0% found this document useful (0 votes)
107 views9 pages

IEEE 802.11s Wireless LAN Mesh Network Technology: Hidenori Aoki, Shinji Takeda, Kengo Yagyu and Akira Yamada

The document summarizes an IEEE journal article about wireless LAN (WLAN) mesh network technology. It describes how WLAN mesh networks consist of WLAN devices that communicate directly with each other without base stations, forming a multi-hop network. It discusses challenges like throughput degradation and proposes coordinating routing, congestion control and other functions at the MAC layer to address these issues and enable high-speed wireless networks without infrastructure. It provides an overview of the WLAN mesh network architecture and its components.

Uploaded by

givangustavo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
107 views9 pages

IEEE 802.11s Wireless LAN Mesh Network Technology: Hidenori Aoki, Shinji Takeda, Kengo Yagyu and Akira Yamada

The document summarizes an IEEE journal article about wireless LAN (WLAN) mesh network technology. It describes how WLAN mesh networks consist of WLAN devices that communicate directly with each other without base stations, forming a multi-hop network. It discusses challenges like throughput degradation and proposes coordinating routing, congestion control and other functions at the MAC layer to address these issues and enable high-speed wireless networks without infrastructure. It provides an overview of the WLAN mesh network architecture and its components.

Uploaded by

givangustavo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

NTT DoCoMo Technical Journal Vol. 8 No.

IEEE 802.11s Wireless


LAN Mesh Network
Technology
Hidenori Aoki, Shinji Takeda,
Kengo Yagyu and Akira Yamada

A wireless LAN (WLAN) mesh network consists of WLAN


devices with relay functions that communicate directly with
each other instead of communicating via base stations. To
solve problems like throughput degradation and congestion,
a technology is proposed that enables coordination between
routing, congestion control, and other functions on the MAC
layer. With this technology, a high-speed wireless network
can easily be constructed even at a location with no network
infrastructure such as a WLAN access point.

1. Introduction
WLAN mesh network technology, which features flexible
broadband network configurations independent of the fixed net-
work, is attracting attention as an elemental technology for
future ubiquitous networks consisting of various types of termi-
nals including digital appliances, personal computers and
mobile terminals [1].
Diverse scenes can be imagined for WLAN mesh networks.
They can be used to achieve home networks, for extending the
coverage area of enterprise WLAN networks, and for construct-
*1
ing ad hoc networks . A WLAN mesh network is formed by
having neighboring terminals connect with each other directly
by wireless means instead of going through centralized control
equipment such as base stations. In this type of network, data
sent out from a terminal arrives at its destination via a sequence
of wireless terminals resulting in a multi-hop wireless network
configuration.
Here, as a wireless system for interconnecting terminals, we
apply WLAN technology conforming to the standard specifica-
*2
tions of IEEE 802.11 , an international WLAN standard [2].

*1 ad hoc network: A network configured by interconnecting mobile terminals with-


out requiring base stations or access points.
*2 IEEE 802.11: An international wireless LAN standard established by the Institute
of Electrical and Electronics Engineers (IEEE), a non-profit association in the
United States.

13
WLAN technology is finding widespread use as a means of on the MAC layer, and proposed the main results of that study
achieving broadband wireless communications, and it is an area to IEEE 802.11s [7].
of ongoing technical innovation especially in Quality of Service The IEEE 802.11s task group was formed in May 2004 to
*3
(QoS) technology [3] and wireless high-speed techniques (600 standardize the technologies that would be needed to deploy
Mbit/s) [4]. WLAN mesh networks. This work involved the creation of
WLAN mesh networks feature higher data transmission rate usage models and requirements necessary for selecting pro-
due to shortened communication distance, expanded network posed technologies and the preparation of formal procedures for
capacity through spatial frequency reuse, automatic network making selections. A Call For Proposal (CFP) was issued in
configuration, and improved robustness due to a route recovery January 2005. By the time of a meeting held in March 2006, 2
mechanism. candidates out of 15 submissions had survived, and these were
However, multi-hop wireless networks are not problem free. eventually combined into a single draft version of a standard
For example, their operation can be affected by hidden termi- specification [8][9] based on the DoCoMo proposal. The plan
*4 *5
nals and exposed terminals that are associated with degrada- from here on is to refine the specifications into a form that will
*6
tion of throughput characteristics, and they also suffer from win final approval. These standardization activities are expected
network congestion [5]. These problems depend heavily on the to be completed by June 2008.
routing protocol used to determine routes and on the radio The following chapters will outline the system architecture
access control scheme and radio resource management scheme of WLAN mesh networks and describe the main technological
*7
implemented on the Medium Access Control (MAC) layer . To components of that architecture.
solve these problems so that the advantages of WLAN mesh
networks can be used to the fullest, it is important that major 2. Overview of WLAN Mesh Networks
functions implemented on the MAC layer operate in coordina- 2.1 Device Types and Network Configuration
tion with the routing protocol in real time [6]. As shown in Figure 1, a WLAN mesh network consists of
Against the above background, we investigated a WLAN Mesh Points (MPs) equipped only with WLAN mesh network
mesh network technology that implements the routing protocol functions, Mesh Access Points (MAPs) equipped with a WLAN

MPP

MP MP

MP

MAP MAP

STA STA STA STA

Figure 1 Configuration of a WLAN mesh network

*3 QoS technology: Techniques for securing optimal bandwidth according to the degrades is called the “hidden terminal problem.”
purpose of communications and guaranteeing the quality required by that type of *5 Exposed terminals: Neighboring terminals whose mutual communications prevent
communications. other terminals from communicating. A phenomenon by which communications
*4 Hidden terminals: Terminals located in areas that cannot receive each other’s sig- are suppressed in this way preventing required throughput from being attained
nals nor determine the other’s communication status. A phenomenon by which and degrading call quality is called the “exposed terminal problem.”
packets submitted simultaneously by hidden terminals collide and call quality

14
NTT DoCoMo Technical Journal Vol. 8 No.2

access point function in addition to MP functions, a MP collo- tional blocks of this architecture.
cated with a mesh Portal (MPP) equipped with a gateway func- 1) Mesh Topology Learning, Routing and Forwarding
tion for connecting to an external network in addition to MP This block contains a function for discovering neighboring
*9
functions, and STAtions (STAs) that are legacy WLAN stations nodes, a function for obtaining radio metrics that provide
having no WLAN mesh network functions. A Wireless information on the quality of wireless links, routing protocol for
*8
Distribution System (WDS) frame is used here to transfer data determining routes to transfer packets to their destinations using
among the MP, MAP and MPP nodes. MAC addresses as identifiers, and a packet forwarding function.
Here, to make efficient use of radio resources, routing protocol
2.2 Usage Model must make use of radio metrics and multiple frequency channels
The IEEE 802.11s standard envisions a small- to medium- in accordance with radio conditions.
scale WLAN mesh network configured with a maximum of 32 2) Mesh Network Measurement
MPs (MAPs included). Practically, each MAP can be connected This block contains functions for calculating radio metrics
to many STAs enabling the entire network to accommodate sev- used by routing protocol and for measuring radio conditions
eral hundred terminals. Multiple WLAN mesh networks can within the WLAN mesh network for use in frequency channel
also be interconnected to further expand network scale. selection.
We expect WLAN mesh network technology to be applica- 3) Mesh Medium Access Coordination
ble to a wide variety of usage environments. These might be This block includes functions for preventing degraded per-
home networks that connect digital appliances, personal com- formance due to hidden and exposed terminals, functions for
puters, and other devices; office networks that make up corpo- performing priority control, congestion control, and admission
rate LANs; college campus networks and public access net- control, and a function for achieving spatial frequency reuse.
works for commercial districts; and ad hoc networks for inter- 4) Mesh Security
connecting mobile terminals [10]. This block contains security functions for protecting data
frames carried on the WLAN mesh network and management
2.3 System Architecture frames used by control functions such as routing protocol. It
Figure 2 shows system architecture for WLAN mesh net- assumes the use of WLAN security schemes defined by the
*10
work technology [11]. The following outlines the main func- IEEE 802.11i standard [12].

Upper Layers
Interworking Mesh Configuration and Management

MAC Mesh Topology Mesh Network Mesh Medium Mesh


802.11s Learning, Routing Measurement Access Coordination Security
WLAN Mesh and Forwarding
(Layer2)

Lower MAC Enhancement for Mesh (11e/n+)

PHY
(Layer1) IEEE802.11 PHY IEEE802.11 a/b/g/j/n

Figure 2 System architecture

*6 Throughput: Effective amount of data transmitted without error per unit time. *9 Radio metrics: Indices used in routing that take the quality of radio links into
*7 MAC layer: A layer that has a control function for preventing packet collisions account.
when sharing communication lines among multiple nodes. This layer is a lower *10 IEEE 802.11i: A standard defining wireless LAN security functions.
sublayer of the data link in the OSI 7-layer model.
*8 WDS frame: Unit of data used for communicating between wireless access points.

15
5) Interworking tocol and radio metric is essential to achieving routing technolo-
As part of the IEEE 802 standard typical of wired Ethernet, gy appropriate for the actual usage environment. At the same
a WLAN mesh network must conform to IEEE 802 network time, the formation of a WLAN mesh network requires that all
architecture. Accordingly, to connect to other networks, a trans- MPs select the same routing protocol and radio metric.
*11
parent bridge function must be implemented in the MPP situ- Consequently, combination of routing protocol and radio
*12
ated at the network boundary, and each WLAN mesh network metric is defined as a profile , and a function is prescribed to
must operate as a broadcast network so that forwarded packets notify the profile that is selected by each MP to neighboring
can be delivered to all terminals connected to the LANs. MPs[17].
6) Mesh Configuration and Management 2) Routing Protocol
This block includes a WLAN interface used for automatic Layer-3 routing protocol, which has been extensively
setting of each MP’s Radio Frequency (RF) parameters (fre- researched for some time, can be broadly divided into two
quency channel selection, transmit power, etc.), for QoS policy types: proactive and reactive [16]. The proactive type establish-
management, etc. es routes beforehand regardless of whether communications are
in progress, while the reactive type establishes routes as needed
3. Details of Elemental Technologies for communication purposes. The characteristics exhibited by
Of the various elemental technologies making up WLAN these schemes are heavily affected by external factors such as
mesh networks, routing technology, congestion control technol- network size and the speed of mobile nodes. Nevertheless, it is
ogy, and dynamic frequency channel allocation technology are desirable that a default routing protocol, which all terminals will
considered to be especially important. These technologies are be required to implement, be capable of minimizing protocol
described below. complexity while exhibiting high performance in diverse usage
environments. With this in mind, we have proposed a scheme
3.1 Routing Technology that builds upon the Ad hoc On-Demand Distance Vector rout-
Routing protocol and radio metrics are important elements ing (AODV) scheme [18], a reactive-type routing protocol. Our
in determining the performance of a WLAN mesh network. To proposed scheme, called Radio Metric AODV (RM-AODV)
date, however, many routing protocols [13] and radio metrics [19], possesses the following features as enhancements to
[14][15] have been proposed, and achieving interoperability AODV.
between devices of different vendors has been a serious prob- a) Support of radio metrics
lem. In addition, the optimal routing protocol or radio metric The proposed routing protocol periodically checks radio
depends on the usage model [16], and to complicate matters conditions with neighboring nodes to select routes that further
even further, future standard technologies and vendor propriety stabilizes and minimizes the radio metric.
protocols are expected to be implemented in the years to come. b) Support of multiple WLAN interfaces
Against the above background, it is important to have a For MPs having multiple WLAN interfaces, the proposed
default routing protocol and radio metric that all devices will be routing protocol includes functions for using them in parallel
required to implement to ensure interoperability, and to have an and for using the interface having the lowest utilization ratio of
extensible framework that enables the implementation of vari- radio resources for any given destination. These functions allow
ous routing protocols and radio metrics optimized for different routing that maximizes system capacity in accordance with con-
usage environments. tinuously changing radio conditions.
1) Extensible Framework c) Support for Legacy 802.11 Stations
A framework that enables flexible selection of a routing pro- A MAP that manages STAs not equipped with routing func-

*11 Transparent bridge: Technology used for interconnecting LANs defined by IEEE *12 Profile: Equipment configuration information. In IEEE802.11s, “profile” refers to
802.1D. It enables terminals belonging to different LANs to be seen by each other routing-related configuration information.
as if they were operating on the same LAN.

16
NTT DoCoMo Technical Journal Vol. 8 No.2

tions enables a STA to participate in a WLAN mesh network by amount of interference, and other factors (Fig. 3 a).
maintaining a route to the destination on behalf of that STA. Next, the source node broadcasts a request packet through-
out the entire network. If, however, the source node happens to
We note here that IEEE 802.11s adopts the Hybrid Wireless be a MAP that accommodates STAs equipped with no routing
Mesh Protocol (HWMP), which incorporates RM-AODV with a protocol, it will send the request packet on behalf of the source
function added for establishing tree-based routes beforehand STA.
[9]. At this time, each relay node adds the value of the radio
3) Radio Metric metric for the upcoming wireless link to the existing value of
The quality of a WLAN mesh network depends on the qual- the radio metric in the request packet so that an accumulated
ity of the wireless links, on interference, and on the utilization radio metric value can be delivered to the destination node. In
ratio of radio resources [20]. To reflect all of these conditions the event that a relay node has more than one WLAN interface
*13
and to achieve easy implementation, we have adopted airtime and the radio metric value is the same for each, the WLAN
as a default radio metric [9]. interface for which the request packet arrives first will be select-
4) RM-AODV Operation Overview ed in order to take the congestion state of each interface into
Figure 3 shows RM-AODV operation. First, an MP per- account (Fig. 3 s).
forms a radio metric exchange with neighboring nodes. The Finally, the destination node selects the route having the
radio metric quantifies the quality of a wireless link as deter- smallest radio metric tabulated over an entire route (Fig. 3 d),
mined by wireless data transmission rate, amount of traffic, and notifies each relay node along that route of this selection

Radio metric changes even between the same Source node sends a request Process selects the WLAN interface
nodes if usage frequency of links differs. packet to all WLAN interfaces. with the lowest radio metric.

12 IF1 12 IF1
10
IF1 IF1 IF1 IF1
10 IF2 11 10 IF2

IF2 10 20 IF2 IF2 10 IF2


IF1 IF1
For identical radio metric
10 20 10
Radio metric increases values, the process selects
IF2 as the communication IF2 the WLAN interface that
distance lengthens. receives the request first.
a Exchange radio metric with each neighboring node sSelect WLAN interface

Selected route: The process informs relay nodes of


Total radio metric value: 20 (=10+10) selected route using a response packet.

12 IF1 IF1
10
IF1 IF1 IF1 IF1
10 IF2 11 IF2

IF2 10 20 IF2 IF2 IF2


IF1 IF1
10 20
IF2 IF2

dSelect destination-node route fFinalize route by response packet

Figure 3 RM-AODV operation overview

*13 Airtime: The actual time taken for packet transmission on a wireless link. Used as
an index for determining paths in IEEE 802.11s.

17
using a response packet (Fig. 3 f). In a manner similar to times the system capacity in terms of throughput by making
request-packet processing, a destination node that happens to be uniform use of multiple WLAN interfaces.
a STA will have its MAP reply with the response packet.
5) Characteristics Evaluation 3.2 Congestion Control Technology
We here present the results of evaluating the proposed pro- In a WLAN mesh network that assumes packet transfer
tocol by computer simulation. among MPs, the buildup of packets at relay equipment can
Figure 4 shows simulation results for 16 MPs placed ran- cause transmission delays and drops in throughput to occur
domly in a 50-m-square area. For comparison purposes, the fig- [21]. To prevent this problem from occurring in an efficient
ure shows characteristics when applying hop count ( the number manner while minimizing revisions to existing specifications for
of relay nodes) versus those for a radio metric as criteria for the MAC layer, proposals have been made for congestion con-
routing, with the results for 1 and 2 WLAN interfaces shown for trol technology that aims to adjust transmission rates between
each. On comparing the conventional scheme using hop count neighboring nodes through signaling [22][23].
and 1 WLAN interface with the proposed scheme using a radio The following outlines a congestion control method that
metric and 2 WLAN interfaces, the latter is found to achieve 2.3 prevents congestion at relay nodes by appropriately setting
parameters known to have a high degree of freedom in the
*14
10 Enhanced Distributed Coordination Access (EDCA) [3]. This
method has been defined as a mandatory function in the
8 IEEE802.11s standard.
1) Principle Behind Generation of Congestion
Throughput (Mbit/s)

Improved by
about 2.3 times Figure 5 shows the mechanism of congestion generation
6
within a WLAN mesh network. The scenario shown depicts two-
way communication between MP1 and MP5 via intermediary
4 MPs. If we compare throughput characteristics for the links near-
est the packet-originating nodes (links L1-2, L5-4) with those for the
1-interface/hop count (conventional) links nearest the packet-destination nodes (links L4-5, L2-1), we see
2 1-interface/radio metric
2-interface/hop count that the latter represents a decrease to about 20% of the former.
2-interface/radio metric (proposed scheme)
We here examine the routes between MP1 and MP3 refer-
0
0 2 4 6 8 10 12 14 ring to Figure 6. If using EDCA as the radio access mecha-
Input data (Mbit/s)
nism, the opportunity for packet transmission would normally
Figure 4 RM-AODV characteristics evaluation be uniform among MP1, MP2 and MP3. In this case, however,
MP2 acts as a relay node requiring

1.42 Mbit/s 0.53 Mbit/s 0.50 Mbit/s 0.26 Mbit/s


it to send packets in both directions.

MP1
L1–2
MP2
L2–3
MP3
L3–4
MP4
L4–5
MP5
This means that MP2 has relatively
lower packet-transmission opportu-
L2–1 L3–2 L4–3 L5–4 nity and that packet buildup and
0.26 Mbit/s 0.50 Mbit/s 0.53 Mbit/s 1.42 Mbit/s transmission-buffer overflow can
occur in that node resulting in sig-
Data is sent and received between the terminals on both ends
nificantly degraded throughput
Figure 5 Mechanism of congestion generation in a WLAN mesh network characteristics.

*14 EDCA: A radio access method for ensuring communication quality on wireless
LAN standardized in IEEE802.11e.

18
NTT DoCoMo Technical Journal Vol. 8 No.2

2) Outline of Congestion Control Technology


MP1 MP2 MP3
Figure 7 shows the proposed conges-
tion control technology. In the figure,
MP(n) receives packets at transmission rate
J(n–1) from upstream node MP(n–1) and Transmission
buffer
sends packets to node MP(n+1) at transmis- Relay packets
Transmission Transmission
sion rate J(n). The following condition must packets packets

be met here for congestion not to occur at Figure 6 Buffer state at time of congestion
relay node MP(n).
Relay node

J (n–1) < J (n)


J (n–1) J (n)
MP (n–1) J (n) MP (n) MP (n+1)
Accordingly, downstream node MP(n)
CCR
needs to convey its maximum transmission message

rate to upstream node MP(n–1), and to do


Figure 7 Overview of congestion control technology
this, it sends a Congestion Control Request
(CCR) packet. The upstream node now MP1

transmits packets at a transmission rate


lower than the one specified in the CCR f1 (data)
MP3 MP4
thereby solving the congestion at the relay
f2 (data)
node and improving end-to-end throughput
as a result. Although Fig. 7 only shows f3 (voice)
MP2
packet traffic in one direction, the same
type of processing would be needed in both
directions in the case of bidirectional traffic. Figure 8 Simulation topology
To make such congestion control tech-
nology as effective as possible, it is important that studies be Because a CCR packet can specify the maximum rate for
made on optimal settings for traffic-observation period and each of the four types of QoS classes specified in [3], it
CCR-sending cycle and on transmission-rate control methods. It becomes possible even when applying congestion control to
is desirable, in particular, that adaptive rate control be per- regulate data traffic flow without having to reduce the through-
formed using EDCA to minimize changes to hardware. put of high-priority traffic such as voice calls.
3) Effect of Congestion Control Technology Figure 9 shows simulations results. It can be seen that the
*15
We here present simulation results for a topology having application of congestion control improves total throughput by
multiple flows in a single network as shown in Figure 8. In this about 30%. Furthermore, since transmission rate can be speci-
simulation, transmission rate is controlled by increasing or fied for each QoS class, high-priority voice traffic (f3) can be
decreasing the value of Arbitration Inter Frame Space Number kept at a fixed rate while improving the throughput of data traf-
*16
(AIFSN) , an EDCA parameter. Symbols f1, f2 in the figure fic (f1, f2) even when applying congestion control by the method
denote data traffic while symbol f 3 denotes voice traffic to presented here.
which a higher QoS class has been set.

*15 Topology: Positional relationship of devices, network configuration, etc.


*16 AIFSN: Time interval before beginning data packet transmission as defined in
EDCA.

19
from occurring requires that each MP select the same frequency
3
With congestion control channel. To this end, frequency channel priority information
2.5 Without congestion control
can be used to enable a common frequency channel to be select-
Link throughput (Mbit/s)

ed for the entire network even if each MP chooses a frequency


2
channel independently. This information is exchanged among
1.5 neighboring nodes and the frequency channel used by the node
with the highest frequency channel priority is selected as the
1 common frequency channel.
2) Multi Channel Mode
0.5
In this mode, it is assumed that each MP possesses multiple
0 WLAN interfaces and that multiple frequency channels will be
f1 f2 f3 Total
used to good effect in a WLAN mesh network. Such an MP is
Figure 9 Throughput characteristics with and without
congestion control
able to dynamically allocate a frequency channel to each wire-
less link in accordance with network topology and traffic condi-
3.3 Dynamic Frequency Channel Allocation Technology tions. In the multi-channel-mode example shown in Figure 10,
Four channels in the 2.4-GHz band and eight channels in the the WLAN interfaces that are to use the same frequency chan-
*17
5-GHz band are available for current WLAN equipment in nel between MPs are grouped together as clusters and a fre-
Japan. In conventional systems, an access point selects an opti- quency channel is determined for each cluster. This framework
mal frequency channel and instructs the terminal awaiting con- for allocating frequency channels can increase network capacity
nection to use that frequency channel. A WLAN mesh network, by load balancing [24] and can even solve the hidden-terminal
however, has a distributed network configuration having no and exposed-terminal problems [25].
equipment that performs centralized control, and it is left to
each MP to decide which frequency channel to use. To form a 4. Conclusion
stable network and increase network capacity in this situation, it This article presented an overview of WLAN mesh net-
is important that a dynamic frequency channel allocation tech- works and described system architecture. It also described the
nology be adopted. The following describes two frequency elemental technologies needed to configure a WLAN mesh net-
channel selection methods defined as mandatory functions in work, namely, routing, congestion control technology, and
IEEE 802.11s [7]. dynamic frequency channel allocation technology. For the
1) Single Channel Mode future, we plan to continue researching WLAN mesh networks
Constructing a stable network to prevent network cutoffs as a platform technology for ubiquitous networks.

Cluster 2
MP1 Cluster 3

Cluster 1

MP4 MP1
MP2 MP4

MP3 MP2 MP3

(a) Single channel mode (b) Multi channel mode

Figure 10 Overview of dynamic frequency channel allocation

*17 Cluster: In this article, a group of wireless LAN interfaces that use the same fre-
quency channel within a wireless LAN mesh network.

20
NTT DoCoMo Technical Journal Vol. 8 No.2

References [16] E. M. Royer and C.K. Toh: “A Review of Current Routing Protocols for
[1] Y. Matsumoto, J. Hagiwara, A. Fujiwara, H. Aoki, A. Yamada, S. Ad Hoc Mobile Wireless Networks,” IEEE Personal Communications,
Takeda, K. Yagyu and F. Nuno: “A Prospective Mesh Network Based Apr. 1999.
Platform for Universal Mobile Communication Services,” IEICE General [17] W. S. Conner and H. Aoki: “Propose Extensible Approach for WLAN
Conference, B-5-245, Mar. 2006. Mesh Standardization,” IEEE802.11-5/0165r1, Mar. 2005.
[2] IEEE Std 802.11-1999 (Reaff 2003), Sep. 2003. [18] C. Perkins and E. Royer: “Ad-hoc On-demand Distance Vector Routing,”
[3] IEEE Std 802.11e, Nov. 2005. in proceedings of 2nd IEEE Workshop on Mobile Computing Systems
[4] IEEE P802.11n/D1.0, Mar. 2006. and Applications, 1999.
[5] S. Sakata, H. Aoki and K. Mase: “Mobile Ad Hoc Networks and Wireless [19] S. Takeda, K. Yagyu, H. Aoki and Y. Matsumoto: “Multi-Interface
LAN Mesh Networks,” IEICE Transactions (B), Vol. J89-B, No. 6, pp. Oriented Radio Metric On-demand Routing Protocol for Layer-2 Mesh
811–823, Jul. 2006 (In Japanese). Networks,” IEICE Technical Report, RCS2005-58, Jul. 2005.
[6] H. Aoki, W. S. Conner, K. Omae and Y. Matsumoto: “Proposal of [20] S. Takeda, K. Yagyu and Y. Matsumoto: “Load Balance Enhancement
IEEE802.11s Layer-2 Mesh Network Architecture,” IEICE Technical Technique for Wireless Mesh Networks,” IEICE Technical Report,
Report, RCS2005-56, Jul. 2005. RCS2004-17, Apr. 2004.
[7] H. Aoki, N. Chari, L. Chu, W. S. Conner, S. M. Faccin, D. Gurevich, V. [21] L. Yang and A. Yamada: “MAC Considerations for Mesh,” IEEE802.11-
Hasty, J. Jetcheva, S. Kangude, S. Saito and R. Taori: “802.11 TGs 04/760r0, Jul. 2004.
Simple Efficient Extensible Mesh (SEE-Mesh) Proposal,” IEEE802.11 doc- [22] A. Yamada, A. Fujiwara and Y. Matsumoto: “EDCA Based Congestion
ument 05/0562r0, Jul. 2005. Control Method for WLAN Mesh Networks,” 2005 IEICE General
[8] D. Eastlake: “Proposals for TGs,” IEEE802.11-05/597r21, Mar. 2006. Conference, B-5-151, Sep.2005.
[9] IEEE P802.11s/D0.01, Mar. 2006. [23] A. Yamada, S. Bahareh, A. Fujiwara and L. Yang: “Simulation Results for
[10] W. S. Conner: “IEEE 802.11 TGs Usage Models,” IEEE802.11- SEEMesh Congestion Control Protocol,” IEEE802.11-05/0568, Jun.
04/662r16, Jan. 2005. 2005.
[11] W. S. Conner: “IEEE 802.11 TGs Functional Requirements and Scope,” [24] K. Yagyu, A. Fujiwara, S. Takeda, K. Omae, H. Aoki and Y. Matsumoto:
IEEE802.11 document 04/1174r13, Jan. 2005. “Topology and Traffic Aware Channel Assignment for Layer-2 Mesh
[12] IEEE P802.11i-2004, Jul. 2004. Networks,” IEICE Technical Report, RCS2005-61, Jul. 2005.
[13] http://www.ietf.org/html.charters/manet-charter.html [25] A. Fujiwara and Y. Matsumoto: “Centralized Channel Allocation
[14] D. D. Couto, D. Aguayo, J. Bicket and R. Morris: “High Throughput Path Technique to Alleviate Exposed Terminal Problem in CSMA/CA-Based
Metric for Multi-Hop Wireless Networks,” ACM MobiCom, Sep. 2003. Mesh Networks,” IEICE TRANS. COMMUN, Vol. E88-B, No. 3, Mar.
[15] R. Draves, J. Padhye and B. Zill: “Routing in Multi-Radio, Multi-Hop 2005.
Wireless Mesh Networks,” ACM MobiCom, Sep. 2004.

21

You might also like