ArticlePDF Available

Abstract and Figures

The wireless local area network standard IEEE 802.11 is the preferred solution for lowcost data services. Key to its success are the 2.4 and 5 GHz unlicensed bands. The transmit power limitations imposed due to regulatory requirements limit the range (coverage) that can be achieved by WLANs in these bands. However, the demand for "larger" wireless infrastructure is emerging, ranging from office/university campuses to city-wide deployments. To overcome the limitations of singlehop communication, data packets need to traverse over multiple wireless hops, and wireless mesh networks are called for. Since 2004 Task Group S has been developing an amendment to the 802.11 standard to exactly address the aforementioned need for multihop communication. Besides introducing wireless frame forwarding and routing capabilities at the MAC layer, the 802.11s amendment brings new interworking and security. In this article, we provide insights into the latest developments in 802.11s and explain how the overall mesh concept fits into the 802 set of networking standards.
Content may be subject to copyright.
Wireless LANs (WLANs) are proliferating and
the desire for ubiquitous wireless connectivity is
driving the demand for coverage extension of
today’s WLANs. However, regulatory limitations
restrict the transmission power of WLAN
We have been here before. Bridging evolved
the Ethernet (802.3) standard from a single-hop
to a multihop system. With bridges, the commu-
nication between end stations is no longer limit-
ed to the same LAN. WLANs are in an early
stage compared to their wired ancestors. The
present 802.11 interconnections rely on wired
networks to carry out bridging functions. For a
number of reasons, this dependency on wired
infrastructure must be eliminated. First, this
dependency is costly and inflexible, as WLAN
coverage cannot be extended beyond the back-
haul deployment. Second, centralized structures
work inefficiently with new applications, such as
wireless gaming, requiring peer-to-peer connec-
tivity. Third, a fixed topology inhibits stations
from choosing a better path for communication.
WLANs can benefit significantly if they evolve
to address these emerging needs.
Wireless mesh networks (WMNs) hold the
promise of a solution. However, existing WMNs
rely on the IP layer to enable multihop communi-
cation and do not provide an inherently wireless
solution. Since wireless links are less reliable than
wired links, a multihop routing protocol operating
in a wireless environment must account for the
nature of the wireless links. As 802.11 does not
specify the interfaces that the IP layer needs to
derive link metrics from the medium access con-
trol (MAC) layer, the ad hoc routing protocols
developed in the Internet Engineering Task
Force’s (IETF’s) Mobile Ad Hoc Networks
(MANET) group are forced to rely on indirect
measurements [1] to observe the radio environ-
ment. However, the acquired link metrics are of
limited accuracy [2], whereas the MAC layer has
adequate knowledge of its radio neighborhood to
make its measurements less outdated and more
precise. Furthermore, for transparent support of
important protocols like Address Resolution Pro-
tocol, Dynamic Host Configuration Protocol,
Spanning Tree, and many more, a WMN must
appear like traditional LAN segments that form
single broadcast domains. In encapsulating layer 2
traffic, IP-based WMNs emulate LAN behavior.
However, this appears to be more like an engi-
neering patch and does not provide a long-term
solution aimed at sustainable scaling of WLANs to
new applications. Since MAC-based multihop
solutions inherently support layer 2 traffic, they
operate transparently to any higher-layer protocol.
To realize the benefits a MAC-based WMN
promises, an integrated mesh networking solu-
tion is under development in IEEE 802.11 Task
Group S. The particular amendment of the
802.11 standard dealing with mesh support,
802.11s [3], describes a WMN concept that intro-
IEEE Wireless Communications • February 2010
104 1536-1284/10/$25.00 © 2010 IEEE
Portal A
Mesh STA A
h other
Mesh STA W
The wireless local area network standard
IEEE 802.11 is the preferred solution for low-
cost data services. Key to its success are the 2.4
and 5 GHz unlicensed bands. The transmit
power limitations imposed due to regulatory
requirements limit the range (coverage) that can
be achieved by WLANs in these bands. Howev-
er, the demand for “larger” wireless infra-
structure is emerging, ranging from
office/university campuses to city-wide deploy-
ments. To overcome the limitations of single-
hop communication, data packets need to
traverse over multiple wireless hops, and wire-
less mesh networks are called for. Since 2004
Task Group S has been developing an amend-
ment to the 802.11 standard to exactly address
the aforementioned need for multihop commu-
nication. Besides introducing wireless frame for-
warding and routing capabilities at the MAC
layer, the 802.11s amendment brings new inter-
working and security. In this article, we provide
insights into the latest developments in 802.11s
and explain how the overall mesh concept fits
into the 802 set of networking standards.
The authors provide
insights into the
latest developments
in 802.11s and
explain how the
overall mesh concept
fits into the 802
set of networking
HIERTZ LAYOUT 2/9/10 1:45 PM Page 104
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
IEEE Wireless Communications • February 2010 105
duces routing capabilities at the MAC layer.
Path selection is used to refer to MAC-address-
based routing and to differentiate it from con-
ventional IP routing.
To understand the general WLAN concept, we
begin by explaining the 802.11 standard [4] briefly.
Next, we provide an outline of the basic building
blocks of 802.11s, extending what is explained in
[5], wherein we discuss interworking, MAC, secu-
rity, and path selection. Subsequently, we discuss
certification activities for mesh solutions in the
Wi-Fi alliance [6]. The last section is dedicated to
implementations such as One Laptop Per Child
(OLPC) [7] and the open80211s [8] project, which
reveal the performance of the 802.11s draft and
provide us with a preview of what can be expected
with 802.11s-certified products.
A station, or STA, is an 802.11-standard-compli-
ant MAC and physical layer (PHY) implementa-
tion [4] and constitutes the basic entity in an
802.11 network. The most elementary 802.11
network, called a basic service set (BSS), can be
formed using two stations. If a station provides
the Integration service to the other stations, this
station is referred to as an access point (AP). If
an AP is present in a BSS, it is referred to as an
infrastructure BSS. To join an infrastructure
BSS, a station associates with the AP. Figure 1
provides an example where AP M is part of the
infrastructure. AP M provides stations B and C
with access to the distribution system (DS). The
DS provides the services that are necessary to
communicate with devices outside the station’s
own BSS. Furthermore, the DS allows APs to
unite multiple BSSs to form an extended service
set (ESS). Within an ESS, stations can roam
from one BSS to another [9]. Today Ethernet
(802.3) usually provides the distribution system
medium (DSM) on which the DS relies. Conse-
quently, in practice, APs collocate with the so-
called portals that provide the integration of
WLANs with non-802.11 networks.
The IEEE 802.11 standard itself does not
provide any details about the DSM. In principle,
the DSM can be wireless too. The 802.11 frame
format (without the extensions highlighted in
Fig. 2) provides four fields necessary for address-
ing over multiple intermediate devices. The
source address indicates the station that generat-
ed the frame (initial hop), and the destination
address indicates the intended receiver (final
hop). Both addresses remain unchanged in a
concatenated set of multiple wireless hops. The
Figure 1. 802.11s enables seamless connectivity among dissimilar 802 networks.
The 802.11s mesh appears as a single logical broadcast domain. Support for spanning tree guarantees loop-free
connectivity with external networks Portals B and C blocked.
802.11s enables mobile mesh
STA Y to seamlessly establish
a new mesh link to mesh STA C
and release mesh links to
mesh STAs A and H.
Via portal D, 802.3 station J
integrates transparently
with the
The 802.11
concept relies on a
central AP that forms a
basic service set (BSS).
Interconnected by 802.11s,
stations can transition to and
from APs K, L, and M within
BSSs K, L, and M, respectively.
Portal A
Mesh STA A
Portal B
Mesh STA B
Portal C
Mesh STA C
Portal D
Mesh STA D
Mesh STA K
Mesh STA J
Mesh BSS
802.11s mesh link (forwarding, may be part of a mesh path, multihop)
802.11s mesh link (non-forwarding, single-hop)
802.11 link within basic service set (BSS)
Link released after transitioning to new location
802.11s mesh
integrates with other
802 networks
(802.3, 802.16, etc.)
Due to its mesh capabilities, mesh STA U connects simultaneously to the printer (mesh STA W) and the storage device
(mesh STA V), and maintains Internet connectivity via mesh STA J. However, as a non-forwarding mesh device, mesh
STA U does not participate in mesh formation. Thus, it does not interconnect mesh STAs W, V, and J.
Mesh STA Y
Mesh STA Y
Mesh STA W
Mesh STA U Mesh STA V
Mesh STA H
Mesh STA G
Mesh STA E Portal E
Mesh STA F
HIERTZ LAYOUT 2/9/10 1:45 PM Page 105
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
transmitting and receiving station addresses,
which denote the stations that actually forward-
ed the frame, change with every hop. The 802.11
frame format provides two additional bits denot-
ed “To DS” and “From DS.” The bit combina-
tions 10 and 01 indicate the traffic entering or
leaving the DS from a BSS, respectively. For the
traffic that is relayed within the DS from one AP
to another, the bit combination 11 is used.1
Interestingly, vendors often use the vernacular
term wireless DS (WDS) for this configuration.
Since the current standard [4] explicitly states that
it does not define the procedures necessary for
WDS implementation, many 802.11 multihop
implementations do not interoperate. 802.11s not
only helps interconnect BSSs wirelessly and there-
by fills the WDS gap; it also enables a new type of
BSS, the so-called mesh BSS (MBSS). In the fol-
lowing we describe the major sections of 802.11s.
Its amendments to interworking, MAC, security,
and path selection make the MBSS a self-con-
tained network that enables applications beyond
what a traditional single-hop WLAN supports.
As a family of standards, interoperability
between the different networking concepts is a
requirement for 802. For seamless integration,
the 802.11s network appears as a single Ethernet
segment to the outside (Fig. 1). The WMN
implements a single broadcast domain and thus
integrates seamlessly with other 802 networks. In
particular, 802.11s supports transparent delivery
of uni-, multi-, and broadcast frames to destina-
tions in- and outside of the MBSS (referred to
as mesh in the following). Devices that form the
mesh are called mesh stations (mesh STAs).
Mesh stations forward frames wirelessly but do
not communicate with non-mesh stations. How-
ever, a mesh station may be collocated with
other 802.11 entities.
Currently, 802.11 categorizes frames as data,
control, or management. Data frames carry high-
er-layer data. Control frames are used for
acknowledgments and reservations. Devices use
management frames to set up, organize, and
maintain a WLAN and the local link. To provide
for multihop, 802.11s extends data and manage-
ment frames by an additional mesh control field,
as shown in Fig. 2. The mesh control field con-
sists of a mesh time to live (TTL) field, a mesh
sequence number, a mesh flags field, and possi-
bly a mesh address extension field. The TTL and
sequence number fields are used to prevent the
frames from looping forever. When mesh sta-
tions communicate over a single hop, their
frames do not carry the mesh control field.
The mesh flags field indicates the presence of
additional MAC addresses in the mesh control
field. The address extension allows for a total of
six address fields in a mesh frame. This is useful
when the source and destination of the frame are
not part of the mesh, but are proxied by mesh sta-
tions. Figure 1 presents an example where mesh
station D proxies non-mesh stations A, B, and J.
Informing other mesh stations of its proxied
devices, mesh station D diverts to itself all frames
destined for A, B, or J. Together with the six-
address scheme, the proxied entities can be iden-
tified as the final destination beyond the
intermediate destination D. In addition, the
extension to six addresses allows for proactive
routing, explained later. Proactive routing divides
a path into two distinct routes to simplify path
selection. In Fig. 3 only mesh station C maintains
paths to all mesh stations. In this case non-mesh
station D’s frames enter the mesh at mesh station
K, traverse to mesh station C (the first route),
and from there to mesh station J (the second
route). An observant reader will note from Fig. 2
that the address extension field allows for the
addition of three addresses, rather than just two.
The rationale for this is that standard manage-
ment frames have three addresses only. Hence, in
the case of multihop mesh management frames,
address 4 is included in the mesh control field
rather than in the standard frame header.
Just as an AP’s beacon frame helps the stations
to detect a BSS and learn about its settings, the
mesh station’s beacon carries information about
the mesh and helps other mesh stations detect
and join the mesh. Mesh stations detect each
IEEE Wireless Communications • February 2010
Figure 2. The 802.11s mesh control field is part of the frame body and provides up to two more address fields.
2 octets
2 octets
Address 1
6 octets
Address 2
6 octets
Address 3
6 octets
2 octets
Address 4
6 octets
QoS control
2 octets
HT control
4 octets
4 octets
Mesh flags
Mesh time
to live
Mesh address extension
1 octet 1 octet 4 octets
2 bits 6 bits
0, 6, 12, or 18 octets
6, 12, 18, or 24
0–7955 octets
1The bit combination 00
is to be used within an
independent BSS that
does not have an AP.
HIERTZ LAYOUT 2/9/10 1:45 PM Page 106
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
IEEE Wireless Communications • February 2010 107
other based on passive scanning (observation of
beacon frames) or active scanning (probe frame
transmission). The mesh-specific beacon and
probe frames contain a Mesh ID (the name of a
mesh), a configuration element that advertises
the mesh services, and parameters supported by
the transmitting mesh station. This functionality
enables mesh stations to search for suitable peers
(e.g., other mesh stations that use the same path
selection protocol and metric). Once such a can-
didate peer has been identified, a mesh station
uses the Mesh Peer Link Management protocol
to establish a peer link with another mesh sta-
tion. Even when the physical link breaks, mesh
stations may keep the peer link status to allow
for quick reconnection. In Fig. 1 mesh station Y
may re-establish connection with mesh station A
or H as soon as it moves in range again.
Mesh stations use a single transceiver only.
Accordingly, a mesh operates in a single fre-
quency channel only. With multitransceiver
devices, however, different frequency channel
meshes can be unified into a single LAN. Figure
4 provides an example where five meshes oper-
ate in four different frequency channels. Mesh
stations C, D, and E collocate within a device
that has three independent transceivers. Incor-
porating an 802 bridge in the device, the collo-
cated mesh stations interconnect and help to
forward frames between their meshes. Conse-
quently, a single WMN can be constituted.
Regulatory bodies have different require-
ments on the frequency bands used by 802.11
and mesh stations are required to comply with
these regulatory requirements. In Europe, for
example, devices must switch to a different chan-
nel (dynamic frequency selection) upon detec-
tion of a radar station in the 5 GHz band. To
prevent a mesh from splitting, a channel selec-
tion protocol allows selection of the new fre-
quency channel. In the absence of a central
coordinator, a distributed algorithm is devel-
oped, based on a 31-bit random channel prece-
dence value, for arbitration.
If mesh station O in Fig. 4 detects a radar
station, it is required to leave its current fre-
quency channel and indicates the new frequency
channel to mesh station N. N forwards the mes-
sage, and thus, mesh stations D and T learn
about the channel switch too. After a predeter-
mined period of time, the mesh channel switch
time, the mesh stations switch to the new chan-
nel. If a mesh station holds a larger channel
precedence value, it broadcasts its value and may
indicate a different frequency channel. Following
the highest announced precedence value, mesh
stations finally coalesce on the new channel.
All beacon frames provide a time reference that
is used for synchronization and power saving.
Power-saving mesh stations are either in light- or
deep-sleep mode. Being in light-sleep mode, a
mesh station switches to full power whenever a
neighbor or the mesh station itself is expected to
transmit a beacon frame. In deep-sleep mode a
Figure 3. The six address scheme provides support for proxied stations and tree-based path selection.
Mesh STA J is the
intended recipient of
STA D’s frame.
STA D associates with
AP K that is collocated
with mesh STA K.
STA D sends a frame
to mesh STA J.
Mesh STA C operates as root mesh STA
as it provides connection to the Internet.
Portal C
Mesh STA C
Mesh STA K
Mesh STA D
Mesh BSS
Mesh STA E
Mesh STA B
Mesh STA A
Mesh STA H
Mesh STA G Mesh STA F
Mesh STA J
If a mesh station
holds a larger
channel precedence
value, it broadcasts
its value and may
indicate a different
frequency channel.
Following the
highest announced
precedence value,
mesh stations finally
coalesce on the
new channel.
HIERTZ LAYOUT 2/9/10 1:45 PM Page 107
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
IEEE Wireless Communications • February 2010
mesh station solely wakes up for its own beacon
frame transmissions. The mesh station can be
informed of buffered traffic during the awake
period that follows the beacon. Synchronization
enables a new kind of distributed reservation
protocol, introduced in the next section.
For medium access, mesh stations implement
the mesh coordination function (MCF). MCF
consists of a mandatory and an optional scheme.
For the mandatory part, MCF relies on the con-
tention-based protocol known as Enhanced Dis-
tributed Channel Access (EDCA), which itself is
an improved variant of the basic 802.11 dis-
tributed coordination function (DCF). Using
DCF, a station transmits a single frame of arbi-
trary length. With EDCA, a station may transmit
multiple frames whose total transmission dura-
tion may not exceed the so-called transmission
opportunity (TXOP) limit. The intended receiv-
er acknowledges any successful frame reception.
Additionally, EDCA differentiates four traffic
categories with different priorities in medium
access and thereby allows for limited support of
quality of service (QoS).
To enhance QoS, MCF describes an optional
medium access protocol called Mesh Coordinat-
ed Channel Access (MCCA). It is a distributed
reservation protocol that allows mesh stations to
avoid frame collisions. With MCCA, mesh sta-
tions reserve TXOPs in the future called MCCA
opportunities (MCCAOPs). An MCCAOP has a
precise start time and duration measured in slots
of 32 μs. To negotiate an MCCAOP, a mesh sta-
tion sends an MCCA setup request message to
the intended receiver. Once established, the
mesh stations advertise the MCCAOP via the
beacon frames. Since mesh stations outside the
beacon reception range could conflict with the
existing MCCAOPs, mesh stations also include
their neighbors’ MCCAOP reservations in the
beacon frame. At the beginning of an MCCA
reservation, mesh stations other than the
MCCAOP owner refrain from channel access.
The owner of the MCCAOP uses standard
EDCA to access the medium, and does not have
priority over stations that do not support MCCA.
Although this compromises efficiency, simula-
tions reveal that high medium utilization can still
be achieved with MCCA in the presence of non-
MCCA devices [10]. After an MCCA transmis-
sion ends, mesh stations use EDCA for medium
contention again.
Access in 802.11 relies on carrier sensing. At a
mesh’s edge, mesh stations have fewer neigh-
bors. and therefore observe an idle wireless
medium more often than mesh stations located
in the core of the mesh. Consequently, edge
mesh stations have a higher probability to trans-
mit. When core mesh stations congest, they can-
not carry the aggregated traffic and drop frames.
This is costly as the mesh frame has already tra-
versed several hops to reach the congested mesh
station. The optional 802.11s congestion control
concept uses a management frame to indicate
the expected duration of congestion and to
request a neighbor mesh station to slow down.
Since it is each mesh station’s choice to issue a
congestion control frame, the notification may
finally ripple back to the traffic source.
With 802.11s, mesh stations perform the dictio-
nary attack-proof Simultaneous Authentication
of Equals (SAE) [11] algorithm. Besides mutual
authentication, SAE provides two mesh stations
with a pairwise master key (PMK) that they use
to encrypt their frame. As its name indicates,
SAE does not rely on a keying hierarchy like tra-
ditional 802.11 encryption [4]. Instead, it imple-
ments a distributed approach that both mesh
stations may initiate simultaneously. Because of
the pairwise encryption, each link is indepen-
dently secured. As a consequence, 802.11s does
not provide end-to-end encryption. Since broad-
cast traffic must reach all authenticated peers, a
mesh station is required to update its broadcast
traffic key with every new peering it establishes.
Within a mesh, all mesh stations use the same
path metric and path selection protocol. For
both, 802.11s defines a mandatory default
scheme. Because of its extensible framework,
they can be replaced by other solutions.
The default metric, termed airtime metric,
indicates a link’s overall cost by taking into
account data rate, overhead, and frame error rate
of a test frame of size 1 kbyte. The default path
selection protocol, Hybrid Wireless Mesh Proto-
col (HWMP), combines the concurrent operation
of a proactive tree-oriented approach with an on-
demand distributed path selection protocol
(derived from the Ad Hoc On Demand Distance
Vector [AODV] protocol [1]). The proactive
mode requires a mesh station to be configured as
a root mesh station. In many scenarios this will
be a mesh station that collocates with a portal
(Fig. 3). As such, the root mesh station constant-
ly propagates routing messages that either estab-
lish and maintain paths to all mesh stations in
Figure 4. Layer 2 bridges in multi-transceiver devices may unify different fre-
quency channel meshes into a single LAN.
HIERTZ LAYOUT 2/9/10 1:45 PM Page 108
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
IEEE Wireless Communications • February 2010 109
the mesh, or simply enable mesh stations to initi-
ate a path to it (red lines in Fig. 3). In the exam-
ple of Fig. 3, mesh station K uses the root mesh
station C to establish an initial path (dotted line)
to mesh station J. Once established, mesh sta-
tions may use the AODV part of HWMP to
avoid the indirection via the root mesh station. In
the present example, K could discover a shorter
path (links marked in grey) via G and H to for-
ward station D’s frames to the destination mesh
station J. Mesh stations also rely on AODV when
a root mesh station is unavailable. When no path
setup messages are propagated proactively, how-
ever, the initial path setup is delayed.
To allow for even simpler configuration, ven-
dors may opt not to implement HWMP at all.
As an example, a battery-limited handheld device
may refrain from frame forwarding to minimize
power consumption. Accordingly, it does not
propagate path information and behaves like an
end station. However, the device is still able to
request the frame forwarding service from neigh-
boring mesh stations.
In October 2006 the Wi-Fi Alliance (WFA)
established the Mesh Marketing Task Group,
chartered to work on a Marketing Requirement
Document, and the specification of a certification
and test plan. To meet customers’ expectations,
WFA’s mesh activities aim at compliance with
the present certification programs. Accordingly,
simple and secure setup of a WFA-certified mesh
needs to comply with the existing WFA pro-
grams. The current Wi-Fi protected setup already
enables security via a push button, secret PIN, or
near-field communication, for example.
To provide compatibility with existing Wi-Fi
devices, WFA’s marketing program requires each
mesh station to incorporate either the AP or sta-
tion functionality too. While Wi-Fi mesh APs
must support frame-forwarding and thereby help
to increase the radio coverage, non-AP mesh sta-
tions may choose to become an end station.
Whereas a non-mesh station connects to a single
AP only, a mesh station may connect with multi-
ple other mesh stations even if it does not for-
ward traffic for others. Consequently, it provides
users the advantage of access to services not
reachable via the AP (mesh station U in Fig. 1).
The OLPC project and open80211s [7] are the
world’s first implementations of 802.11s. In the
next two sections we briefly introduce the pro-
ject goals and experiences gained from real-
world setups.
Developed by the OLPC Foundation, the XO
laptop aims to serve as a learning tool for chil-
dren living in developing countries where a com-
munication infrastructure is unlikely to exist.
With WLAN embedded in the XO, the decision
to implement 802.11s was self-evident. Based on
an early draft of 802.11s, the XO omits certain
functions of 802.11s such as encryption or proac-
tive routing. One of the challenges faced by
OLPC was to ensure at all times a minimal node
density, which is critical for the proper operation
of a mesh network. Two design choices were
made to address this issue [7].
First, the OLPC mesh does not implement
any access control mechanism. Each node can
receive and forward traffic from any other mesh-
capable node, thus avoiding a possible fragmen-
tation of the network caused by incompatible
access credentials. As there is no authentication
at the mesh layer, XOs must rely on upper layers
for confidentiality.
Second, the mesh protocol stack is embedded
in the wireless network card. With this architec-
ture, the entire 802.11s code can operate inde-
pendently of the host CPU. Thus, the XO works
as a mesh station even when in power-save
mode; that is, a laptop transitioning into power-
save mode will not adversely affect other stu-
dents who may rely on a single student’s
provision of Internet connectivity.
Due to its distributed nature, OLPC assumes
that a root mesh station is never available. Thus,
the XO would not benefit from implementing a
tree structure. Consequently, the XO solely
implements HWMP’s AODV part.
Several presentations at road shows and live
demonstrations have shown the capabilities of
the present OLPC mesh implementation. To
date, OLPC has shipped over 1.2 million mesh-
capable laptops around the globe.
open80211s [7] is a vendor-neutral implementa-
tion of 802.11s for the Linux operating system
(Fig. 5). Since 802.11s introduces only minimal
changes to the MAC layer, the 802.11s stack can
be almost fully implemented in software and
made to run on legacy 802.11 cards. The goal of
the project is to closely track the 802.11s draft
and support the interoperability of different
802.11s implementations. The availability of the
source code helps to identify and resolve design
problems, and resolve ambiguities in the protocol
being specified. Performance measurements are
routinely taken before each release. Figure 6a
shows the path discovery time for different path
lengths, where we measure a 12-node open80211s
testbed wherein all the nodes are in radio range
Figure 5. The open80211s stack integrates into the Linux kernel.
802.1 bridging
Hardware driver
802.11 station 802.11 AP 802.11s mesh STA
HIERTZ LAYOUT 2/9/10 1:45 PM Page 109
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
IEEE Wireless Communications • February 2010
of each other, and manual address filter settings
enforce multihop topologies. We measure the
discovery duration from the time a path setup is
requested until the time the path is actually
established. As expected, routes are resolved in
linear time, and the variance increases with the
length of the path. Figure 6b shows the average
end-to-end data rate measurement results in an
environment where the default 802.11 parame-
ters are used for measurements, and where each
experiment is run six times for 10 s per measure-
ment in an uncontrolled wireless environment.
As previously mentioned, non-802.11s prod-
ucts rely on the unspecified WDS to enable mul-
tihop networking. However, without path
selection mechanisms, suboptimal path length or
undesired increase in path length occurs. Figure
6 illustrates scenarios with more than four to five
hops that are usually beyond the application sce-
nario of a single-channel WMN. Note that
throughput decreases rapidly with the number of
hops in the fully connected topology of the
experiment. Under normal conditions, the rout-
ing protocol used in 802.11s would resolve paths
that are limited to one or two hops only. The
open80211s stack has been part of the Linux
kernel since version 2.6.26 (July 2008).
The wired Internet liberated users from dial-up
connections and leased lines. By interconnecting
autonomous systems, the decentralized Internet
enabled new services and business. With 802.11s,
a similar development has started. The classical
single-hop connectivity no longer addresses the
needs of an ever-growing user base, and mesh
technology is the natural evolution.
However, with mesh networking, 802.11 enters
uncharted territories, and 802.11s does not yet
define the definite solution. Further improvements
are necessary to increase efficiency and enable a
degree of QoS that users are willing to accept.
Judging from the current status of the ongoing
standardization process, it seems that the finaliza-
tion of the 802.11s standard may be expected next
year. Remaining problem areas are congestion
control, channel selection, and medium access.
The congestion control mechanism requires a
mesh station to specify the congestion duration
to its neighbors. However, due to varying radio
conditions, the link speed changes; a congested
mesh station can hardly predict this. Even worse,
legacy devices in overlapping BSSs cannot com-
ply with congestion messages. As a result, neigh-
boring BSSs receive an increased share of the
wireless medium, and the congested mesh sta-
tion is not really helped. Finally, conditions that
trigger congestion control, as well as neighbor
mesh station reaction, are left unspecified.
Another concern addresses the current 802.11s
method for distributed frequency channel selection.
Based on random values that are needed for arbi-
tration, the propagation of a common frequency
channel has limited reliability. Unfortunately, the
selection scheme cannot guarantee that all mesh
stations are informed. Since a mesh station switches
at the latest after 255 ms, large networks are subject
to partitioning when different mesh stations
increase the precedence value and the message
cannot propagate back to the originator in time.
Finally, there remains the challenge of medi-
um access. Currently, WLAN deployments rely
on a wired backbone where APs need not be in
radio range and hence do not share a frequency
channel. However, to form a WMN, this changes.
With the wireless medium being shared among
neighbors, the environment becomes much more
interference-prone. Furthermore, mesh stations
have no priority over other 802.11 devices, and a
mesh suffers severely from any overlapping BSS.
Even if a mesh AP incorporates multiple
transceivers to separate the BSS and mesh net-
work into different frequency bands (2.4 and 5
GHz), the WMN carries the aggregation of
locally generated and forwarded traffic. Accord-
ingly, the WMN is threatened with saturation.
Experiments with real 802.11s deployments sub-
stantiate these limitations of the EDCA-based
medium access mechanisms when used in a
WMN environment. Schemes tailored to mesh-
specific needs, such as the MCCA scheme, possi-
Figure 6. open80211s performance measurements.
Number of hops
Discovery time (ms)
2 3 4 5 6 7 8 9 10 11
Number of hops
Throughput (Mb/s)
2 3 4 5 6 7 8 9 10 11
Regression Measured
HIERTZ LAYOUT 2/9/10 1:45 PM Page 110
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
IEEE Wireless Communications • February 2010 111
bly combined with path selection and flow con-
trol, are quite likely to bring benefits.
Aided by the experiences gained in the imple-
mentation projects, IEEE 802.11s is in the pro-
cess of tackling these challenges. Once
overcome, mesh technology will be an inherent
part of any future wireless standard. Today a
novelty, users are likely to take the ability to
communicate without a wired infrastructure for
granted in the future: convenient deployment
and spontaneous connectivity — anytime, any-
where. To provide for this, 802.11s will not
remain the only mesh solution. With 802.11s, the
fascinating WMN adventure has just begun.
[1] C. Perkins, E. Belding-Royer, and S. Das, “Ad Hoc On-
Demand Distance Vector (AODV) Routing,” IETF RFC
3561, Jul. 2003.
[2] I. F. Akyildiz, X. Wang, and W. Wang, “Wireless Mesh
Networks: A Survey,” Comp. Networks and ISDN Sys.,
vol. 47, no. 4, Mar. 2005.
[3] “Draft Standard for Information Technology — Telecom-
munications and Information Exchange Between Sys-
tems — LAN/MAN Specific Requirements — Part 11:
Wireless Medium Access Control (MAC) and Physical
Layer (PHY) specifications: Amendment 10: Mesh Net-
working,” IEEE unapproved draft, IEEE P802.11s/D4.0,
Dec. 2009.
[4] IEEE P802.11-2007, “Information Technology —
Telecommunications and Information Exchange
Between Systems — Local and Metropolitan Area Net-
works — Specific Requirements — Part 11: Wireless
Medium Access Control (MAC) and Physical Layer (PHY)
Specifications,” JunE 2007.
[5] S. M. Faccin et al., “Mesh WLAN Networks: Concept
and System Design,” IEEE Wireless Commun., vol. 13,
no. 2, Apr. 2006.
[6] Wi-Fi Alliance, available:
[7] J. Cardona, “Wireless Meshing with the One Laptop Per
[8] open80211s,
[9] B. Walke, S. Mangold, and L. Berlemann, IEEE 802 Wire-
less Systems: Protocols, Multi-Hop Mesh/Relaying, Per-
formance and Spectrum Coexistence, Wiley, Nov. 2006.
[10] Y. Chen, and S. Emeott, “MDA Simulation Study:
Robustness to Non-MDA Interferers,” IEEE 802 Plenary
Meeting, Submission 11-07-0356-00-000s, Orlando, FL,
Mar. 2007.
[11] D. Harkins, “Simultaneous Authentication of Equals: A
Secure, Password-Based Key Exchange for Mesh Net-
works,” SENSORCOMM ’08 — 2nd InGut’l. Conf. Sen-
sor Technologies and Apps., Aug. 2008.
GUIDO R. HIERTZ ( received his Dipl.-Ing.
degree in electrical engineering from RWTH Aachen Univer-
sity, Germany. Working toward his Ph.D. in the Depart-
ment of Communication Networks, he contributed to
various research projects and authored several papers at
IEEE conferences. He is a voting member of IEEE 802.11
and Vice Chair of IEEE 802.11s. He is a charger member of
the industry forum Wi-Mesh Alliance that created the initial
draft of IEEE 802.11s jointly with the industry group SEE-
Mesh. Since 2009 he is the head of research and develop-
ment in the Rental Series Department at Riedel
Communications, Wuppertal, Germany.
DEE DENTENEER received an M.Sc. in statistics from the Uni-
versity of Utrecht and a Ph.D. in applied probability (queu-
ing analysis) from Eindhoven University ot Technology, the
Netherlands. From 1984 to 1988 he worked at the Dutch
Central Statistical Office, where he designed the Blaise lan-
guage, a language for questionnaire description. Since
1988 he has been employed at Philips Research in Eind-
hoven, since 2000 as a principal research scientist. At
Philips he has worked on the application of mathematics in
various industrial research projects such as MPEG encod-
ing, speech recognition, secure biometrics, and data trans-
mission systems. He is Chair of IEEE 802.11s. His current
research interest is in the performance analysis and stan-
dardization of wireless mesh networks.
SEBASTIAN MAX studied computer science at RWTH Aachen
University, and received his diploma degree with distinction
in 2005. Since then he has been with the Chair of Commu-
nication Networks (ComNets) at RWTH Aachen University,
where he is working toward his Ph.D. He holds the
Research College (Graduiertenkolleg) Software for Mobile
Communication Systems scholarship of the German
Research Foundation (DFG). His main research field is wire-
less mesh networks for city-wide Internet access.
RAKESH TAORI is a principal engineer in the Digital Media
and Communications R&D Centre at Samsung Electronics,
Suwon, South Korea. He is currently involved in research,
development, and standardization of the 4G air interface
technologies pertaining to the MAC layer. He was an
active contributor to the standardization of multihop relay
in the IEEE 802.16 systems and is now contributing to the
development of the IEEE 802.16m amendment: the
advanced OFDMA air interface for the IEEE 802.16 system.
Prior to joining the Samsung DMC R&D Centre, he held
research positions at Samsung Research (2004–2008,
South Korea), Ericsson Research (2000–2004, Sweden and
the Netherlands), and Philips Research Laboratories
(1992–2000, the Netherlands). Over the past 17 years he
has performed research and standardization work in the
area of media coding and wireless systems. In the area of
media coding his primary focus was low-bit-rate paramet-
ric coding of speech and audio signals. In the area of
wireless systems he has contributed to research and stan-
dardization in wireless PANs (Bluetooth and UWB) and
wireless LANs (802.11s, WLAN mesh), and is currently
active in the area of wireless MANs (802.16m, advanced
air interface). He has contributed to several standardiza-
tion organizations — MPEG, ITU-T, ETSI, Bluetooth SIG,
IEEE, and the WiMAX Forum — and has served these
organizations in various roles. From August 2004 to
November 2005 he served as Chair of the Technical Steer-
ing Committee of the WiMedia Alliance. He obtained his
B.Eng. degree in control and computer engineering and
M.Phil degree in digital signal processing and communica-
tions from the University of Westminster, London, United
JAVIER CARDONA is the co-founder and CEO of cozybit Inc.,
an engineering consulting firm in the field of wireless com-
munications. He was one of the implementers of the OLPC
mesh stack and open802.11s. He holds an M.S. in telecom-
munications from Universitat Politecnica de Catalunya,
Spain, and an M.Eng. in embedded systems design from
LARS BERLEMANN ( is program
manager at T-Mobile International, leading the market
introduction program for NGMN. He holds a Ph.D. and
Diploma degree in electrical engineering from RWTH
Aachen University as well as a Diploma degree in business
and economics from the same university. He is author of
the textbooks Cognitive Radio for Dynamic Spectrum
Access (Wiley, 2009) and IEEE 802 Wireless Systems: Proto-
cols, Multi-Hop Mesh Relaying, Performance and Spectrum
Coexistence (Wiley, 2006). He has published a multitude of
reviewed publications including several journal articles, and
was scientific organizer of European Wireless Conference
2005 and IEEE PIMRC 2005. He has been a guest lecturer
on mobile radio networks at the Technical University of
Dortmund since 2007.
BERNHARD WALKE is directing the ComNets Research Group
at RWTH Aachen University, Germany, focusing on 4G air
interface design and performance evaluation, besides
developing system-level simulation tools like openWNS. He
contributed, together with his Ph.D. students, revolutionary
concepts that are being used in standardized mobile radio
networks, such as the packet data traffic channel of GPRS
operating on a GSM traffice channel; the fast radio link
establishment for GPRS (later named TBF), a concept also
used in UMTS; the MAC frame applied in WiMAX and 3GPP
LTE systems for radio resource allocation; and the concept
of fixed decode-and-forward relays used in broadband cel-
lular radio networks like WiMAX and 3GPP LTE/LTE-A.
Besides that, his group has contributed a number of
improvements now implemented in IEEE 802 standards. His
work is published in six textbooks and about 260 peer
reviewed conference and journal papers. Prior to joining
academia, he worked for 18 years in industry at EADS AG.
He holds a doctorate in information engineering from the
University of Stuttgart, Germany.
Aided by the
experiences gained
in the implementa-
tion projects,
IEEE 802.11s is in
the process of
tackling these
challenges. Once
overcome, mesh
technology will be
an inherent part
of any future
wireless standard.
HIERTZ LAYOUT 2/9/10 1:45 PM Page 111
Authorized licensed use limited to: Javier Cardona. Downloaded on February 26,2010 at 21:10:57 EST from IEEE Xplore. Restrictions apply.
... As the research interest into generalised MANETs diminished in favour of applicationspecific ones, some of the applications previously considered for MANETs solidified while others evaporated. The big push to have MANETs integrate with existing infrastructure became successful with the emergence of Mesh networking [81], which has seen both commercial success [154][152] [166] and acceptance within industry standards [109]. While MANETs are not yet widely deployed for natural disaster recovery as it was hoped, the technology has been adopted to create communication systems that work in the cases of other infrastructure shutdowns. ...
Mobile ad hoc networks (MANETs) are the core technology that provides the US military with adaptable and reliable battlefield communications. These self-organising networks are ideal for rapidly changing scenarios that require connectivity even under hostile conditions. The excitement and promise that these networks generated transferred to the civilian space as well, fueling over a decade of research. However, they only experienced limited success in this new setting, leading to a fragmentation into several application-oriented sub-fields that dealt with narrower sets of constraints. In this dissertation, I postulate that the unique properties of these networks makes them much more error-prone than initially considered. Consequently, a focus on improving their capabilities rather than on mitigating their faulty nature made the design of general-purpose MANETs increasingly challenging. I support this argument through an extensive set of experiments that is informed by analysis of the literature, history, and properties of these networks. Ultimately, my work contributes to the field in three fronts: Firstly, motivated by the multiple challenges that research in this area presents, I designed and built MeshSim, a real-time network simulator. This new platform focuses on code-fidelity and enables me to follow a data-driven experimental cycle. Secondly, I present the Reactive Gossip Routing family of protocols, designed to provide reliable and scalable routing by mitigating the MANET properties that lead to faults. Using MeshSim, I evaluate these protocols experimentally under increasingly harsher conditions and compare their effectivity against the incumbents in the literature. Finally, I demonstrate through experimentation that distance-vector routes are ill suited for MANETs due to a geographical-spreading effect they induce, a result that extends to many routing metrics when used in shortest-path algorithms.
... The wireless local area network (WLAN) standard amendment IEEE 802.11s [1,2] integrates basic mesh functions directly into the WLAN link layer, providing the basis for interoperable mesh solutions to realize smart city applications [3][4][5]. For this purpose, 802.11s defines mandatory mechanisms for spontaneous networking and message forwarding (routing) between mesh nodes. ...
Full-text available
WLAN mesh networks are one of the key technologies for upcoming smart city applications and are characterized by a flexible and low-cost deployment. The standard amendment IEEE 802.11s introduces low-level mesh interoperability at the WLAN MAC layer. However, scalability limitations imposed by management traffic overhead, routing delays, medium contention, and interference are common issues in wireless mesh networks and also apply to IEEE 802.11s networks. Possible solutions proposed in the literature recommend a divide-and-conquer scheme that partitions the network into clusters and forms smaller collision and broadcast domains by assigning orthogonal channels. We present CHaChA (Clustering Heuristic and Channel Assignment), a distributed cross-layer approach for cluster formation and channel assignment that directly integrates the default IEEE 802.11s mesh protocol information and operating modes, retaining unrestricted compliance to the WLAN standard. Our concept proposes further mechanisms for dynamic cluster adaptation, including subsequent cluster joining, isolation and fault detection, and node roaming for cluster balancing. The practical performance of CHaChA is demonstrated in a real-world 802.11s testbed. We first investigate clustering reproducibility, duration, and communication overhead in static network scenarios of different sizes. We then validate our concepts for dynamic cluster adaptation, considering topology changes that are likely to occur during long-term network operation and maintenance.
Conference Paper
Full-text available
Multihop cellular networks used to remove the limitations on the cell capacity and the coverage, dead spot, hot spot issues & to get high throughput. In the Mobile stations near cell edge the interference of cochannel problem becomes severe; it is one of the major things to affect performance of the network. So in this paper we going to address these problems and also we concentrated on the traffic congestion and quality of service. Here we present a survey on reducing these issues by implementing orthogonal frequency division multiplexing access technique and for the interference coordination, Intercell interference coordination (ICIC) had been investigated. For the analysis purpose we focused on IEEE 802.16j/m specification in the downlink of OFDMA-based MCNs with time division duplex mode & we used novel frequency reuse scheme for the reducing interference and we provide hand over mechanisms and investigation continues importantly for the coverage of more users in MCNs.
Full-text available
IEEE 802.11 consists of one of the most used wireless access technologies, which can be found in almost all consumer electronics devices available. Recently, Wake-up Radio (WuR) systems have emerged as a solution for energy-efficient communications. WuR mechanisms rely on using a secondary low-power radio interface that is always in the active operation mode and is in charge of switching the primary interface, used for main data exchange, from the power-saving state to the active mode. In this paper, we present a WuR solution based on IEEE 802.11 technology employing transmissions of legacy frames by an IEEE 802.11 standard-compliant transmitter during a Transmission Opportunity (TXOP) period. Unlike other proposals available in the literature, the WuR system presented in this paper exploits the PHY characteristics of modern IEEE 802.11 radios, where different signal bandwidths can be used on a per-packet basis. The proposal is validated through the Matlab software tool, and extensive simulation results are presented in a wide variety of scenario configurations. Moreover, insights are provided on the feasibility of the WuR proposal for its implementation in real hardware. Our approach allows the transmission of complex Wake-up Radio signals (i.e., including address field and other binary data) from legacy Wi-Fi devices (from IEEE 802.11n-2009 on), avoiding hardware or even firmware modifications intended to alter standard MAC/PHY behavior, and achieving a bit rate of up to 33 kbps.
Throughout the next decade, 802 wireless systems will become an integral part of fourth generation (4G) cellular communication systems, where the convergence of wireless and cellular networks will materialize through support of interworking and seamless roaming across dissimilar wireless and cellular radio access technologies. IEEE 802 Wireless Systems clearly describes the leading systems, covering IEEE 802.11 WLAN, IEEE 802.15 WPAN, IEEE 802.16 WMAN systems' architecture, standards and protocols (including mesh) with an instructive approach allowing individuals unfamiliar with wireless systems to follow and understand these technologies. Ranging from digital radio transmission fundamentals, duplex, multiplexing and switching to medium access control, radio spectrum regulation, coexistence and spectrum sharing, this book also offers new solutions to broadband multi-hop networking for cellular and ad hoc operation. The book Gives a comprehensive overview and performance evaluation of IEEE 802.11, 802.15 and 802.16 Includes a tutorial like introduction to the basics of wireless communication Discusses challenges in mesh/multi-hop relaying networks and provides profound solutions for their realization with 802 Wireless Systems Covers spectrum sharing on different levels and provides solutions for coexistence, cooperation and interworking of 802 Wireless Systems that are following the same or different standards, but share the same spectrum Includes a detailed overview and introduction on cognitive radio and dynamic spectrum access Accompanying website contains simulation software and provides slides of the figures and tables from the book ready for course presentation This book is an essential text for advanced undergraduate students with a basic working knowledge of wireless communication, graduate students and engineers working in the field of wireless communications.
Wireless mesh networks (WMNs) consist of mesh routers and mesh clients, where mesh routers have minimal mobility and form the backbone of WMNs. They provide network access for both mesh and conventional clients. The integration of WMNs with other networks such as the Internet, cellular, IEEE 802.11, IEEE 802.15, IEEE 802.16, sensor networks, etc., can be accomplished through the gateway and bridging functions in the mesh routers. Mesh clients can be either stationary or mobile, and can form a client mesh network among themselves and with mesh routers. WMNs are anticipated to resolve the limitations and to significantly improve the performance of ad hoc networks, wireless local area networks (WLANs), wireless personal area networks (WPANs), and wireless metropolitan area networks (WMANs). They are undergoing rapid progress and inspiring numerous deployments. WMNs will deliver wireless services for a large variety of applications in personal, local, campus, and metropolitan areas. Despite recent advances in wireless mesh networking, many research challenges remain in all protocol layers. This paper presents a detailed study on recent advances and open research issues in WMNs. System architectures and applications of WMNs are described, followed by discussing the critical factors influencing protocol design. Theoretical network capacity and the state-of-the-art protocols for WMNs are explored with an objective to point out a number of open research issues. Finally, testbeds, industrial practice, and current standard activities related to WMNs are highlighted.
Conference Paper
We propose a simple protocol for authentication using only a password. The result of the protocol is a cryptographically strong shared secret for securing other data-- e.g. network communication. SAE is resistant to passive attack, active attack, and dictionary attack. It provides a secure alternative to using certificates or when a centralized authority is not available. It is a peer-to-peer protocol, has no asymmetry, and supports simultaneous initiation. It is therefore well-suited for use in mesh networks. It supports the ability to tradeoff speed for strength of the resulting shared key. SAE has been implemented for 802.11-based mesh networks and can easily be adapted to other wireless mesh technology.
In recent years WLAN technology has become the common wireless access technology for mobile computing. Additional to infrastructure access to WLAN networks, peer-to-peer and mesh networking are currently gaining in interest. Mesh networking techniques using WLAN are being standardized in IEEE 802.11s. This article describes use cases, the main technical issues, and a set of potential solutions for mesh network development. Furthermore, an overview of the standardization activities in IEEE 802.11s is presented. The key technical aspects of mesh networks identified are topology creation, routing, medium access control, security, quality of service, and power efficiency
MDA Simulation Study: Robustness to Non-MDA Interferers
  • Y Chen
  • S Emeott
Y. Chen, and S. Emeott, " MDA Simulation Study: Robustness to Non-MDA Interferers, " IEEE 802 Plenary Meeting, Submission 11-07-0356-00-000s, Orlando, FL, Mar. 2007.
Wireless Meshing with the One Laptop Per Child
  • J Cardona
J. Cardona, " Wireless Meshing with the One Laptop Per Child, " [8] open80211s,