6

I am ‘A’ type of sensor with ‘B’ serial number

My location is ‘C’

I have detected RATs ‘D’, ‘E’ and ‘F’ at ‘G’, ‘H’, and ‘I’ frequency

I have found ‘J’ signal autocorrelation function at ‘K’ frequency

(Perhaps future addition) I have ‘L’, ‘M’, ‘N’ radio configuration capability

I wonder which devices

are in the area that I might

be able to communicate

with through the

opportunistic formation of

“cognitive radio” links?

Let’s check with IEEE

1900.6, the

communication

subsystem of which I am

connected to…

CE/DA

Device 1

(S and CE embedded)

Device 2

(S embedded)

Great! If his serial is ‘B’

then he is hosted by ‘O’

type of device, which I can

connect to! This is one

connection option at

location ‘C’

S = Sensor

CE = Cognitive Engine

DA = Data Archive Also, I now know that there are devices

transmitting RATs ‘E’ and ‘F’

somewhere near location ‘C’, and I am

able to communicate with those

devices or networks as I am capable of

RATs ‘E’ and ‘F’

But there’s more! That

autocorrelation function ‘J’ found at

location ‘C’ looks like RAT ‘P’, e.g.,

due to the time duration between its

peaks. I could also connect with that

But wait! There is also a

DA in this 1900.6

system. Bet there is a lot

of information there!

Let’s find out

Now I know lots of things! I can connect with ‘Q’

network at location ‘R’, ‘S’ network at location ‘T’,

‘U’ device at location ‘V’. I know all the RATs and

link capabilities which I can associate with at given

locations, and can match that to my expected future

traffic capabilities and mobility, etc

Over S-S Interface (e.g., collaborative sensing scenario)

Request

I can even have a fair idea of cognitive radio

ad-hoc networking possibilities (e.g., routes

and prospective link capabilities over

multiple hops) and use this knowledge in

collaboration with other devices to

autonomously form such networks

15

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Opportunistic usage of Wi-Fi access points (including in TV white space!) to

enable power saving modes for cellular network equipment (powering down cells

where possible and sectorization switching—20% Wi-Fi access point deployment)

Example: Offload to Wi-Fi enabling Cellular Power Saving Modes

16

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Opportunistic usage of Wi-Fi access points (including in TV white space!) to

enable power saving modes for cellular network equipment (powering down cells

where possible and sectorization switching—5% Wi-Fi access point deployment)

Example: Offload to Wi-Fi enabling Cellular Power Saving Modes

17

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Results on previous slides obtained through simulations using following coverage

analyses as basis: S. Kawade and M. Nekovee, “Broadband wireless delivery using

an inside-out TV white space network architecture,” IEEE Globecom 2011

• Further detail can be obtained in A. Aijaz, O. Holland, P. Pangalos, H. Aghvami, H.

Bogucka, “Energy Savings for Mobile Communication Networks through Dynamic

Spectrum and Traffic Load Management,” to appear in Green Communications:

Theoretical Fundamentals, Algorithms and Applications, CRC Press, 2012

• Further related work has been presented in ICC 2012: A. Aijaz, O. Holland, P.

Pangalos, and H. Aghvami, “Energy Savings for Cellular Access Network through

Wi-Fi Offloading”

Example: Offload to Wi-Fi enabling Cellular Power Saving Modes

18

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Mix of FTP, HTTP and video streaming traffic, 15%, 45% and 40% respectively

…

…

Example: Offload to Wi-Fi enabling Cellular Power Saving Modes

19

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Example: Offload to Wi-Fi enabling Cellular Power Saving Modes

• Opportunistic reallocation between frequency bands/networks to enable power

saving modes (base station powering down and sectorization switching)

• Can also extend to network-side reconfiguration decisions

(power consumption

model similar to macro

case on slide 5)

20

IEEE VTS-UKRI Dublin Meeting

26 July 2012

• Using cognition on the

network side (fuzzy cognitive

maps) to learn about traffic

variations on make decisions

on power saving modes

Example: Offload to Wi-Fi enabling Cellular Power Saving Modes

• Cumulative energy

consumption and blocking

rate

21

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Conclusion • Big energy consumption issues caused by data volume increases

– Capacity provision ultimately will require greater frequency reuse and smaller

cells (under assumption of the same spectrum)

– Presents energy issues, both operational and embodied

• Presented opportunistic cognitive radio networking as a means to

save energy by facilitating power saving modes

• Discussed various scenarios in which such solutions might apply

• Shown performance examples indicating very significant savings

• Future prospects

– “Green communications” research has to consider from-the-socket power

rather than just minimising transmission power (is beginning to happen to

some extent) as well as embodied energy (hardly considered thus far)

– Solution such as presented here help address/consider both such issues

22

IEEE VTS-UKRI Dublin Meeting

26 July 2012

References [1] O. Holland, T. Dodgson, A. H. Aghvami., and H. Bogucka, “Intra-Operator Dynamic Spectrum

Management for Energy Efficiency,” IEEE Communications Magazine, to appear

[2] O. Holland, O. Cabral, F. Velez, A. Aijaz, P. Pangalos and A. H. Aghvami, “Opportunistic Load and

Spectrum Management for Mobile Communications Energy Efficiency,” IEEE PIMRC 2011, Toronto,

Canada, Sept. 2011

[3] O. Holland, C. Facchini, A. H. Aghvami, O. Cabral, and F. Velez, “Opportunistic Spectrum and Load

Management for Green Radio,” chapter appearing in: E. Hossein, V. Bhargava, G. Fettweis, 2011,

Green Radio Communication Networks, Cambridge University Press, 2011

[4] O. Holland, Vasilis Friderikos, A. H. Aghvami, “Green Spectrum Management for Mobile Operators,”

IEEE Globecom, Miami, FL, USA, December 2010

[5] O. Holland et al., “Intra-Operator Spectrum Sharing Concepts for Energy Efficiency and Throughput

Enhancement,” CogART 2010, Rome, Italy, November 2010 (invited paper)

[6] A. Aijaz, O. Holland, P. Pangalos, A.H. Aghvami, “Energy Savings for Cellular Access Network

through Wi-Fi Offloading,” IEEE ICC 2012, Ottawa, ON, Canada, June 2012

[7] A. Aijaz, O. Holland, P. Pangalos, H. Aghvami, H. Bogucka, “Energy Savings for Mobile

Communication Networks through Dynamic Spectrum and Traffic Load Management,” appearing in

Green Communications: Theoretical Fundamentals, Algorithms, and Applications, Auerbach

Publications, CRC Press, Taylor & Francis Group

[8] C. Facchini, O. Holland, F. Granelli, N. Fonseca, A. H. Aghvami, “Dynamic Green Self-Configuration

of 3G Base Stations using Fuzzy Cognitive Maps,” submitted to Elsevier Computer Networks

23

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Acknowledgement • This work has been supported by the ICT-

ACROPOLIS Network of Excellence, www.ict-

acropolis.eu, FP7 project number 257626

24

IEEE VTS-UKRI Dublin Meeting

26 July 2012

Thank you!

[email protected]