a novel dimension of cooperation in 4g
TRANSCRIPT
-
8/2/2019 A Novel Dimension of Cooperation in 4G
1/12
A NovelDimensionof Cooperationin 4GUSMAN JAVAID, TINKU RASHEED, DJAMAL-EDDINE MEDDOUR, TOUFIK AHMED, AND NEELI R. PRASAD
The Fourth Generation
(4G) o wireless net-
works is no longer a
dream; it is knocking
now at the doors o our
inormation village. 4G promises
to oer a vast range and diversity
o converged devices, services, and
networks and to revolutionize the
way we communicate. 4G would
inuence todays networking ar-
chitecture where the interuser
communication is realized with the
help o thirdparty communication
inrastructure. In 4G, the central-
ized thirdparty controlled net-
working architecture can emerge
into a hybrid model, where a part
o usertouser interaction would beenvisaged by short/medium range
wireless communication systems.
Moreover, 4G will not only o-
er ultrahigh datarates but would
also enable a ubiquitous computing
paradigm, particularly interest-
ing or the enduser with the help
o various personalized and user-
riendly services. This increase in
shortrange communication among
users and the introduction o such
Digital Object Identifier 10.1109/MTS.2008.918037todd davidson/stockillustration rf/getty images
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008 19324529/08/$25.002008IEEE | 29
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
2/12
personalized services could orm a the emergence o client/server ar- wireless communications were not
personal ubiquitous environment chitecture, with a relatively small sure about the useulness o their
(PUE) around the user. Since in number o privileged servers serv- work and were underestimating
such environments, multiple us- ing a huge mass o consumer hosts. the power o wireless. They might
ers will come closer (without any This architecture was the opposite have envisaged that without the es-
thirdparty barriers); their coop- o the undamental design o the In- sence o cooperation and sharing,
eration will be the key to the suc- ternet, i.e., a cooperative network no technology can be economically
cess o 4G. o peers. However, in late 1990s, and socially viable. Cooperation inwith the appearance o musicshar- wireless technologies is the key to
Importance of Cooperation ing applications such as Napster, discovering a variety o unoreseenHistorical Perspective the Internet experienced another innovative applications [1], [2].The development o the modern day drastic change, and the architectur This latter aspect is the core rea-
computer was the result o techno- al design o the Internet reverted to son why cooperation is gradually
logical advancements and the hu- its original peertopeer notion. increasing with increments in the
man need to compute. In the begin The millions o hosts connected to generation o mobile systems, i.e.,
ning o the computing wave when the Internet, inspired by the culture 1G, 2G, 3G, and now 4G [3]. Today
computers were frst introduced, o cooperation and openness, start- we are at the doorstep o 4G sys-
they were gigantic and were gen- ed connecting to each other direct- tems, where collaborative services,
erally reerred to as mainrames ly, orming collaborative groups, technologies, environments, and so
or as a central data repository, sharing their resources to become on, are the major areas o researchlinked to users through less power- usercreated powerul inormation concern.
ul devices such as workstations. It clusters.
was a general belie that these com Currently, peertopeer applica Towards Cooperation in 4Gputing devices were specialized tions are using the Internet much In contrast to what was origi-
machines developed to ulfll spe- as it was originally intended: a nally expected, the uture is not
cifc high computational needs and common platorm or hosts to col- limited to cellular systems and
were o no use or an ordinary user. laborate and to share inormation 4G should not be exclusively un
However, this notion was totally as equal computing peers. In 1908, derstood as a linear extension o
incorrect, and computers proved Nicola Tesla said The wireless art 3G [5]. In concrete terms, 4G is
to be userriendly and inexpen- oers greater possibilities than any more about services than ultra-
sive, extendable to meeting a large invention or discovery heretoore highspeed broadband wireless
range o user needs. Most dramatic made, and we can expect with connectivity. As predicted in
was the emergence o the Internet certitude that in the next ew years, [6] , keeping the cellular core,
which glued together so called per- wonders will be brought by its ap- the network architecture in 4G
sonal computers and introduced a plications. And so, Teslas words wil l be predominant ly extended
computer culture o cooperation, are true even today. Wireless com- to shortrange cooperative com-
sharing, openness, and trust. The munication is nothing less than munication systems. Apar t rom
tradition o cooperation ostered magic or someone who does not coverage extension, power and
by the Internet and its marriage to know how it works. It enables us to spectral eiciency, increased ca-
personal computers, gave birth to communicate anytime, anywhere pacity, and reliabil ity, this enor-
the personal computing paradigm. (there is a signal), in many orms mous lexibil ity at the user end
The personal computing age our- (data, voice ). However, wireless will help in the development o a
ished aster than any other domain, technology is not limited to com- personal ubiquitous environment
connecting hundreds o millions o munication; it can oer much more around the user. This envi ron
people all over the world, making than just a phone call. The limits ment is indeed the dream o Mark
their work available or others on o wireless communication are still Weiser, the ather o ubiquitous
the global inormation village, i.e., unpredictable and unimagined. computing [1]. The 4G service
the Internet. The ather o radio communica- and technology inrastructure
The goal o the original Inter- tion Heinrich Hertz once said I do will induce user devices to orm
net was to provide a unifed com- not think that the wireless waves I cooperative groups and share in-
munication platorm or dierent have discovered will have any prac- ormat ion and resources in order
kinds o devices and networks as tical applications. The inventor o to atta in mutua l sociotechnical
well as uture technologies, where the frst wireless telegraph system beneits. A whole collect ion o
every single host would be an equal Guglielmo Marconi said Have I unoreseen 4G cooperat ive ser-
player. However, the undamen- done the world good, or have I add- vices will enable 4G technologiestal design radically changed with ed a menace? These early giants o to recede into the background o
30 | IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
3/12
our lives [1], making us a part o
an intell igent and ubiquitous per-
sonal substrate.
Until recently, the cooperative
services in 4G systems have re-
ceived signifcant attention due to
their high degree o technological
and social exibility, considerablereedom o choice or the user, and
most importantly, the potential
megarevenues or industrial play
ers. In this article, we ocus on the
services side o cooperation in 4G
systems and discuss how these per-
sonalized personal/group services
will make use o the multitude o
wireless systems and networks
available under the auspices o 4G.
Futuristic CooperativeServices in 4GThe widely agreed upon rule or
success in 4G telecommunication
markets is to visualize a coopera-
tive service chain o multiple sup-
pliers satisying the evergrowing
requirements o end customers [9].
The evolution o 4G systems in a
multidimensional world provided a
rich platorm or deriving advanced
and innovative useroriented and
cooperative services. Embossed to
highlevel perspectives and equally
leveraging on technical dimensions,
we recognize several aspects o co-
operative services; those related
to personal (or group centric) ser-
vices, intelligent transport network
services, cooperative community
networks, and largescale ad hoc
network services. As shown in Fig.
1, these cooperative and heteroge-
neous services account or efcient
4G convergence platorms that ren-
der clearcut benefts in terms o
bandwidth, coverage, power con-
sumption, and spectrum usage.
The personal and groupcentric
communication models put orth
a multitude o interesting services,
benefting rom the cooperative
clouds ormed as a result o multi-
level social groups based on selor-
ganizing common objectives [10].
Within this context, various com-pelling services or smarthome
networking, cooperative healthcare,
etc., are shaping up. One such ser-
vice is the cooperative distribution o
media in stationary home networks,
where transparency, enabled by a
seamless and intelligent platorm,
equips the home network to become
an interdependent service ecosystemor the consumer [11]. Other services
in group communication that exploit
collaborative behavior include sym-
bolic resource sharing among com-
munication groups (or example,
usercentric dynamic content sharing
similar to popular web services like
MySpace or YouTube), and ubiq-
uitous and collaborative healthcare
monitoring at home or hospitals [6].
The Intelligent Transport network is
also an interesting setting or provid-ing collaborative 4G services rom a
user perspective, presented in Fig. 2.
IP Core
1G / 2G / 3G4G
NetworkWi-Fi / Wi-Max / UWB ...
4G UserServices Personal/Group Intelligent Cooperative
Services Transport Community
Networks
Fig. 1. Cooperation in 4G services perspective.
The most interesting among
these services is the development
o evolutionary cooperative multi-
player games as a massive collab-
orative constellation or vehicular
networks [12]. These selevolving
games are targeted at intelligent
transport networks that range romprivate vehicle owners to public
transportation system users. Other
envisaged services include various
locationbased services oered on
a cooperative basis, where consum-
ers may either locate their intended
ootage leveraging the collaborative
platorm, or where they can market
their business availing themselves
o cooperative advertisement op-
tions. This creates an open service
ecosystem benefcial or the entireservice value chain in vehicular
transportation networks [13].
Large-scaleAd-hoc
Networks
Fig. 2. Intelligent Transport Systems (ITS) A 4G service.
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008 | 31
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
4/12
32 | IEEE TECHNOLOGY AND SOCIETY MAGAZINE
made appealing progress, par
ticularly in the ield o wireless
sensor networks. Many distrib
uted applications are envisaged
in sensor networks where collab
orative computing [19] assumes
a distributed and cooperative ash
ion. It is this usercentric coopera
tion and similar issues that account
or the development o cooperative,
ubiquitous, personal communica
tion models.
ing groups. The
eration in personal/group services
may take various dimensions rang
ing rom technology and services
to sociophysiological
order to expand our discussion on
the subject o cooperation, we clas
siy into usercentric
and groupcentric cooperation.
addition, the cooperation at these
magnitudes helps us to move or
ward towards the development o
PUEs [18].
User-CentricCooperation in 4GThere is a large array o actors in
the 4G service arena such as the
user, the service/content provider,
the network operator,
bodies, researchers and so on, who
have stakes in 4Gs success. How
ever, economicallyuser is a major player, a center oFig. 3. Cooperative wireless community networks in 4G.
Wireless community networks
(commercial, public, and non
proit), as shown in Fig. 3, have
matured through the continuing
evolution o mesh networks [17],
which are now exploiting hetero-
geneity in a third generation mesh
context with the use o multipleradios (including dierent radios
or downlinkuplink), dynamic
intererence detection and avoid-
ance mechanisms, and automatic
location updating mechanisms
[14]. Along with the introduc-
tion o intercommunity net-
working eatures this has given
new dimensions to collaborative
service distribution in commu-
nity networks. New dimensions
include communitybased IPTVservices, cooperative webradio,
and collective surveillance, etc.,
apart rom common service at-
tributes like resource sharing
among users. In general, large-
scale user cooperation is an im-
portant aspect o the success o
community networks triggering
the collaborative serviceproit
chain and introducing competi-
tive dierentiation. Mobile ad
hoc networks applications have
the center stage; smart messag-
ing services or sensors, collab-
orative objects tracking, etc., to
name a ew [15].
In the search or niche markets
and opportunities or 4G, large
organizations and policy mak-
ers converge to agree that the 4Glandscape will not just be about
defning higher data rates or newer
air interaces, but rather will be
shaped by the increasing integra-
tion and interconnection o het-
erogeneous systems, with dierent
devices processing inormation or
a variety o purposes, a mix o in-
rastructures supporting transmis-
sion and a multitude o applications
working in parallel making the
most efcient use o the spectrum[16]. On the contrary, users are get-
ting more varied in the services
that they require and the modes
with which they preer to commu-
nicate and cooperate, which also
hugely inuences the uture o 4G
commercialization. These develop-
ments have led us to think along the
lines o personal/group services as
the most appealing and predomi-
nant platorm or the development
o 4G, where users collaborate in
Cooperative Personal/GroupServices in 4G SystemsThe marriage o the ourishing
personal computing paradigm and
the networking world, gave birth
to a new era o computing called
ubiquitous computing [1]. The
ubiquitous computing paradigmcan also be seen as a byproduct o
4G systems. 4G is not the name o
a single technology [7], rather it
is a cooperative platorm [5], [8]
where a large range o heteroge-
neous wireless networks and ser-
vices coexist. Under the auspices o
4G, diverse devices, networks, and
service elements fnd their way into
the lie o the end user. The integra-
tion o 4G elements into the end
user environment should ideally gounnoticed to the user, so that the
technology eventually ocuses on
the user rather than the user ocus-
ing on the diversity and complexity
o the technology around him. It is
clear that this preerably invisible
and intelligent world o calm 4G
technology [1] integrated into the
users world is only possible with
cooperation, sharing, openness, and
trust within the users own devices,
and among the users that are orm-
notion o coop-
aspects. In
cooperation
In
regulatory
speaking, the
| SPRING 2008
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
5/12
the entire 4G globe, whereas the
other actors join hands to meet the
expectations o the end user. Tak-
ing the technological dimension,
in the last ew years, a number o
heterogeneous devices emerged
and networked, ranging rom mo-
bile communication equipmentto home electronics. This proli-
namically adjusting themselves ac-
cordingly. For instance, i the user
receives a video call while at home
sitting in his TV lounge, the mobile
phone should intelligently detect the
activity/mood o the user and should
propose to transer the video ow on
the higher resolution display placedin ront o him. Both o these di-
towards personalization and user-
centric cooperation, we generalize
the concept o PCs and extend it
towards personal networks (PNs)
[18], frst introduced in the EU
IST MAGNET project. A PN is
a system/network owned and op-
erated by one person, i.e., the PNowner. The PN owner is the sole
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008 | 33
eration results in a large range o
choices available to the user to
communicate in highly diverse
environments. As a result, in a 4G
system, the user is surrounded by a
variety o devices oering a mul-
tiplicity o services [21], as shown
in Fig. 4. Moreover, the utilization
o these devices and services dra-
matically changes with the change
in a users environment. Thereore,the devices and services in the 4G
world should have strong adapta-
tion capabilities.
Personalization [3] is a key
term here. Since every user is
unique in their roles, tastes, and
preerences, 4G systems should be
intelligent enough to ully under-
stand the user and adapt the net-
work and service elements accord-
ing to the users preerences.
In a usercentric model, the user
is the ocus o the whole system.
The cooperation among his het-
erogeneous devices and his envi-
ronment is vital or the seamless
working o the entire 4G system.
Here, we reer to cooperation in
two dimensions. At frst, the de
vices themselves need to cooper-
ate, or instance, while the user is
busy working on his laptop and he
receives an important voice mes-
sage on his mobile phone, the mo-
bile phone should track the activity
o the user in order to notiy him
about the voice message. To this
end, irrespective o their specifca-
tions, the users devices should be
able to cooperate in order to help
the user in his daily lie.
Second, the devices should co-
operate with the users environ
ment. Since user preerences vary
with a changing environment, thedevices should be capable o dy-
mensions o cooperation are only
possible when the 4G systems en-
circling the distinct end user ully
understand the sociophysiological
and the technological potentials
and limitations o cooperation.
The personalization aspects
o 4G systems are largely similar
to the early concept o personal
computers (PCs). In the 4G era,
authority in his personal inter-
connected devices and can use
the PN in the way he wants. The
personal devices may be located,
both in his close vicinity and at re-
mote locations. Fig. 5 presents the
PN o Bob, which is composed o
his home, ofce, and car clusters.
The owner o the PN can add new
devices or personalized services
Fig. 4. User-centric cooperation.
u
d
e
Fig. 5. Bobs personal network.
Bb H c
Bb o c
Bb c c
Bb
i
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
6/12
to their personal network at will.
For its owner, the PN is a heaven
o personalized services in cyber-
space. But to the outside world, it
appears as a black box.
Group-Centric
Cooperation in 4GHuman nature does not promote
living in isolation. The emer-
gence o communication networks
is alone proo o it. Groupcentric
cooperation is also reerred to as
cooperation among the end users
who are organized in groups. This
is in many ways undamentally
opposite to usercentric coopera-
tion, where only the users devices
and environments cooperate, and
their cooperation appears as a darkcloud to the outside world (to other
users). In act, the 4G services that
can be made available to a single
user (with usercentric coopera-
tion) are limited, and users need to
cooperate with each other to extend
their global services repository. In
addition, many serviceoriented
patterns need to extend the bound
In order to promote groupcen-
tric cooperation in 4G systems,
the concept o personal network
ederations (PNF) [22] has been
recently introduced in the EU IST
MAGNET Beyond project (Phase
II o the MAGNET project). PN
Fs address the interactions be-tween multiple PN users with
common interests or a range o
diverse services. A PN ederation
can be defned as a secure im-
promptu, situationaware or be-
orehandagreed cooperation be-
tween a subset o relevant devices
belonging to dierent PNs or the
purpose o achieving a common
goal or service by orming an e-
fcient collaboration.
Consider the PNF B in Fig. 6: asimple example o PNF is the ed-
eration o PNs belonging to a group
o students in a classroom, sharing
lecture notes.
Based on how the coopera-
tion between devices in dierent
PNs is realized in order to estab-
lish the ederation, we can di-
erentiate between inrastructure
rastructure network. As shown
in Fig. 6, the inrastructure PNF,
i.e., PNF A, is ormed between
user 1 and user 2, who are located
across the inrastructure network.
On the other hand, in a spontane-
ous, adhoc PNF, the ederation
is ormed in the absence o a fxedinrastructure. This type o eder-
ation mostly occurs when nearby
users collaborate within a edera-
tion. PNF B in Fig. 6 presents a
spontaneous PNF ormed among
user 3, 4, and 5.
Towards Personal UbiquitousEnvironments in 4G SystemsAs discussed in the previous
sections, both usercentric and
groupcentric cooperation are re-quired in order to meet the long-
term expectations o a 4Genabled
ubiquitous computing world. Co-
operation among the users, their
devices, and their environments
results in the development o a
personal ubiquitous environment
around the user, which permits
ubiquitous global access to a
34 | IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008
aries o usercentric cooperation
and involve the secure interaction
o multiple users having common
interests or various proessional
and private services.
and spontaneous PN ederations
[22]. In an inrastructurebased
ederation, a PNF is established
between devices in PN clusters
that are all connected to an in-
vast number and variety o inor-
mation resources [18], [20]. This
uniorm and comprehensive sense
o cooperation results into a vast
base o services or all the users
who are the part o this PUE vil-
lage. In the language o personal
networking, we collectively defne
PN (personal network) and PN-
F (PN ederation) as a personal
ubiquitous environment (PUE).
As shown in Fig. 7, three us
ers join hands to share devices,
services, and environments to
orm a cooperative group (PUE
/PNF). In PUE space, the users
believe in the essence o open-
ness and sharing not only or
their selcentric goals but also
or the global beneits o the
entire cooperative community.
Those users, who are satisied
with their own proper resources
and do not have any intention o
cooperating, stay in their own
usercentric environments, i.e.,PN, as highlighted in Fig. 7.Fig. 6. Personal network federation architectures.
Pn-f a(i)
Pn-f B(sp)
u 3u 4
u 5
u 1
u 2
u 6
i
s
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
7/12
Socio-TechnicalDimensions Potentialsand Limits of CooperationSociotechnical dimensions may
either limit or support coopera-
tion in 4G collaborative services.
We ocus on the PUE as one o
the oremost services arena in 4G.The PUE is particularly interest-
ing in terms o social implications
o cooperation because the user in
a PUE is totally ree to cooperate
within his own network and with
others, without going through cer-
tain rigid sets o obligations rom
the service provider or the network
operator (which is the case in cellu-
lar/inrastructurebased networks
today). In PUEs, we consider a sce-
nario in which a group o users arelocated within each others spatial
proximity and are open to cooper-
ate and share services and applica-
tions. However, some basic ques-
tions may arise here. For example,
why does the user want to extend
his PUE in order to accommodate
other users, what is he interested in
and more importantly, what would
he be able to get ater orming the
PNF with other users, and fnally,
what price might the user have to
pay or these services. We base our
discussion around three undamen-
tal stances outlined in the ollow-
groups, i the user eels satisfed
with the services he has in his own
PN, no desire to cooperate and to
orm groups will arise. The user
shall only devise ways into coop-
eration when he looks or some
service that his own PN (or current
PUE) cannot oer. The users in-tent to cooperate can be classifed
in several ways: purposedriven
cooperation vs. opportunitydriven
cooperation, shortlived coopera-
tion vs. longerterm cooperation,
and proactive cooperation vs. reac-
tive cooperation.
Purposedriven cooperation
means that the cooperative strate-
gies are explicitly defned beore-
hand, whereas opportunitydriven
means that the users cooperatespontaneously when interesting
circumstances to do so arise. In
both cases, and especially in the
second, inormation about the us-
ers context/environment/activities
can play an important role. Next,
depending on the lietime o the
cooperative groups, we can make
the distinction between very short-
lived cooperation and longerterm
cooperation. This distinction will
have its implications on the com-
plexity o the solutions to establish
the cooperative groups. In the case
o shortlived cooperative groups,
solutions to set up and manage the
cooperation need to be lightweight
and simple. Longer term coopera-
tion opens up many more opportu-
nities to introduce more complex
and powerul management and de-
inition mechanisms. Finally, basedon the way the cooperation process
is carried out, both proactive and
reactive cooperative groups are
possible. Proactive implies that
the cooperative groups are estab-
lished in anticipation o the use o
the common goals or services pro-
vided by the cooperation strategies
o each group user. Last but not the
least, reactive cooperative groups
are established upon request or
when the opportunity arises.
Formation ofCooperative GroupsIn precise terms, a cooperative
group is a unction o cooperation
strategies defned by each partici-
pant o the group. First the group
members defne their local strate-
gies and exchange them with the
other members. The exchange o
strategies is similar to negotiation
between the endusers, i.e., what
each o the users wants to pro
vide and consume as a part o the
ing three subsections.
Before theCooperation BeginsThe PUE o a user frst consti-
tutes his own devices and services
available in his PN. The user is the
sole authority to extend his PUE
(to orm a PNF) in order to ac-
commodate the services and the
devices available to other users in
their own PNs. However, beore
really moving towards cooperat-
ing and orming groups, the user
frst looks at his motivation to co-
operate. Adam Smith, the ather
o modern economics said, Every
man, as long as he does not violate
the laws o justice, is let perectly
ree to pursue his own interest hisown way. In terms o cooperative Fig. 7. Some personal networks interact; some do not.
P ubqe
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008 | 35
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
8/12
A cooperative group canhave multiple convergence points.
cooperative group. For instance, as gotiate on the terms and conditions
shown in Fig. 8, there are three dis o the PNF. As an outcome o thistinct PNs who want to orm a co negotiation, all o the potential co-
operative group (a PN Federation). operative group (PNF) members
o strategies), reerred to in Fig. 8
as the convergence point. Once
the convergence point is attained,
i.e., the common strategies or the
cooperative groups are defned,
then the cooperative groups are ac-
tually ormed.
Cooperative groups may varyon dierent scales such as age,
proession, likes, needs, culture,
36 | IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008
Beore orming the group, they ne converge at a certain point (a group and so on. Thus, it is less likely
at times that they converge on a
single point. The derivation o
common strategies or the entire
group gets more complicated and
and any increase in the number
o members o the cooperative
group. Moreover, even i the
members fnally converge to cer-
tain agreed upon strategies o thegroup, the time it would take to
orm a group would be consider-
able. Thereore, it may be quite
efcient that certain group mem-
bers converge on some strategies
and do not converge on others. Or
it is also possible that the coop-
erative group defnes one single
strategy as a general strategy
or the group, and other specifc
strategies or cooperation among
group members.
To this end, a cooperative group
can have multiple convergence
points. As in Fig. 9, PN1 defnes
two disjointed convergence points
with each o the other PNs (i.e., PN-
2 and PN3) in the group. In con-
crete service terms, in the scenario
considered in Fig. 9, the coopera-
tive group is ormed by the PN1 to
consume/provide service to each o
the other PNs, whereas other PNs,
i.e., PN2 and PN3 might not be
interested in each others services.
Thereore, in order or the group to
achieve its goal, the convergence
points o PN1 with other PNs are
essential. However, in this case, a
more complex problem is to pro
vide a secure and efcient interace
between each o the convergence
points defned within the scope o
the cooperative groups. Moreover,
during the lietime o the coopera-tive group, due to the dynamism o
Fig. 8. Cooperation among PNs.
Pn-3 Pn-3
Pn-3
c
Fig. 9. Cooperation among PNs.
cPn-1/Pn-2
c
Pn-1/Pn-3
Pn-3 Pn-2
Pn-1
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
9/12
the group and its members, individ-
ual strategies can change. Coping
with this dynamism in cooperative
group environments is also a hard
nut to crack.
Sharing Strategies in
Cooperative GroupsIn order to ully understand the
sharing strategies in cooperative
groups, it is interesting to observe
how the economics o cooperation
work in society in general. Coop-
eration reers to the practice o
people or greater entities working
in common with commonly agreed
upon goals and possibly methods,
instead o working separately in
competition [23]. In society, we
cooperate when we want to ac-complish something that we can
not achieve working alone. In con-
trast, sometimes we cooperate not
or obvious shortterm benefts but
or longterm gains. For instance,
User A relays the trafc o User B
so that in uture, User B would be
in a position to ask User A to relay
his trafc. This type o cooperation
involves business, cultural, and
relationship development aspects.
Whatsoever the reason behind the
cooperative behaviors is, coopera-
tion does not come or ree and we
always have to pay a certain price
or it. The cost and the gains o
cooperation can take many orms
ranging rom resources (man, mon-
ey, machines) to moral and ethical
support, reerred as the potentials
o cooperation.
Even i all members o a
group beneit rom the coopera-
tive group, individual selinter-
est may not avor cooperation.
This theory o noncooperative
behavior or selinterest in a
cooperative group is reerred as
prisoners dilemma [25]. There
can be several reasons to be non-
cooperative in a group. One o
the major reasons is associated
with the utility o being the part
o the group. Everyone wants to
have the best thing under the costconstraints he has. Thereore, the
We cooperate when we want toaccomplish something that we cannot achieve working alone.
user would be cooperative to a
certain limit where his total util-
ity o being cooperative is great-
er or equal to the cost he is pay-
ing as a part o the cooperative
group. Since the total utility and
related cost are associated with
the satisaction o the user, once
the cost bypasses the total utility
the users satisaction starts de-
creasing, and he becomes more
egoistic, or a less cooperativemember o the group.
In this section, we discuss the
potentials and limits o coopera-
tion by applying Nash Equilibrium
(NE) theory (part o game theory)
to PUE concepts. John Nash intro-
duced the concept o Nash equi-
librium in his doctorate thesis and
showed or the frst time in his dis-
sertation, Noncooperative games
(1950), that Nash equilibria must
exist or all fnite games with any
number o players [24]. In PUE,
where dierent PNs join hands to
orm a cooperative group (PNF)
in order to share certain services,
let (S, f) be a cooperative group,
where S is the set o strategy pro-
fles andfis the set o payo profles.
Let s i be a strategy profle o all
group members except or member
i. When each member o the group i
e{1n} chooses strategyxiresult-
ing in strategy proflex = (x1,...,xn)
then member i obtains payofi(x).
Note that the payo depends on
the strategy profle chosen, i.e., on
the strategy chosen by member i as
well as the strategies chosen by all
the other members. A strategy pro-
flex* eS is a Nash equilibrium ino unilateral deviation in strategy
by any single member is proftable,
that is, i or all i,
* * * *fi ( x i, x i ) fi ( x i, x i ). (1)
In descriptive terms, i there
is a set o group strategies with
the property that no group mem-
ber can beneft by changing his
strategy while the other members
keep their strategies unchanged,
then that set o agreedupon group
strategies and their corresponding
payos constitute the Nash Equi-
librium in the cooperative group.
Thereore, in a Nash Equilibrium
none o the group members canunilaterally change his strategy to
increase his payo.
We have analyzed the poten-
tials and limits o cooperation
with the help o NE theory under
multiple group members (PNs)
scenarios, who orm a Personal
Network Federation (cooperative
group). Moreover, contradicting
the basic NE concept, we have
also studied the scenarios where
multiple equilibrium points are
possible. In our study, the coop-
erative group strategies model
reerred to as consume/provide
is based on basic supply/demand
economics theory [26].
Potentials of
Cooperation in 4Gs Personal
Ubiquitous Environments
The potentials o cooperation in
PUEs are associated with the stra-
tegic satisaction o each coopera-
tive group member. This implies
that the percentages o his local
strategies are reected in the com-
mon group strategy. As discussed
earlier, towards the ormation o
cooperative groups (PNF), each
group member (PN) prepares his
proper local strategy and then ex
changes it with the other potential
group members. A group member
who frst initiates the group orma-tion process is reerred as a group
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008 | 37
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
-
8/2/2019 A Novel Dimension of Cooperation in 4G
10/12
The derivation of commonstrategies for the entire groupgets more complicated with anyincrease in the number of members
of the cooperative group.
creator (PNF creator). In concrete (QoS), economic, etc.) or certain
terms, a local strategy contains services, how much time is he will-
the inormation related to the par ing to remain a member o the co-
strategy is prepared. This strategy is
the convergence point in the entire
cooperative group space. I all the
group members agreed on a certain
group strategy to the extent that
none o the group members wants
to unilaterally change his strategy
to increase his payo, we can saythat the cooperative group has at-
tained a Nash Equilibrium point
as shown in Fig. 10. The Xaxis in
Fig. 10 represents the consume
strategy, whereas the provide
strategy is on the Yaxis. Here,
38 | IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008
ticipation o the member such as
which services he wants to con
sume/provide, what are his preer-
ences (security, quality o service
operative group, and so on.
Ater an exhaustive exchange o
local strategies among the coopera-
tive group members, a fnal group
three members (PNs) join hands to
orm a cooperative group (PNF).
PN1 is the creator o the coopera-
tive group. Ater some negotiation
o their own local consume/provide
willingness, they all agree on a cer-tain point, which is marked in Fig.
10(a) as the Equilibrium point.
Fig. 10(b) highlights a much di-
erent behavior o PN1 (creator)
in the cooperative
PN1 is the initiator o the group,
it is quite possible that he might be
more open to provide as much ser-
vice as possible to the group with
comparatively very limited desired
to consume services. This behavior
is much justifable in the society, as
a manager or the rontliner is nor-
mally the center o ocus o a group
and his behavior has a strong im-
pact on the strategies o the other
group members. Thereore, or the
success o a group, the initial strat-
egy defned by the mentor o the
group is highly important.
It is important to note that the
Nash Equilibrium point presents the
minimum set o provide strate-
gies owned by all the cooperative
group members. It is o course pos-
sible that, at the later stage o coop-
eration, one o the members may
express a generous attitude and pro
vide more services by keeping his
consume strategy constant, as can
be the case with PN1 in Fig. 10(b).
To this end, the equilibrium point
will shit keeping the entire groups
consume strategies constant. This
phenomenon o moving an equilib-rium point with a variable provide
Fig. 10. (a) (b) Potentials of cooperation in PUE Consumer/provider strategies.
Provide
Provide
Pn 1 (c)
Pn 3
Pn 2
eqb
Pn 1 (c)
Pn 3
Pn 2
c c
() (b)
Fig. 11. Potentials of cooperation in PUE Multiple equilibrium strategies.
Provide
c
Pn 1 (c)
eqb 2
eqb 1
Pn 2
Pn 3 Pn 4
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
group. Since
-
8/2/2019 A Novel Dimension of Cooperation in 4G
11/12
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008 | 39
strategy o one member, and con-
stant consume strategies o all the
other members, can be clearly stud-
ied on a threedimensional graph,
where theXaxis is consume strat-
egy, the Yaxis is provide strategy,
andZaxis is equilibrium.
As we normally see in soci-ety, some players in a group have
their own proper stakes associated
with only certain members or cer-
tain goals o the group. They stay
with the group only or such lim-
ited benefts as a part o the entire
cooperative groups ecosystem. To
this end, a group may have mul-
tiple equilibrium points satisying
all group members as a whole or
some o them. As shown in Fig. 11,
a group consists o our memberssuch as PN1, PN2, PN3, and PN-
4. Lets assume that PN4s interest
in the group is only associated with
some services oered by PN1 and
he is not interested in any other
service. In this respect, as in Fig.
11, we have two equilibrium points
such as Equilibrium1 among all
members except PN4 and Equi-
librim2 between PN1 and PN-
4. In multiple equilibrium group
cases, it is important that both the
equilibrium strategies should have
a certain level o interace among
them. As in the example in Fig. 11,
a strong communication between
both strategic equilibrium points
would monitor and control the ac-
curate working o the group. For
example, here in this example, this
interace ensures that PN4 only
consumes the services o PN1 as
defned by Equilibrium2 and does
not interact with any other group
services made available by other
members in the group.
Limits of Cooperation in
4Gs Personal Ubiquitous
Environments
Sometimes certain group members
either cooperate in a way that their
cooperation is not useul or the
group or they behave in a totally
noncooperative way (becomingegoistic). Both o these situations
are extremely egoistic in their co-
operative behaviors, i.e., they are
inclined towards consuming much
more service than oering to other
group members. This case is again
best or himsel, right? Adam Smith
was wrong! Themessage: Sometimes
it is better to cooperate. Coopera-
tion is the buzz word in the commu-
nications industry today driving the
Fig. 12. (a) (b) Limits of cooperation in PUE Egoist/ineffective cooperation.
Provide
Provide
Pn 1 (c)Pn 1 (c)
Pn 2
Pn 2
Pn 3
Pn 3
eqb
c c() (b)
Authorized licensed use limited to: Vodafone. Downloaded on November 26, 2008 at 05:33 from IEEE Xplore. Restrictions apply.
Cooperation does not comefor free and we always have topay a certain price for it.
limit cooperation in the PUE. The
ormer case is discussed in Fig.
12(a). A cooperative group has
three potential members such as
PN1, PN2, and PN3. The strat-
egies defned by PN1 and PN2
cause them to settle down to a cer-
tain equilibrium point, whereas PN-
3 is not party to the common equi-
librium. In Fig. 12(a), the strategy
o PN3 is represented by a straight
line parallel to the Yaxis (provide).This implies that or PN3, while he
is a very cooperative member o the
group, his cooperation is not inter-
esting or the other group mem-
bers. For example, PN3 is provid-
ing services that are not needed by
the other members. In this case, an
ideal equilibrium point among all
the group members is blocked by
the local strategies o PN3.
The latter case, where some
members become egoistic, is dis-
cussed in Fig. 12(b). In Fig. 12(b),
again three PNs are potential co-
operative group members. Here
we clearly see that PN2 and PN3
a bottleneck in the ormation o a
cooperative group with certain es-
sential equilibrium point(s).
One way to overcome this bot-
tleneck in cooperative groups is to
reward more or cooperative atti-
tude and to punish more or a non-
cooperative attitude. In the absence
o any reward/punish mechanism,
the noncooperative behavior will
have a titortat eect on the entire
group. For instance, i a coopera-tive members partner deects rom
cooperative behavior, the group
responds in a similar noncoop-
erative way towards other partners.
This chronic behavior will rapidly
spread within the group, and it
might end with a total noncoop-
erative group, where no member is
willing to cooperate.
Vision of theFuture for 4G SystemsIn the Hollywood flm A Beautiul
Mind, John Nash said that Adam
Smith said the best result comes rom
everyone in the group doing whats
-
8/2/2019 A Novel Dimension of Cooperation in 4G
12/12
notion o cooperative mechanisms
in uture heterogeneous systems,
including 4G. The 4G landscape is
so diverse, and the industry leaders
and strategic leaders accept that 4G
is not only about improved data rates
or diverse air interaces and unifed
standards, but rather is going to beshaped by increasing integration,
collaboration, and interconnection o
heterogeneous systems. On the con-
trary. The widely agreed upon rule
or success in 4G telecommunication
markets is to visualize a cooperative
service chain o multiple suppliers
satisying the evergrowing require-
ments o end customers. This inter-
twined and inspiring direction could
acilitate the realization o a large
scale cooperative and o ubiquitouswireless communities.
Furthermore, or personal or
group communication environ-
ments particularly, PUEs could
eventually be orerunners or
exploiting the theoretical limits
o cooperative systems, enabling
the provision o niche coopera-
tive systems and services. This
potential capability needs to be
explored in much detail, under-
standing sociotechnical aspects
and potential limits o coop-
eration, and developing efcient
models to develop and nurture co-
operative societies. Several socio-
economic aspects need to care-
ully observed and studied, such
as human actors with respect to
human nature, egocentric human
behaviors, and social actors such
as the eects on society, econom-
ic competition, etc.
From a service perspective, we
believe that the uture o coopera-
tive services in 4G largely depends
on the result o cooperation o major
players in industry including service
providers and vendors, etc., on one
hand and policy makers, academia,
etc., on the other. From a technology
perspective, our opinion is that the
largescale integration o coexist-
ing applications and incorporation
o emerging technologies, exiblemodels or spectrum allocation,
40 |
etc., should be considered in depth.
Finally, encouraging healthy inter-
working between application re-
search and technology research, and
supporting seamless cooperation
should be the vision o the uture o
4G systems.
Author InformationUsman Javaid and DjamalEd-
dine Meddour are with France
Telecom R&D, Lannion, France,
[email protected]. Tinku
Rasheed is with the CreateNet
Research Center, Trento, Italy. Toufk
Ahmed is with LaBRI Labs, Univer-
sity o Bordeaux I, Talence, France.
Neeli R. Prasad is with the Center or
TeleInFrastruktur (CTIF), Aalborg
University, Aalborg, Denmark.
References[1] M. Weiser, Does ubiquitous com-puting need interace agents? No, pre-
sented at M.I.T. Media Lab Symp. on UserInterace Agents, Invited presentation,
Oct. 1992.
[2] J.M. Pereira, Fourth Generation: Now,
it is Personal, presented at 11th IEEE Int.Symp. Personal, Indoor, and Mobile Radio
Communications (PIMRC), London, U.K.,Sept. 2000.
[3] S. Frattasi et al., 4G: A usercentric sys-
tem, presented at Mobile eConerence (Me),Aug. 2004.
[4] S. Frattasi et al., A pragmatic methodol-
ogy to design 4G: From the user to the tech-nology, presented at 5th IEEE Int. Con. on
Networking (ICN), Reunion Island, France,
Apr. 2005.
[5] E. Bohlin et al., The uture o mobile
communications in the EU: Assessing thepotential o 4G, European Science and Tech-
nology Observatory (ESTO), ESTO Publica-
tions, Feb. 2004; http://esto.jrc.es/detailshort.
cm?ID_report=1199.
[6] S. Frattasi et al., Cooperative services or
4G, presented at 14th IST Mobile and Wire-less Communications Summit, Dresden, Ger-
many, June 2005.
[7] K. Murota, NTT DoCoMo, Mobile com-
munications trends in Japan and DoCoMos
activities towards 21st century, presented at4th ACTS Mobile Communications Summit,
Sorrento, Italy, June 1999.
[8] M. Katz and F. Fitzek, Cooperative tech-
niques and principles enabling uture 4G wire-less networks, presented at IEEE EUROCON
2005, Serbia & Montenegro, Belgrade, Nov.2005.
[9] G. Roussos, Endtoend service architec-
tures or 4G mobile systems, in The Path to4G Mobile. London, U.K.: IIR, 2003.
[10] C. Sandvig, Assessing cooperative ac-
tion in 802.11 networks, presented at 31st Int.Con. Communication, Inormation, and In-
ternet Policy, Washington, DC, Sept. 2003.
[11] Digital Living Network Al liance, DLNAoverview and vision whitepaper, DLNA,
2006.
[12] B. Knutsson et al., Peertopeer support
or massively multiplayer games, presented at23rd Ann. IEEE Con. Computer Communi-
cations (Inocom), Hong Kong, 2004.
[13] Wind River, Automotive Services Prac-tice; http://www.windriver.com/services/au -
tomotivepractice/, accessed Jan. 5, 2008.
[14] Mesh Dynamics, Third Generation MeshTechnology; http://www.meshdynamics.com/
third_generation.html, accessed Jan. 5, 2008.
[15] C. Borcea et al., Cooperative computingor distributed embedded systems, in Proc.
22nd Int. Conf. Distributed Computing Sys-
tems (ICDCS), Vienna, Austria, 2002.
[16] European Commission ICT Report, Note
on the uture o mobile communications in theEU, Sevilla, Spain, Nov. 2004.
[17] H. Moustaa et al., A panorama on wire-less mesh networks: Architectures, applica-
tions and technical challenges, presented atInt. Wkshop. on Wireless Mesh: Moving to-
wards Applications (Wimeshnets), Waterloo,
Canada, Aug. 2006.
[18] U. Javaid et al., Personal network rout-ing protocol (PNRP) or personal ubiquitous
environments, presented at IEEE Int. Con.on Communications (ICC), Glasgow, U.K.,
June 2007.
[19] T. Rasheed et al., Clusterquality based
hybrid routing in large scale mobile multihopnetworks, presented at IEEE Wireless Com-
munications and Networking Con. (WCNC),Hong Kong, Mar. 2007.
[20] U. Javaid et al., A proflebased personal
network architecture or personal ubiquitousenvironments, presented at IEEE Vehicular
Technology Con. (VTC), Dublin, Ireland,
Apr. 2007.
[21] U. Javaid et al., Towards universal con-vergence in heterogeneous wireless networks
using ad hoc connectivity, presented at 9thInt. Con. Wireless Personal Multimedia
Communications (WPMC), San Diego, CA,
Sept. 2006.
[22] J. Hoebeke et al., Personal networks eder-
ations, presented at 15th IST Mobile and Wire-
less Summit, Myconos, Greece, June 2006.
[23] Wikipedia, The Free Encyclopedia Co-operation; http://en.wikipedia.org/wiki/Co-
operation, accessed Jan. 5, 2008.
[24] J. Nash, Noncooperative games,Annals
Mathematics, vol. 54, no. 2, pp. 286295, 1951.
[25] C. Holt and M. Capra, Classroom games:
A prisoners dilemma, J. Economic Educa-tion, Sum. 2000.
[26] S. Landsburg, Supply, demand, and
equilibrium, in Price Theory and Applica-
tions, 5th ed., Southwestern , 2002, ch. 1 pp.126.
IEEE TECHNOLOGY AND SOCIETY MAGAZINE | SPRING 2008
mailto:[email protected]://esto.jrc.es/detailshorthttp://www.windriver.com/services/auhttp:///reader/full/http://www.meshdynamics.comhttp://en.wikipedia.org/wiki/Comailto:[email protected]://esto.jrc.es/detailshorthttp://www.windriver.com/services/auhttp:///reader/full/http://www.meshdynamics.comhttp://en.wikipedia.org/wiki/Co