COMMUNICATE GREEN

Energy Efﬁcient Mobile Communication

Abdulbaki Uzun, Axel K

upper

Service-centric Networking, Deutsche Telekom Laboratories, TU Berlin, Berlin, Germany

Hans J. Einsiedler

Innovation Development, Deutsche Telekom Laboratories, Berlin, Germany

Keywords:

Mobile computing, Green ICT, Energy efﬁciency, Context-aware computing.

Abstract:

In order to fulﬁll today’s high demands on mobile network usage, mobile network providers in Germany have

around 100.000 base stations working 24/7. The permanent availability of those network components causes

a signiﬁcant energy consumption. Through an adaptive and context-aware power management in mobile

networks, a considerable amount of energy can be saved by maintaining the high quality of experience at

the same time. There are a lot of contextual information present in mobile network components and end

devices, which can help to calculate decisions for a dynamic de- and reactivation of network components. This

position paper discusses the idea of a context entity that aggregates and processes various types of contextual

information, and the integration of it into the mobile network architecture.

1 INTRODUCTION

Modern and future radio networks, such as WiFi,

UMTS or LTE, provide good connectivity and high

data transfer rates to the mobile user. In order to ful-

ﬁll these requirements, mobile network providers in

Germany have around 100.000 base stations working

24/7 with a total power consumption of 760 GWh to

3040 GWh per year. However, these capacities are not

always fully utilized, e.g., in rural areas or at night.

Furthermore, mobile services that are mostly in use

today (e.g., telephony, SMS, etc.), do not have high

data rate requirements. Even though the demand on

bandwidth for mobile Internet increases continuously,

a lot of the services used today can be implemented

with older networks (e.g. GSM or EDGE) without

any noticeable loss in quality. These examples show

that there are massive potentials to save energy in ra-

dio access and core networks. This can be done by dy-

namically de- and reactivating network components

(e.g., base stations) or by adaptively reconﬁguring the

network to the user’s needs based on a certain infor-

mation basis gained from the network.

The main objective of the project Communicate

Green is the development of an adaptive, context-

aware and technology-comprehensive power manage-

ment for modern radio networks, with which a consid-

erable amount of energy can be saved by maintaining

the high quality of experience at the same time. There

are a lot of contextual information present in mobile

network components and end devices, which can be

useful in order to calculate decisions for a dynamic

de- and reactivation of network components, such as

the current load of a cell, the average daily load of

a cell, the number of users in a cell, service usage

proﬁles or QoS requirements. This paper mainly dis-

cusses the idea of developing a context entity that ag-

gregates and processes various types of contextual in-

formation coming from mobile network components

as well as end devices, and the integration of this en-

tity into the mobile network architecture.

2 OBJECTIVES

The project Communicate Green is about implement-

ing an adaptive and context-aware power manage-

ment in radio networks having a global view on dif-

ferent overlay networks. To achieve a high power efﬁ-

ciency throughout different mobile network technolo-

gies, all parts of the system are going to be optimized.

Radio network infrastructures are most of the time

fully active even when they are not fully utilized (see

302

Uzun A., Küpper A. and J. Einsiedler H..

COMMUNICATE GREEN - Energy Efﬁcient Mobile Communication.

DOI: 10.5220/0003399403020305

In Proceedings of the 1st International Conference on Pervasive and Embedded Computing and Communication Systems (PECCS-2011), pages

302-305

ISBN: 978-989-8425-48-5

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 1). Only in some regions, where the utilization

is foreseen, base stations are temporarily switched

off. The parameters are not very sophisticated and

dynamic, such as the time of the day or special knowl-

edge about certain activities in the region. A simple

optimization can be achieved by adjusting the param-

eters of a mobile network to the actual operating state

and by providing only the needed capacity. Possible

approaches are the dynamic de- and reactivation of

base stations and the transfer of coverage to neigh-

boring base stations.

Figure 1: Daily load proﬁle of a telecommunication net-

work in January 2009, Ericsson GmbH.

The potential of optimizing mobile networks in

terms of energy efﬁciency can be increased by cre-

ating integrated power management concepts for het-

erogeneous radio networks. One possible approach is

the adaptive conﬁguration and selection of radio net-

works based on decreasing the power consumption by

maintaining a high quality of experience for the user.

Figure 2 shows an exemplary night mode, in which

3G/4G cells and WiFi HotSpots are conﬁgured in a

way that only one base station is active in order to

guarantee basic services and a few WiFi HotSpots that

deliver broadband data rates.

Figure 2: Exemplary deactivation of network nodes.

Changes in radio access networks as they are men-

tioned above directly affect the core network. In order

to assure that all functions in the core network can be

delivered without any loss in the quality of service,

the core network has to be optimized as well. Virtual-

ization of functions is the key approach here.

The basis for these network optimizations is built

by various types of information present in mobile net-

work components, end devices and other sources –

here named as context. A list of relevant contextual

information will be determined by examining differ-

ent mobile network context sources. Furthermore, a

service classiﬁcation scheme will be developed show-

ing what kind of service can be delivered by what

kind of radio technology. The main focus of our work

in the project is the development of a context entity

that aggregates and processes contextual information

in order to calculate decisions for network optimiza-

tions. By applying various context acquisition ap-

proaches, both the network as well as the user view

will be included into the power management process.

3 RELATED WORK

The problem of energy consumption is also recog-

nized by other project groups. The EARTH project

(EARTH, 2010), for example, funded by the EU, in-

vestigates the energy efﬁciency of mobile communi-

cation systems in order to develop a new generation

of energy efﬁcient equipment, components, deploy-

ment strategies and energy aware network manage-

ment solutions. It mainly focuses on LTE, its evolu-

tion LTE-A and beyond. Communicate Green, how-

ever, considers all available radio technologies (e.g.,

GSM, UMTS or WiFi) and applies energy efﬁcient

decision algorithms on available hardware and net-

works based on network, service and user context.

Therefore, a context entity containing context

quality, provisioning, modeling and reasoning mecha-

nisms plays a big role in this project. A lot of research

has been done in this ﬁeld so far. (Flor

een et al., 2005)

describe a context management framework developed

within the MobiLife project (MobiLife, 2006), which

handles context information, modeling and reason-

ing for mobile applications and services. For this

purpose, essential functionalities are split into differ-

ent components, which can be conﬁgured for differ-

ent tasks and reasoning methods. (Mannweiler et al.,

2009) uses the concept of a context management ar-

chitecture in order to apply context information for

intelligent radio network access (IRNA) purposes. As

presented by (Schneider et al., 2010), this architecture

is also able to integrate any kind of context provider

(e.g., sensor or mobile devices) in a ”plug-and-play”

manner irrespective of type, manufacturer or accuracy

in order to collect context information useful for the

mobile communication industry and network opera-

COMMUNICATE GREEN - Energy Efficient Mobile Communication

303

tors.

A good overview about context modeling ap-

proaches and reasoning techniques is given by (Bet-

tinia et al., 2010). The authors come to the conclu-

sion that hybrid context modeling approaches are a

promising direction discussing also a possible hybrid

architecture.

4 ARCHITECTURE

The architecture comprises two parts: the functional

and the system architecture.

Figure 3: Functional Architecture.

The functional architecture (as illustrated in Fig-

ure 3) encapsulates the functions of the system to

logical entities and illustrates the relations between

them. An integral part of the functional architecture

is the data – here named as context information – with

which decisions for a dynamic de- and reactivation

of network components are calculated. Based on the

context model and service classiﬁcation scheme men-

tioned above, various types of contextual information

can be considered.

The context entity collects and aggregates con-

textual information by applying two different ap-

proaches. In the net-centric context acquisition ap-

proach, data is only extracted from network compo-

nents. This approach has the advantage that already

available data (e.g., coarse location data delivered by

the HLR, VLR or RNC) can easily be integrated into

the context aggregation process. However, the user’s

view is completely ignored in this approach, so that

deactivations calculated by the system might lead to a

bad quality of experience for the user. In order to re-

duce this problem, context is also acquired using the

hybrid approach that considers context data from end

devices as well.

The decision entity uses the collected contextual

data in order to make a decision concerning de- or re-

activation of components, whereas the control entity

controls and manages de- and reactivations.

Figure 4: System Architecture.

In the system architecture (see Figure 4), the func-

tional entities are mapped to system components.

Data is directly extracted from network components

and from end devices via dedicated client applications

on modern smartphones. However, due to the mass of

different and diverse mobile end devices subscribed to

the mobile network, it is quite impossible to consider

all mobile users and operating systems. The UMTS

SIM Application Toolkit provides potentials to avoid

this problem since it allows the implementation of

value-added services on the SIM card that can acquire

different information from the network regardless of

the mobile phone in use. Because of this, context data

acquisiton possibilities via the USAT will be analyzed

and implemented as well.

The context, decision and control entity are

merged to one component called the Power Manage-

ment Component (PMC). Various options to integrate

this component into the topological structure of the

network are going to be examined. In a central archi-

tecture, for example, each mobile network provider

has a PMC in his core network controlling and man-

aging the whole infrastructure of the network. Using a

hierarchical approach, on the other hand, would lead

to many PMC nodes in the access networks controlled

by a central PMC in the core network. In a distributed

implementation, the PMC nodes are only in the access

networks organizing themselves when de- and reacti-

vating network components.

PECCS 2011 - International Conference on Pervasive and Embedded Computing and Communication Systems

304

5 EXPECTED RESULTS

The power consumption of base stations differs from

manufacturer to manufacturer. Typical values are be-

tween 750W and 3000W

per base station.

Table 1: Number of base stations in Germany.

Provider GSM UMTS Total

Deutsche Telekom 25000 11000 36000

Vodafone 20000 13000 33000

O2 17000 10000 27000

E-Plus 18804 6616 25420

80804 40616 121420

Taking the base station numbers (O2, 2010)

(KPN, 2009) (Vodafone, 2009) (Flatrate To Go, 2009)

in Table 1 as a reference, the total power consump-

tion of base stations in Germany is between 760 GWh

and 3040 GWh per year meaning that 410.000 to

1.645.000 tons of carbon dioxide is emitted in Ger-

many per year

(BDEW, 2008). For 2020, (The Cli-

mate Group, 2008) estimates a carbon emission of

349.000.000 tons for the telecommunication infras-

tructure, whereas the mobile environment will con-

tribute with 51%. (Remark: In 2002, the carbon

emission of the telecommunication infrastructure was

151.000.000 tons and the mobile environment cov-

ered 43% of it.) With the optimizations being done

in the mobile network, we estimate a reduction of the

operating time of network components up to 40% –

60%, i.e., a reduction of the carbon dioxide emission

up to 328.000 to 1.974.000 tons per year.

6 CONCLUSIONS

The objective of the project Communicate Green

comprises the development of an adaptive and

context-aware power management. The decision and

adaptation algorithms that are going to be imple-

mented throughout the project, are going to save en-

ergy in mobile networks by dynamically de- and re-

activating network components and by reconﬁguring

the network to user needs. Optimizations will be pro-

posed for single and heterogeneous radio technolo-

gies as well as for the core network. Our part of the

project concentrates on the development of a context

Numbers extracted from data sheets of the companies

Ericsson, Motorola, Nokia Siemens Networks and Huawei

Calculation is based on the average value for carbon

dioxide emission in Germany 2007, 541 g/kWh

entity that collects and processes contextual informa-

tion from various context sources in order to build the

data basis for decision calculations. The expected re-

sults as listed above show that the projects approach

is innovative and promises an enormous reduction of

the power consumption in future mobile networks.

ACKNOWLEDGEMENTS

We thank all partners in the consortium of the

Communicate Green project funded by the Federal

Ministry of Economics and Technology (BMWi):

Deutsche Telekom Laboratories, Ericsson GmbH,

Fraunhofer Gesellschaft zur F

orderung der ange-

wandten Forschung e.V., Universit

at Paderborn,

Technische Universit

at Berlin.

REFERENCES

BDEW (2008). Datenerhebung 2007 - Bundesmix 2007.

http://tinyurl.com/36gmp3g, Accessed: 2010/12/15.

Bettinia et al. (2010). A survey of context modelling and

reasoning techniques. Pervasive and Mobile Comput-

ing, 6(2):161–180.

EARTH (2010). Energy Aware Radio and Network

Technologies. https://www.ict-earth.eu/, Accessed:

2010/12/14.

Flatrate To Go (2009). UMTS Flatrates auf dem Vormarsch.

http://tinyurl.com/2ukcdmo, Accessed: 2010/12/14.

Flor

een et al. (2005). Towards a context management

framework for mobilife. In In IST Mobile & Wireless

Communications Summit.

KPN (2009). Annual Report. http://tinyurl.com/35ywzkc,

Accessed: 2010/12/14.

Mannweiler, C., Klein, A., Schneider, J., and Schotten, H.

(2009). Exploiting user and network context for intel-

ligent radio network access. In Proceedings of ICUMT

International Conference on Ultra Modern Telecom-

munications, St. Petersburg, Russia.

MobiLife (2006). Life goes mobile. http://www.ist-

mobilife.org/, Accessed: 2010/12/15.

O2 (2010). Press Release, O2 with highest customer

growth in 2009. http://tinyurl.com/23damjm, Ac-

cessed: 2010/12/14.

Schneider et al. (2010). Environmental context detection for

context-aware systems. ICaST - ICST’s Global Com-

munity Magazine. http://tinyurl.com/6cjdws2.

The Climate Group (2008). Smart 2020: Enabling the low

carbon economy in the information age. Technical re-

port, Global eSustainability Initiative (GeSI).

Vodafone (2009). Press Release, F

unf Jahre UMTS.

http://tinyurl.com/36bpwxe, Accessed: 2010/12/14.

COMMUNICATE GREEN - Energy Efficient Mobile Communication

305