ROLE BASED CROSS-LAYER COMMUNITIES ON WMN
João Paulo Barraca, Susana Sargento and Rui Aguiar
Instituto de Telecomunicações, Campus Universitário de Santiago, Aveiro, Portugal
Department of Electronics, Telecomunications and Informatics, Universidade de Aveiro, Aveiro, Portugal
Keywords: Communities, cross-layer, cooperation, access control, mesh networks.
Abstract: The community notion can be exploited as a rational concept leading users to cooperate in sharing resources
on Wireless Mesh Networks. We propose a novel concept for self-organizing networks, where multiple
entities (network elements or users) collaborate to achieve common goals, and in particular, to establish the
basic connectivity and service delivery infrastructures. The resulting architecture is based on wireless mesh
communications, with different entities taking different roles in the communities in a cross-layer approach.
These communities can collaborate, leading to increasingly complex and geographically extended scenarios.
1 INTRODUCTION
Wireless mesh networks are able to provide
broadband access using an adaptable wireless
infrastructure, directly tied to user interaction. Mesh
networks exploit ad-hoc networks concepts where
self-organization, cooperation and distributed
operation are vital concepts. Its usage scenarios are
usually related to the provision of broadband
network access to distant rural areas or in dense
urban environments (Borcoci, 2007), being actively
exploited by several companies.
In rural areas, far from any wired network
access, meshed operation using technologies such as
802.16 can cover areas as large as 1000km
2
,
providing reasonable broadband access to
geographically distant users. On the other hand, in
metropolitan environments, meshed technologies
enable the deployment of cost effective networks
with added benefits in terms of its adaptability to
user density, self-organization and self-healing
capabilities, as well as the capacity of favouring
content creation by users. This last benefit is
expected to be the centre of next generation
networks where the user focuses on direct user
interaction rather than content provision by large
content providers.
A fully distributed architecture enables cost
effective scalability at the cost of lower
predictability as stated in (Dressle, 2006). Mesh
networks can evolve by combining the thousands of
access points (AP) existing in metropolitan areas
into a single adaptable network. In terms of
adaptability and resilience this is the ideal move;
however, the approach leads to a large degree of
heterogeneity in access conditions as well as
decrease in reliability. Furthermore, it is not
expected that both users and operators deploying
wireless equipment let the mesh operate in a fully
distributed (uncontrolled) manner, retaining some
control either by personal or monetary reasons.
To address the management of these
environments in a distributed and efficient way, this
paper proposes a community-oriented architecture,
developed inside the IST-WIP (IST-WIP) project.
The community approach envisioned considers that
communities are formed across different layers with
the possibility of having different objectives, e.g.,
wireless resource sharing, routing support, exchange
of specific distributed application-layer services. In
this sense, we propose a cross-layer approach for the
communities’ creation and management: the
communication between several nodes requires a
cross-layer interaction and agreements between
different types of communities, where this
interaction is addressed in a similar way across
layers. We also describe the benefits and the main
impact, on a community driven architecture, and the
application of the generalized community concept.
This paper is organized as follows. Section 2
presents a simplified vision of the mesh network
architecture. Section 3 discusses the communities’
concepts, the roles and rules applied to the
community elements, and the cross-layer approach.
85
Paulo Barraca J., Sargento S. and Aguiar R. (2007).
ROLE BASED CROSS-LAYER COMMUNITIES ON WMN.
In Proceedings of the Second International Conference on Wireless Information Networks and Systems, pages 85-92
DOI: 10.5220/0002150000850092
Copyright
c
 SciTePress

The management process of the communities is
addressed in section 4 for intra-community, and in
section 5 for inter-community interactions. Section 6
describes some interesting business models achieved
with this approach, and section 7 presents the final
conclusions.
2 WIP ARCHITECTURE
The WIP global network is structured based on a
wireless backhaul, optimized for high performance
forwarding between fixed access points that provide
overall connectivity to highly mobile terminals
organized into spontaneous sub-networks. The
wireless backhaul will make use of sophisticated
techniques for high performance transmission
(directional antennas, multiple radios), and
technologies such as 802.16. For practical reasons,
mobile terminals need to live with more traditional
wireless solutions such as omni-directional antennas
and standard 802.11a/b/g wireless equipments on
off-the-shelf laptops.
The wireless backhaul presents some similarities
with the current Internet: access points are fixed
end-points and its role recalls the functions of the
core Internet. In fact, multiple accesses to current
Internet can be pruned in this backhaul. However,
for spontaneous sub-network formation of mobile
terminals, we also need components designed from
scratch to deal with mobility and self-organization.
Figure 1: Topological WIP architecture.
The nodes in the WIP architecture contribute to
the operation of the network by participating in
traffic forwarding and by providing some local
resources and services. Their contribution may
depend on the state of local resources and,
fundamentally, on the node’ willingness to share
them. Their operation is autonomic, i.e. they operate
without explicit human intervention (e.g. nodes may
choose frequency bands, time schedules and
topologically valid addresses…).
In order for end-users to decide to participate
and contribute to the wide coverage of this radio
internet, it is important that it is designed so as to be
an attractive alternative to the wired Internet for
private or group communications. Towards this end,
the WIP project is focused on developing
approaches concerning mobility management and
routing avoiding the restrictions imposed by the
current Internet architecture. Moreover, it aims at
building an architecture that enables the spontaneous
creation of several types of communities (existing or
novel ones) allowing them to self-organize and
customize the network functionality to suit their own
needs. Also, this community notion can be used in
the creation and efficient operation of a mesh
network. First, as a general concept, the notion of a
community could be used for efficiently
implementing aspects of the network (or lower)
layers’ functionality. Additionally, carefully
designed communities could provide adequate
incentives to users (and nodes) to participate in the
WIP network sharing their available resources (their
wireless access points, their Internet access), which
in most cases is necessary for ensuring a wide
coverage of this type of wireless networks.
3 COMMUNITIES
In our generalized community concept, communities
are defined as a set of entities that collaborate or
cooperate. The concept is traditionally used to
denominate a group of users united by a common
interest, and collaborating towards a predefined
objective. Examples of such collaborating
communities can range from the web based
encyclopedias (Wikipedia), forums where users
exchange ideas, or comment movies (IMDB), to file
sharing networks (Emule) or distributed computing
services (SETI@Home). A seldom considered
collaboration based on communities is related to the
network elements themselves: some routing
protocols (Haas, 2002) already had similar
collaboration concepts, exploiting the ability of the
network to organize and cluster. However, currently
at this level there is no formal description of the
participation rules and interaction is somewhat
limited to a rigid task (routing, QoS, monitoring or
charging).
The next sub-sections present the roles (and its
rules) we apply to community entities, and the cross-
layer approach.
WINSYS 2007 - International Conference on Wireless Information Networks and Systems
86

3.1 Roles and Rules
Rules are vital to communities as they define the
behaviour nodes should follow, resources they can
use, and the configuration parameters they should
apply. The agreement on the same set of rules will
create a coherent environment, even if members
have different resources or provide different
services. All communications between nodes in the
Internet share a “common context” with rules and
roles (server-client). In our approach, the rules
composing a role do not define complete
communication mechanisms (such as IP, TCP or
802.11). Instead they define the policies governing
the underlying mechanisms as well as the available
services.
Examples of rules would be the wireless channel
to use, the key and mechanism to cipher packets, the
bandwidth to use for P2P file sharing, or the
permissions regarding a particular service.
Taking as an example a neighbourhood
community grouping all APs and terminals in a city,
this community contains different equipments
operating autonomously and with different
resources. The rules applied to each community
member may suffer some changes. In this scenario,
as an example, APs can have a rule stating the
possibility of selecting the wireless channel; on the
other hand, laptops can be forbidden to choose the
wireless channel due to their reduced knowledge of
the spectrum, or participating in routing due to their
higher instability. However, they can support
enhanced services like processing and storage,
which may be unavailable to routers due to lack of
resources.
Figure 2: Heterogeneous neighbourhood community (from
Google Earth).
All the set of rules which determine a functional
entity compose a role. Depending on the interests of
the members and resources available, members can
always decide to act according to a given role. This
will imply that the community will accept the
member to perform such role, and the member will
act accordingly to the rules composing the role.
Roles can also be composed by no rules. A role with
no rules is still useful as it may condition the access
to resources, while not enforcing any specific
behaviour.
Some roles can result in actions requiring higher
trust than others (e.g., the role of a public storage
will require high trust due to privacy requirements).
A hierarchical environment is thus formed, requiring
the members to correctly perform some roles in
order to be eligible for more important roles. Also,
trust information must be obtained from each
interaction and later used in the process of role
delegation.
3.2 Cross-layer Approach
Communities can coexist at different layers,
contextualizing all types of communications; in this
sense, the community mechanisms act vertically
managing the control plane of the stack.
Also important on user driven networks, topology
will influence user interactions as new users sharing
the same interests become reachable. Conversely,
users’ interest will shape the formation and
organization of network elements in order to
optimize operation.
In a cross-layer community environment, access
points and backhaul routers cluster in communities
in order to optimize the scarce wireless resources,
and users establish their upper layer communities
above the created network infrastructure. Services
related with media distribution or distributed
processing could also create their own communities
on top of the network support. Such behaviour is
currently much exploited by communities exploited
by viral marketing campaigns which promote
upcoming movies, games or products.
From this vision, network and application/user
level communities co-exist in the same scenario. For
communication to be possible, user level
communities require the existence of another
community providing the actual physical delivery or
a direct interface with the network stack. In Figure 3,
users create an application community (A) sharing
music. However, they will only be able to
communicate using two different communities to
route packets (B and C). In this case, for
ROLE BASED CROSS-LAYER COMMUNITIES ON WMN
87

communication, it is formed a concatenation of
multiple communities for transport support.
Figure 3: Community interactions.
The example can be managed considering
interconnection between communities where some
communities request delivery from other
communities. The user community (A) asks to the
routing community (B) to send a message to other
participants of the user community. The B
community accepts routing and performs the same
request to C. Note that such exchanges may be
confined to QoS parameters or traffic amounts and
even require some form of payment/rewarding
across communities. Also, some restrictions on
authorizations can be applied. These concepts will
be described in section 5.
4 COMMUNITY MANAGEMENT
The basis of the community management employed
on our architecture is derived from solutions like
RBAC (Sandhu, 1996) and more recently dRBAC
(Freudenthal, 2002) and RT0 (Li, 2001). All these
solutions propose mechanisms to control the access
to resources and services by the roles the entities
possess. Entities create roles and delegate them to
other entities when requested or required. Depending
on the entity requesting the permissions, the role can
be delegated with modifications from the original
role. These modifications are used to further restrict
the permissions of the requesting entity. It is
important to notice that an entity can never delegate
more permissions than the ones it already has.
4.1 Community Bootstrap
Entities bootstrap communities by creating a
community description (called doctrine) with a set of
roles and eventually some general rules and
attributes. These rules may restrict the dimension of
the community, its physical location or impose any
requirement on participation. The entity creating the
community will have a special function as it owns
all primitive roles. After new members arrive, these
roles can be delegated. Depending on the purpose of
the community, management can be distributed only
if the owner allows such operation. This is expressed
by the right of delegation. If no right of delegation is
ever given to members, only the owner has the
capability of authorizing new entities to join or enrol
new roles.
Due to the dynamic behaviour of mesh networks,
where some nodes are mobile, a distributed
approach is required, and it is important to delegate
the assignment right to other trustworthy entities as
it increases the scalability and resilience of the
network. If a community is to emulate a service or
network provider, it is expected that only one or a
small number of entities to actually have the
capability to delegate new roles.
Because delegation chains need to be verified
until the first issuer, it is desirable to have short
delegation chains. These top nodes are first assigned
by the creator of the community and should be
chosen taking into account its attributes on
processing capabilities, stability, low mobility, and
above all, trust.
4.2 Clustering and Redundancy
When delegating a top role to an entity, the
community creator will also transmit information
about some other entities sharing the same role and
assignment permissions. With this information, the
enrolling entity, following a cluster-based approach,
will exchange information about the amount of
delegations given and the percentage of
computation, memory and wireless resources
available.
Delegations requests are then load balanced
between the several entities using this information.
Such mechanism will actually balance the delegation
tree and avoid the creation of long delegation chains.
In the occurrence of a delegation request which
should have a negative response, the answer is sent
directly not requiring forwarding to the other
entities.
Other aspect considered is the support for
delegation redundancy. In the case the top issuer of
the delegation chain is unreachable, it will be harder
for active entities to validate the delegation chain.
Without this validation, it will be impossible to
assure the delegation is still valid.
Entities may try to obtain same delegations from
different issuers following different paths. The result
will be the same delegation will be verifiable
following different paths. Due to overhead
constrains, this operation should be preformed only
when the community is idle.
WINSYS 2007 - International Conference on Wireless Information Networks and Systems
88

4.3 Community membership
The Member role is the one every entity must enrol
in order to participate on the community. When a
node wishes to join a given community it must
locate a member and issue a request to join. If the
member has been delegated the capability of
assignment for this role, and the candidate follows
the requirements of the Member role, the member
may allow it to participate. Otherwise the member
may indicate other entity (such as the issuer of its
Member delegation) for the candidate to contact.
Figure 4 depicts a sequence chart occurring when
Joana and Peter want to join the community C.
Maria was already authorized by John, which is the
community creator. Notice that this example could
be performed through different types of entities,
running lower layer protocols, such as APs.
Figure 4: Joining a community through delegation.
All members store the delegations provided to
them on their Wallet. As proposed in (Freudenthal,
2002), this Wallet acts as a repository for the
delegations issued. Entities may also publish their
delegations to the rest of the network creating a
distributed Wallet. Entities are capable of searching
other’s Wallets or enumerate proofs by a set of
parameters. Moreover, when asked for a proof, any
entity can either provide the proof, or return a
pointer to another Wallet containing the required
proof. This process continues until the proof is found
or the entity returns no pointer.
Distributed algorithms already exist in the
literature to efficiently locate objects on a distributed
Wallet, having especial relevancy the ones based on
Distributed Hash Tables. It should also be noticed
that the same proof can be simultaneously mirrored
at different Wallets. Such behaviour will enable both
redundancy and faster location of the required
information.
4.4 Resource Management
One important functionality for mesh networks is the
possibility of adaptation to network conditions.
Current proposals address this issue by providing
solutions each managing a specific layer of the IP
stack. Solutions based on zone routing (Haas, 2002)
cluster route dissemination information taking in
consideration the location of nodes. Others
efficiently allocate the best wireless channel based
on local estimation of interference (Ramachandran,
2006).
The community doctrine allows the description
of formal rules stating the configurations to follow,
but most importantly, they allow the real time
negotiation of the best parameters with subsequent
propagation of changes. This can be performed
independently of the stack layer their resources refer
to. Moreover, changes can be either global or affect
only the entities belonging to the same role. One
example is dynamic negotiation of the wireless
parameters for backhaul transport, without changing
the same parameters at the user part. Today this is
possible, but not formally integrated and following a
role based, and secure model.
In the scenarios we envision, management of
resources can be performed using different methods.
For a start, the creator of the community, having
control over all roles and configurations issued, can
force any rule into the community. Such action will
result in a centralized management and should be
only used when the community follows such
management. Because members are free to create
their own communities, if the community creator
oppresses its members, it risks to loose control over
the members.
In community with partially distributed
management, entities can request their delegation
issuers for a specific change either to a role or to the
community. Because the specific attribute or rule
can be defined either by any entity in the delegation
chain, the request is propagated until it reaches the
entity responsible. On a community with centralized
management, this entity is the community creator.
If an offending rule or attribute is stated by a
self-signed delegation (an entity somewhere in the
middle of the delegation chain), this entity decides
upon the request and either changes the role or
denies the request.
If management is distributed among a list of
entities, and the request targets a community
ROLE BASED CROSS-LAYER COMMUNITIES ON WMN
89

attribute or rule, an election process takes place. The
first managing node to receive the request checks if
the request is valid. This can concern the verification
of the number of elections requested by the entity, or
any other constrain stated on the doctrine. Then, it
requests all the entities with the management role to
vote. These entities also forward the request to other
entities they know and wait for a reply. Duplicate
requests for the same election are obviously
dropped. Each entity then takes a decision and
replies with a signed response to the entity which
sent or forwarded the request. Votes are kept at each
forwarding entity and the accumulated result is
propagated. Any entity can request the individual
votes to check if the values reported are according to
the votes received. The entity leading the election
will then receive the result of all votes and issue a
verdict to the requesting entity. It will also send the
result to the voting entities which should apply it. In
this phase it is still possible to request verification of
the votes reported. Corruption of such information
will result in severe punishment for the offending
entity and cancelling of the election. If the election
ran favourably to the request, roles are updated with
new delegations being issued, replacing existing
delegations.
The case of fully distributed management, the
process occurs in the same manner, only it affects all
entities. Because this process consumes much
processing and bandwidth resources, the assignment
of the voting right should be restricted. Alternatively
the number of elections per entity should also be
restricted. In small community with only a few
entities, the process is efficient and capable of
rapidly propagating changes to their neighbours.
5 INTER COMMUNITY
AGREEMENTS
One of the main advantages of wireless mesh or ad-
hoc technologies is their high adaptability to events,
along with the fact that distributed algorithms can be
deployed in order to manage the network in an
autonomic manner. These networks are self-
organized and self-managed, making use of
contextualized management as defined in their
community doctrine. Mesh communities are able to
dynamically, and without user intervention,
negotiate and trade services or perform peering
agreements. The issues regarding distributed
operation and self-management capabilities are
related to resources and trust.
For a node to be able to properly decide if an
agreement is to be established, it must possess
knowledge on current community status, which thus
may require a distributed knowledge base. Also it
must have the permissions to actually perform such
decision or take part on the negotiation. More
resourceful peers will be responsible for actually
storing and maintaining the information consistency.
Following the role based access model,
interoperation between communities is expressed as
the existence of a valid delegation. The result is all
inter-community management is performed in the
same manner. So, overlay communities are treated in
a similar manner to neighbourhood communities.
The main difference between the two is members of
two neighbouring communities are only members of
one community. If one community is overlaid on
another, at least one entity belongs to both
communities at the same time. Thus it may have
different permissions inherent to the role of member.
Peter Maria JohnJoana
Maria -> A.forwarding
C.member -> A.forwarding?
C.member -> A.forwarding'
Store in Wallet
Maria -> C.member?
[Maria -> C.member] John
Store in Wallet
[Maria -> A.forwarding] Peter
Data Packet
Forward
Store in Wallet
Figure 5: Forwarding of packets between different
communities.
In the example depicted in Figure 5, Maria and
John are members of the community C while Joana
and Peter are members of community A. When
Maria asks Peter to forward a packet, first Peter
checks if Maria is really a member of C. Then,
because he has no delegation authorizing
communication with other communities, it asks
Joana about the delegation. In the request it asks if
packets from a member of community C can be
forwarded by a member of A. Joana replies with a
delegation stating that any member of C can use A
to forward packets. Also, this delegation can be
further assigned to other members.
Several other communities may be available for
the forwarding of traffic of A, and the one offering
better conditions will be the one chosen. For
example, another community F may have a role
WINSYS 2007 - International Conference on Wireless Information Networks and Systems
90

stating that, for community A, 5Mb/sec are
available. However, for community C, only 2
Mb/sec are allowed. Therefore, only 2 Mb/sec can
be actually used between A and C using F to
forward. This is one example of negotiation and
access control; other examples can include the time
period of forwarding, the type of traffic, or any other
attributes or restrictions applied to this community
concept.
After all delegations are stored on Peter’s Wallet,
it forwards the packets. In this case, Joana, as the
community creator (this is a simplified community
with only one top node), could have limited to only
authorizing forwarding during a certain time period
or using a given bandwidth.
Again, notice that the names Joana, Peter, etc.,
may not be the user names, but names assigned to
specific nodes in the network or even unique random
identifiers.
6 BUSINESS OPORTUNITIES
Following the proposed communities’ management
mechanisms, new and novel business models can be
built on top of the communities’ concept.
First, the proposed mechanisms enable new
nodes/users to join the community automatically and
in a self-organized approach. This enables both the
support of access control for new nodes in the
network and for adaptable shared communication
contexts. This approach can then be used to build
micro operators, wireless and mesh based, with core
nodes fixed and already established in the
community, with possible network extensions
through new nodes. For this business model to be
possible, incentive mechanisms should be in place to
enforce cooperation of new nodes. Moreover,
through the communities’ management and access
control concepts proposed, only authorized users can
access the community (network) and the services
available, emulating the same behaviour of access to
the operator services. Moreover, access constrains
may take in consideration reputation or resources,
and be applied to any layer of the communication
stack. Finally, the inter-community management
following a cross-layer approach enables the
creation of a network operator through the support
of interaction between different layer communities.
The delegation concept described can also have a
large impact in the support of new business models.
As an example, one community can sell network
resources (e.g. bandwidth) to another community,
which contains specific delegation roles to enable
the reselling of these (or a set of) resources to
different communities. This can be applied, for
example, to federation agreements between inter-
domain operators, which define service level
agreements (SLAs) between each other (the selling
of resources for traffic traversing their inter-domain
connections), and even end-to-end agreements
which contain end-to-end resources available for a
specific set of services, through the reselling of the
resources to the several domains on the end-to-end
path. Another example is again the support of micro-
operators, whose resources can be achieved through
this delegation process.
7 CONCLUSIONS
This paper presents a community-based approach to
the definition of next generation user-centric
communications. Nodes and users are encouraged to
cooperate at all levels, sharing their resources both at
the application and communication layers. The
social relationships existing between users will
promote increased trust in local environments,
allowing wireless techniques based on individually
owned APs to become a trusted communication
environment. Mesh networking is particularly
adequate to this approach, allowing users to roam
freely, while their APs establish long-term
communication backbones.
These multi-level communities can define their
own policies and establish cooperation agreements
with other communities – both at the same and at
different levels of the communication stack.
Roles and delegations are of vital importance to
community management, with all the organization
being managed according to these concepts.
Furthermore, such methods allow the creation of
extended business models where, instead of the
service providers, users and their relations are the
driven forces for network operation.
Current work is focused on further refining the
community architecture and the integration of QoS
and mobility mechanisms. Results from prototype
implementations and real world deployments, on
metropolitan scenarios, will further contribute to the
evaluation of the solutions proposed.
ROLE BASED CROSS-LAYER COMMUNITIES ON WMN
91

ACKNOWLEDGEMENTS
This work has been partially supported by the
European Commission project IST-WIP under
contract 27402.
REFERENCES
D. Banerjee, S. Saha, S. Sen, P. Dasgupta, 2005,
"Reciprocal Resource Sharing in P2P Environments",
Proeedings. of the 4th International Conference on
Autonomous Agents and Multi-Agent Systems
(AAMAS'05), Utrecht, The Netherlands, pp. 853-869.
E. Borcoci, E. Guainella, M. Katz, P. Neves, 2007,
“WiMAX technology support for applications in
environmental monitoring, fire prevention and
telemedicine”, IEEE Mobile WiMAX 2007,
Orlando/USA, pp. 25-29.
L. Buttyan, J. Hubaux,, 2000, “Enforcing Service
Availability in Mobile Ad-Hoc WANs”, In Proc. of
the IEEE/ACM Workshop on Mobile Ad Hoc
Networking and Computing.
B. Cohen, 2003, “Incentives build robustness in
Bittorrent”, In Workshop on Economics of Peer-to-
Peer System.
B. Dressler, 2006, "Self-Organization in Ad Hoc
Networks: Overview and Classification," University of
Erlangen, Dept. of Computer Science 7, Technical
Report.
E. Freudenthal, T. Pesin, L. Port, E. Keenan, V.
Karamcheti, 2002, “dRBAC: Distributed Role-based
Access Control for Dynamic Coalition Environments”,
In Proc. of the
International Conference on Distributed Computing
Systems, Vienna.
P. Golle, K. Leyton-Brown, I. Mironov, and M.
Lillibridge. 2001, “Incentives for sharing in peer-to-
peer networks”, In Proc. of the Second International
Workshop on Electronic Commerce, pages 75—87.
Z. Haas, M. Pearlman, P. Samar, 2002, “The Zone
Routing Protocol (ZRP) for Ad Hoc Networks”,
Internet Draft, draft- ietf-manet-zone-zrp-04.txt.
G. Hardin., 1968 “The Tragedy of the commons”. Science
162, pp. 1243-1248.
Q. He, D. Wu, and P. Khosla. 2004, “SORI: a secure and
objective reputation-based incentive scheme for ad hoc
networks.” In IEEE Wireless Communications and
Neworking Conference (WCNC 2004)
D. Hughes, G. Coulson, and J. Walkerdine, Freeriding,
2005, “Gnutella Revisited: the Bell Tolls?”, IEEE
Distributed Systems Online, Vol. 6, No. 6.
N. Li, W. Winsborough, J. Mitchell. 2001, “Distributed
credential chain discovery in trust management”, In
Proc. of ACM Conf. on Computer and
Communications Security.
K. Ramachandran, E. Belding, K. Almeroth, M.
Buddhikot, 2006, “Interference-Aware Channel
Assignment in Multi-Radio Wireless Mesh
Networks”, IEEE Infocom, Barcelona, Spain.
R. Sandhu, E. Coyne, H. Feinstein, C. Youman, 1996,
“Role-based access control models”, IEEE Computer,
20(2):38–47.
B. Yu, M.P. Singh, K. Sycara, 2004 “Developing trust in
large-scale peer-to-peer systems”, In Proc. of First
IEEE Symposium on Multi-Agent Security and
Survivability., pp. 1-10.
EMule Project: www.emule-project.net
IMDB - The Internet Movie Database (IMDb):
www.imdb.com
IST-WIP - An All-Wireless Mobile Network Architecture:
www.ist-wip.org
SETI@Home: setiathome.berkeley.edu
Wikipedia – The free encyclopaedia: www.wikipedia.com
WINSYS 2007 - International Conference on Wireless Information Networks and Systems
92