Wireless Mesh Networks: a technology for

community-deployed networks in
developing countries
Marvin Sánchez G.
Dept of Communication Systems,
Royal Inst of Technology (KTH),
Electrum 418, 164 40 Kista, Sweden.
Email: marvins@radio.kth.se
Oscar Somarriba J.
Dept of Communication Systems,
Royal Inst of Technology (KTH),
Electrum 418, 164 40 Kista, Sweden.
Email: oscars@radio.kth.se
Jens Zander
Dept of Communication Systems,
Royal Inst of Technology (KTH),
Electrum 418, 164 40 Kista, Sweden.
Email: jens.zander@radio.kth.se

Abstract
In this paper we present a methodology for planning and capacity evaluation of community-deployed wireless mesh networks
in rural areas, based on radio equipment parameters for WiMax technologies on 3.4 GHz frequency band, and also on 2.4 GHz
bands (WiFi physical layer). The capacity is described by the maximum end-to-end transmission rate (throughput) provided to each
node that composes the network utilizing interference-based link scheduling. A community-deployed scenario for Internet access
is studied utilizing 13 sites in rural communities located in the north center highland region of Nicaragua in Central America. To
reduce the Internet service cost shared common access points are used in mesh configuration under asymmetric traffic demands.
The radio propagation environment is estimated utilizing free-available tools (digital maps and freeware program) that can be used
by the community. Next, the (upper-bound) capacity resulting from the community-deployed approach is computed by finding the
link transmission schedule applying nonlinear optimization. Finally, in this work we also motivates the need to further develop
easy-to-use free-software tools coupled with mesh network products to make this kind of community deployment ubiquitous.

I. I NTRODUCTION
Since 1990s important policy reforms has been done by countries of Central America. The reforms have fostered private
sector investment in telecommunication services and infrastructure providing benefits to their citizens who now, at least in
urban areas, have access to several telecommunication services. However, although we have observed tremendous growth in
infrastructure and services (voice and data), it seems to be economically difficult to extend these benefits to rural areas without
the financial support of the government [?]. So far, the rural areas far from the main cities have been unattractive to private
investors due to low population density, long distance, and usually irregular terrain together with low incomes of potential
users. In particular, this problem continues to be a challenge to the governments of developing countries, whose ultimate goal
is to eliminate the digital divide.
In countries with so many needs, one may think that investment in telecommunications might have low priority level.
However, for instance the Nicaraguan government recognizes that the development of and access to telecommunication means
is a key enabler [?] in promoting job creation, knowledge-based growth, business innovation, access to valuable information, and
can be utilized to improve education and health-care assistance. Fig.1 shows the general vision of the Regulator of Nicaragua
(TELCOR). To guaranty universal access to telecommunications to all citizens of the country requires reducing what is called
by the regulator the market efficiency gap and the real access gap (aimed to be financed by the Telecommunication funds
called, FITEL by its name in spanish).
In this context wireless mesh networks are an appealing technology to provide rural area communication. The used of
this technology could influence important changes to the current mechanisms used regarding the financial support done by
governments. By taking advantage of the self-organizing capability of mesh networks the communities themselves or even new
potential local players can take an active part in the solution to their own needs of telecommunications. However, even in the
foreseen scenario where potential users are assumed to be fixed, urge the availability of easy-to-use and free-of-charge tools
for network planning and capacity estimation as the one presented in this paper. Recent work on this area have make used of
simulation tools to analyze the performance in mesh networks like the one in [?] and [?], however their work has focus on the
performance analysis and testbed rather than a common frame work to facilitate community network planning. The general
contribution of our work is on providing a general methodology of planning for community-deployed mesh networks making
used of the free software called ”Radio mobile” that incorporate topographical information (with free available cartography)
combined with the use google earth. Using this tools we will present an approach that could be used for the network deployment
and we will present a methodology for capacity evaluation. The capacity is described by the maximum end-to-end transmission
rate (throughput) provided to each node that composes the network.
The remainder of the article is organized as follows. An introduction on wireless mesh networks is presented, why they can
be used as an alternative for fixed broadband wireless access, and different issues that need to be solved for them to work
Fig. 1. The general vision of the Nicaraguan Regulator regarding universal access (version of the original written in spanish, www.telcor.gob.ni).

properly. Next we present the system model and methodology we developed to study the capacity analysis of a rural region
in our country. We then describe the community-deployed networks and capacity evaluation of them. Finally, we have the
concluding remarks and discussions of our study.
II. M ETHOD OF A NALYSIS
We will present as an illustrative example the deployment scenario for Internet access for 13 sites in rural communities of
Nicaragua (telecenter candidate sites provided by the Project Coordination Unit (UCP) of the regulatory body of Nicaragua,
TELCOR).
The communities are located in the north central highlands region of the country, at the departments of Estelı́, Madriz and
Nueva Segovia. Fig. 2 shows the geographical locations for the communities of interest utilizing a digital map of the terrain.
The general research question of interest that we would like to answer is:
• What is the capacity that a communiy-deployed MWN can provide subject to low cost deployment constrain for internet
access for the 13 community telecenters?
For low-cost user-deployment we consider the following:
• Share internet access: for low-cost internet service (network service recurrent cost) nodes share access provider.
• Economy of scale: Two cases are considered
1) Utilization of radio equipment parameters on the 2.4GHz frequency bands (WiFi physical layer). Similar rules to the
FCC part 15 have been adopted in Nicaragua throug the rules in AA001-2006 [?].
2) Utilization of radio equipment parameters for WiMax technologies on 3.4GHz frequency band. This bands have been
allocated in Nicaragua for Broadband Wireless access technologies (Primary usage).
 We follow a similar approach to the one presented by the authors in [?] but also considering the use of mesh networking
with parameters for WiMax technologies like the one utilized in [?].
A. Capacity Evaluation Example
To share common Internet access points in a mesh configuration we assume asymmetric traffic demand from each node to
a gateway node connected to the internet and viceversa. Under this assumption the average traffic from node i to the internet
gateway, node g, is defined as λgi and from the gateway to node i as λig , hence the total network traffic load is given by:
X X
λ= λgi + λig
∀i6=g ∀i6=g

We assume that the average traffic from a node to the internet is 10% of the traffic from the internet to that node. This
is a reasonable assumption if for instance the gateway is connected via an ADSL service over the PSTN and we share this
connection providing similar capacity to all telecenters.
The user-deployed scenario is analyzed estimating the radio propagation environment of located nodes. The path-losses in
the network are derived utilizing the digital map GTOPO30 with the Longley-Rice model [?] as implemented by the Radio
Mobile freeware program by VE2DBE [?].
The (upper-bound) capacity resulting from a user-deployed approach is evaluated by founding the link transmission schedule
applying nonlinear optimization [?], [?].
B. Analysis and User-deployed Methodology:
We can summarize the general methodology applied in the following sections by the following steps:
I) Network Deployment:
i. Determine the radio network parameters that will be used
ii. Define the Internet gateway to be used (e.g. closer available land-based infrastructure)
iii. Determine the link-budget between stations including the gateway (Gmatrix)
iv. Determine the current network topology
v. If there are nodes can not reach the gateway, add digital-repeater nodes (digipeaters) to extend the connectivity by
multihoping and go to step Iiii
II) Capacity analysis:
i. Define the external traffic load demand from each node to the Internet gateway and viceversa

Fig. 2. Rural communities of Nicaragua for case study (120km × 80km)

 ii. Determine the routing matrix
iii. From steps IIi and IIii estimate the link traffic load of the network.
iv. Compute the equivalent link path routing matrix in order to estimate the real relative traffic load Rmesh (in this step,
we remove those links that we are not going to use in the communication, i.e. weak links)
v. Determine the network capacity by interference-based scheduling (STDMA)
Recreation of the user-deployment is related to step Ii to Iv where users try to connect their communities to the Internet
gateway. At the early stage, the MWN start composed only by the nodes located at the communities. Then, progressively from
the gateway towards the communities, the users add digital-repeaters to connect the network.
To determine the network topology we assume that the hardware utilized has the physical layer parameters either shown in
table Ior in table II.
In the first case (Wi-Fi Technology), we consider those parameters are based on utilizing equipment with similar physical
layer to IEEE 802.11g operating on the 2.4GHz frequency band [?] but under FCC part 15 rules. Note that the receiver
sensitivity utilized is conservative as compared for instance with the specifications for the equipment in [?]. To find the path
losses, we assume utilization of the first channel on IEEE802.11g that corresponds to operate on the frequency range: 2401
MHz- 2423MHz. To determine the radio propagation path losses we have used the Radio Mobile freeware software with 20
meters antennas height.
Next we turn to WiMax technology, we also consider those parameters are based on utilizing equipment with similar physical
layer to IEEE 802.16 mesh mode operating on the 3.4GHz frequency band [?] but under FCC part 15 rules. Mesh Mode is
an optional topology for user-to-user communication in non-line of sight IEEE 802.16a. It is included in the standard to allow
overlapping, ad hoc networks in the unlicensed spectrum and extend the edges of the wireless range at low cost. Mesh support
has recently been extended into the licensed bands too.
To find the path losses, we assume utilization of the XXX?? channel on IEEE802.16x that corresponds to operate on the
frequency range: 3401 MHz- 3423MHz ??. To determine the radio propagation path losses we have used the Radio Mobile
freeware software with 20 meters antennas height.

TABLE I
P HYSICAL LAYER PARAMETERS

Frequency [GHz] 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4
Data rate 1 2 6 9 12 18 24 36 48 54
(R_set)[Mbps] *
Modulation 'CCK' 'CCK' 'BPSK' 'BPSK' 'QPSK' 'QPSK' '16QAM' '16QAM' '64QAM' '64QAM'
Code rate 1 1 1/2 3/4 1/2 3/4 1/2 3/4 2/3 3/4
Peak output 30 30 30 30 30 30 30 30 30 30
power** [dBm]
Transmitter 6 6 6 6 6 6 6 6 6 6
antenna gain
[dBi]
EIRP [dBm] 36 36 36 36 36 36 36 36 36 36
Receiver antenna 0 0 0 0 0 0 0 0 0 0
gain [dBi]
Bandwidth 22 22 22 22 22 22 22 22 22 22
[MHz]
Thermal Noise -174 -174 -174 -174 -174 -174 -174 -174 -174 -174
[dBm/Hz]
Receiver Noise 10 10 10 10 10 10 10 10 10 10
Figure [dB]
Noise Power -90.6 -90.6 -90.6 -90.6 -90.6 -90.6 -90.6 -90.6 -90.6 -90.6
[dBm]
Interference 2 2 2 2 2 2 2 2 2 2
Margin [dB]
Minimum Es/N0 0.6 3.6 4 5 7 9 12 16 20 21
[dB]
Sensitivity -88 -85 -84.6 -83.6 -81.6 -79.6 -76.6 -72.6 -68.6 -67.6
(minimum Rx.
Power)
[dBm]***
Maximum 124 121 120.6 119.6 117.6 115.6 112.6 108.6 104.6 103.6
allowed path
loss[dB]
*Rates 5.5 and 11 are omitted assuming that all nodes use parameters similar to the one utilize at the physical layer of equipment
based on IEEE802.11g.
** According to FCC part 15.247 (b)(1) and (b)(3)(i): the power must be reduced by 1dB for every 3dB that the directional gain of the
antenna exceeds 6dBi. For omnidirectional antennas FCC part 15.247(b)(3) and (b)(3)(iii) apply, i.e. the maximum EIRP must be
36dBm (4W).
*** By increasing the interference margin the minimum required receive power is increased and the network topology could be
modify.
C. Traffic model
We assume asymmetric traffic demand from each node to the gateway node connected to the internet. Under this assumption
the average traffic from node i to the internet gateway, node g, is defined as λgi and from the gateway to node i as λig . Hence,
the total network traffic load is given by:
X g X
λ= λi + λig
∀i6=g ∀i6=g

We assume that the average traffic from nodes to the internet is 10% the traffic from the internet to the nodes. We think
this is a reasonable assumption if for instance the gateway is connected via an ADSL (Asymmetric Digital Subscriber Line)
service and we would like to share this connection providing similar service (target to be equal) to all telecenters connected
to the same gateway. In ADSL over Plain Old Telephone Service (POTS), ITU G.992.1 Annex A standard, the downstream
rate is 12Mbps and the upstream rate is 1.3Mbps (which is 10.8% the downstream rate). With this assumption the total traffic
load can be rewritten by:
P g P i
λ= λi + λg
∀i6=g ∀i6=g (1)
= (N − 1)(λ(U L) + λ(DL) ) = (N − 1)(0.1λ(DL) + λ(DL) ) = 1.1λ(DL) (N − 1)
where λ(U L) is the uplink traffic (from node to the gateway), λ(DL) is the downlink traffic (from the internet to node) and
(N-1) the number of node interconnected to the internet gateway. Hence, we can also rewrite the downlink allocation per node
as:
λ
λig = λ(DL) = , ∀i
1.1(N − 1)

TABLE II
P HYSICAL LAYER PARAMETERS

Frequency [GHz] 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.4
Data rate 1.41 2.12 2.82 4.23 5.64 8.47 11.29 12.71
[Mbps]*
Modulation 'BPSK' 'BPSK' 'QPSK' 'QPSK' '16QAM' '16QAM' '64QAM' '64QAM'
Code rate 1/2 3/4 1/2 3/4 1/2 3/4 2/3 3/4
Peak output 28 28 28 28 28 28 28 28
power** [dBm]
Transmitter 17 17 17 17 17 17 17 17
antenna gain
[dBi]
EIRP [dBm] 45 45 45 45 45 45 45 45
Receiver antenna 17 17 17 17 17 17 17 17
gain [dBi]
Bandwidth 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5
[MHz]
Thermal Noise -174 -174 -174 -174 -174 -174 -174 -174
[dBm/Hz]
Receiver Noise 10 10 10 10 10 10 10 10
Figure [dB]
Noise Power -88.56 -88.56 -88.56 -88.56 -88.56 -88.56 -88.56 -88.56
[dBm]
Interference 2 2 2 2 2 2 2 2
Margin [dB]
Minimum Es/N0 -3.44 -1.44 -0.44 2.60 5.60 8.60 13.60 14.60
[dB]
Sensitivity -100 -98 -97 -94 -91 -88 -83 -82
(minimum Rx.
Power)
[dBm]***
Maximum 162 160 159 156 153 150 145 144
allowed path
loss[dB]
*Assumming that all the nodes use parameters similar to the one utilize at the physical layer of equipment BreezeMax from Alvarion based
on IEEE802.16 mode Mesh.
** According to FCC part 15.247 (b)(1) and (b)(3)(i): the power must be reduced by 1dB for every 3dB that the directional gain of the
antenna exceeds 6dBi. For omnidirectional antennas FCC part 15.247(b)(3) and (b)(3)(iii) apply, i.e. the maximum EIRP must be
36dBm (4W).
*** By increasing the interference margin the minimum required receive power is increased and the network topology could be
modify.
 In the following, to better explain the application of the method we utilize a simple network example.
D. Example of the method with a 4-nodes network
To illustrate the methodology we utilize the 4-nodes ring network shown in Fig. 2. The distance between nodes is set to
15km. For this illustrative example we use a distance dependent radio propagation model to find the path gain matrix. We use
the parameters from table I with omnidirectional antennas with 6dBi gain. With this parameters the G matrix is given by:
 
∞ - 129.8 - 134.3 - 129.8
 - 129.8 ∞ - 129.8 - 134.3 
G=  - 134.3 - 129.8
 [dB]
∞ - 129.8 
- 129.8 - 134.3 - 129.8 ∞
Hence, using the parameters in aforementioned table, the received power is:
 
∞ −87.8 −92.3 −87.8
 −87.8 ∞ −87.8 −92.3 
Pr = Pt + Gr + Gt + G =   −92.3 −87.8
 [dBm]
∞ −87.8 
−87.8 −92.3 −87.8 ∞
where Pt is the transmitted power, Gt and Gr are the antenna gains for transmission and reception, respectively, and G denotes
the path gain matrix.
Since the received sensitivity is −88dBm, the feasible links are those for which Pr ≥ −88dBm (sensitivity at rate
R0 = 1Mbps). The resulting network topology and routing table using Dijkstra algorithm with equal cost for all links (minimum
hop routing) are shown in Fig. 3. The node-arc incident matrix matrix, A to represent the network topology is given by:
 
−1 −1 1 0 0 0 1 0
 1 0 −1 −1 1 0 0 0 
A=  0

0 0 1 −1 −1 0 1 
0 1 0 0 0 1 −1 −1
From the above matrix we can verify that the (directed) link labels used are the following:

link label 1 2 3 4 5 6 7 8
Link (i, j) (2, 1) (4, 1) (1,2) (3,2) (2,3) (4,3) (1,4) (3,4)

3 1

 
0 2 2 4
 1 0 3 1 
r=
 2

2 0 4 
1 1 3 0
4

Fig. 3. Four nodes ring network

Gateway, routing, and link traffic

In this example we arbitrary define node 3 as the gateway for Internet access. Therefore the external traffic load in matrix
form can be written as:
  
0 0 0.1 0
 0 0 0.1 0  (DL)
Λ=  1 1 0 1 λ


0 0 0.1 0
To express the traffic load in vector form we label only the (S,D) pairs with traffic load higher than zero as p = 1 . . . 6 with
the following labels:

Label p 1 2 3 4 5 6
(S,D) (1,3) (2,3) (3,1) (3,2) (3,4) (4,3)
Hence the traffic load in vector form is given by:

£ ¤T
Λ= 0.1 0.1 1 1 1 0.1 λ(DL)
Using the source to destination pairs in Λ, we get the corresponding link-path routing matrix R:
 
0 0 1 0 0 0
 0 0 0 0 0 0 
 
 1 0 0 0 0 0 
 
 0 0 1 1 0 0 
R=  

 1 1 0 0 0 0 
 0 0 0 0 0 1 
 
 0 0 0 0 0 0 
0 0 0 0 1 0
Here we can see that for source to destination labeled 1, links 3 and 5 are utilized for routing (i.e. these link labels correspond
to links (1,2) and (2,3)). Source to destination labeled 2 utilizes only link 5, etc.
From the routing and external traffic we can find the traffic load demand for each link by Π = RΛ. Note that in this case
links 2 and 7 do not have traffic demand since links 1 and 3 are the one selected for the downlink and uplink traffic for node
1. Since they are carrying no traffic they can be removed from been scheduled for transmission. In fact, in our implementation
we redefine an equivalent link path routing matrix to account for the relative traffic load removing those links that are not
carrying traffic. This equivalent matrix is define and result for this example in:
 
0 0 1 0 0 0
 0.1 0 0 0 0 0 
 
 0 0 1 1 0 0 
Rmesh = R ∗ diag(Λλ(DL) =1 ) =   

 0.1 0.1 0 0 0 0 
 0 0 0 0 0 0.1 
0 0 0 0 1 0
Note that for this we redefined matrix, the mesh-link labels corresponding to it are:

Mesh link labels 1 2 3 4 5 6

Link(i,j) (2,1) (1,2) (3,2) (2,3) (4,3) (3,4)

STDMA Scheduling for mesh

To find the interference-based scheduling for constant transmission power and variable rate systems we follow a procedure
similar to the one described in [?]. Using the available rate sets and required SINR in table I we find the sets of cliques
containing links having the property that all links in the same clique can transmit simultaneously selecting one of the available
data rates.
T
The columns of S can be linearly combined to create the STDMA schedule by defining the vector of wights α = [α1 . . . αK ]
corresponding to the fraction of the time that each column vector of S is activated within a STDMA frame. Hence, for a given
α the allocated link capacity, c = [c1 . . . cL ]T is given by
K
X
c = Sα ; αk = 1. (2)
k=1
 TABLE III
G EOGRAPHICAL COORDINATES FOR I NTERNET G ATEWAYS

Node name Latitute Longitude Altitude

ENITEL (San Juan de Limay ) 13 10’23”N 08636’23” W 948.6ft
San Juan del Rio Coco 1333’00”N 086 10’ 00”W 2908.5ft

The capacity allocation can be done through a scheduling algorithm. The scheduling algorithm allocates slots and transmission
rates depending on the amount of traffic passing though each link and the objective function to be maximized. In order to
find the interference-based schedule for max-min fair allocation we utilized the column generation method [?], [?] for constant
transmission power and variable transmission rate as formulated in [?] but with the redefined mesh equations derived in this
paper. That is, we find the columns S that solve the following optimization problem:

maximize λmin
subject to λi ≥ λmin ∀i;
Rmesh Λmesh ≤ Sα ;
P
αk = 1 ;
k
λmin ≥ 0; Λmesh ≥ 0 ; 0 ≤ αk ≤ 1 .
Capacity evaluation
By solving the above problem with constant transmission power and variable rate using the nonlinear optimization method
the following link capacity allocation is found:
    
1 0 0 0 0 0 0.2273 0.2273
 0 1 0 0 0 0  0.0227   0.0227 
    
 0 0 1 0 0 0  0.4545   0.4545 
C = Sα =   0 0 0 1 0 0


=
 
 [Mbps]

  0.0455   0.0455 
 0 0 0 0 1 0  0.0227   0.0227 
0 0 0 0 0 1 0.2273 0.2273
The allocated end-to-end rates in matrix form can be written by:
 
0 0 0.0227 0
 0 0 0.0227 0 
Λ=  0.227 0.227
 [Mbps]

0 0.227
0 0 0.0227 0

In summary, by the transmission schedule, the resulting end-to-end uplink transmission rate is 22.7kbps and for the downlink
is 227kbps to each node. The total end-to-end network capacity is then 750kbps.
In the above example, if we reduce the distance between nodes to 10km and by increasing the interference margin to 5dB,
the topology is maintained. In that case, the transmission rates in the matrix S increases from 1 to 12Mbps. That result in a
throughput gain of 10 with respect to the previous case, i.e. the downlink to each user increases to 2.27Mbps and the uplink
to 227kbps.
III. U SER - DEPLOYED AND C APACITY E VALUATION FOR THE RURAL C OMMUNITIES IN N ICARAGUA
A. Description of User-deployed approach
We assume a simple user-deployed behavior. The community starts by locating their own nodes at their locations (telecenters)
and the shared Internet access node (Gateway) located at the closes land-based infrastructure. At the communities under
consideration there are two municipalities head-end (towns) where it is possible to have access to land-based Internet services:
San Juan de Limay and San Juan del Rio Coco. The geographical coordinates are shown in the table III-A. In San Juan de
Limay the operator ENITEL has a 63 meters height tower, and we assume that a similar tower could be utilized (or is available)
in San Juan del Rio Coco.
By utilizing these gateways we can group the communities into two potential user-deployed subnetworks as shown in Fig.4.
We call subnetwork 1 to the set of communities that are geographically closer to ENITEL in San Juan de Limay, and subnetwork
2 to the set of communities closer to San Juan del Rio Coco.
The user-deployed approach corresponding to each subnetwork is then applied. The stages for connecting the communities in
subnetwork 1 are illustrated in Fig.5 when utilizing 6dBi omnidirectional antennas (transmitter and receiver) and the parameters
in table I.
 Fig. 4. Subnetwork 1 and subnetwork 2

aa cc
bb

7 7
7 10

6 6
2 6 2
2
1 1
1
5 5
5 9
9

PSfrag replacements PSfrag replacements PSfrag replacements

4 4 4

Stage 1 8
Stage 1 8
Stage 1 8

Stage 2 3 Stage 2 3 Stage 2 3

Stage 3 Stage 3 Stage 3

Fig. 5. User-deployed stages for subnetwork 1. The gateway is indicated by a square (node 8), digipeaters by rhombus (nodes 9 and 10), and the nodes at
the communities are represented by a circle.

After stage 1 and its resulting connectivity, the users add a relaying node to connect node 5 (El Carrizo) and node 6 (El
Angel 1) to ENITEL(San Juan de Limay). For that, they select an intermediate (high) point for the relaying node. We refer
to relaying nodes that don’t generate external traffic as Digipeater (Digi). On the next stage (stage 3), node 7 (El Mojon) has
to be connected by also adding another digipeater. One high point in the region can be located that connect to node 5. The
positions for the digipeaters are summarized in the table IV.

TABLE IV
G EOGRAPHICAL COORDINATES FOR DIGIPEATERS IN SUBNETWORK 1

Node name Latitute Longitude Altitude

Digi-1(node 9) 1301’56”N 08638’23”W 3218.5ft
Digi-2(node 10) 1318’15”N 08640’24”W 3802.5ft
 Fig. 6. User-deployed (subnetwork 1)

B. Capacity evaluation for subnetwork 1

We apply the described method before together with the above user-deployed network and also utilizing beam-steering
antenna systems. The capacity evaluations are summarized in the following table:
TABLE V
C APACITY EVALUATION FOR SUBNETWORK 1(*)
Antenna Interference Antenna Uplink Downlink Total
System Margin [dB] Gain [dBi] (rate/node) (rate/node) rate
Omni 2 6 114.8kbps 1.15Mbps 8.84Mbps
Beam-steering 2 12 233.1kbps 2.33Mbps 17.95Mbps
(HPBW=30)
Beam-steering 5 12 516.7kbps 5.167Mbps 39.79Mbps
(HPBW=30)
* Communities and digipeaters: San Luis, La Fraternidad, San Lorenzo, La Grecia, El Carrizo, El Angel 1, El Mojon, ENITEL (San Juan
de Limay), Digi-1, Digi-2.

We can note that utilizing an interference margin of 5dB and Beam-steering antennas produces higher end-to-end data
rate. This is because when utilizing IM=2dB with beam-steering antennas, the network topology changes with respect to the
omnidirectional case making link (3,5) feasible. Link (3,5) is a weak link utilized for routing traffic towards the gateway (node
8) and to satisfy its traffic demand its activation consumes much network capacity.
In summary, by the simple user-deployed approach for the subnetwork 1, the (upper bound) allocated end-to-end downlink
transmission rate is up to 5.167Mbps per node (telecenter) and for the uplink is up to 516.7kbps per node.
C. User-deployed subnetwork 2
To connect the remaining communities we assume that an Internet gateway can be installed at the municipality head-end
called San Juan del Rio Coco. Next, following the user-deployed approach described before we add two digipeaters in the
locations summarized in the following table:
For the gateway at San Juan del Rio Coco we assume 63m tower height. As before, all nodes in the network are assumed to
be installed utilizing 20m antenna height. Figure 7 shows the locations of the communities and the locations of the digipeaters
needed for connectivity to the Internet gateway to be located in San Juan del Rio Coco.
 TABLE VI
L OCATIONS FOR DIGIPEATERS IN SUBNETWORK 2

Node name Latitute Longitude Altitude

Digi-3 1331’45.1”N 08609’27.4”W 3438.3ft
Digi-4 1334’59.6”N 08659’02.8”W 3676.2ft

Fig. 7. User-deployed (subnetwork 2)

when utilizing omnidirectional antennas with the selected points there is not fully connectivity for all nodes that composes
the network. The utilization of beam-steering antennas with an Interference margin of 2dB make possible to connect all the
communities to the Internet gateway. The increment on the interference margin is done in order to allowed more resistance
to interference and by this way we remove links with low transmission rate (we avoid them to be used for routing traffic and
to be scheduled). However, when increasing the interference margin to 5dB we found that the community called El Jobo gets
disconnected. Fig.8 shows the network topology when utiling beam-steering antennas with 2dB interference margin.
Note that there is a long range connection of 36.2km from El Ojoche (node 1) to San Miguel (node 4). This is because
there is line of sight between these two points as illustrated in Fig.9.

Pt=28dBm, Sensitivity = −88 dBm

3 9

7 6
8 2

Fig. 8. Network topology (subnetwork 2). Node 7 (San Juan del Rio Coco) is the Internet gateway and nodes 8 and 9 are the digipeaters.
 Fig. 9. Terrain profile between El Ojoche and San Miguel.

D. The capacity evaluations for subnetwork 2

We apply the described method before together with the above user-deployed network and also utilizing beam-steering
antenna systems. The capacity evaluations are summarized in table VII.
TABLE VII
C APACITY EVALUATION FOR SUBNETWORK 2(**)
Antenna Interference Antenna Uplink Downlink Total
System Margin [dB] Gain [dBi] (rate/node) (rate/node) rate
Omni 2 6 - - -
Beam-steering 2 12 388.5kbps 3.9Mbps 25.6Mbps
(HPBW=30)
Beam-steering 5 12 - - -
(HPBW=30)
** Communities and digipeaters: : El Ojoche, San Marcanda, El Varrillal, San Miguel, Cristo Rey, El Jobo , San Juan del Rio Coco, Digi-3,
Digi-4.

Hence, for this user-deployed subnetwork, the resulting (upper bound) end-to-end downlink transmission rate is up to about
3.9Mbps per node (community) and 388.5kbps per node for the uplink.
IV. C ONCLUDING R EMARKS
In this paper we presented a methodology to utilize free-available tools that can be used for community-planning of mesh
networking. The evaluation results shown by an example of planning and capacity evaluation in rural communities of Nicaragua
demonstrate that mesh networking is an appealing technology to allow that the community themselves or even new potential
players take active part in the solution to their own needs. This work also shows the need to develop easy-to-use free-software
tools coupled with mesh network products to make the community deployment of mesh networks ubiquitous.
In this work we have introduced a novel methodology for user-deployed Multihop Wireless Networks in rural areas with
low density users, which allow the perform analysis and design of these kind networks. MWN are an important alternative for
fixed wireless broadband access to cover remotely located places in rough terrain.
For easy analysis we divide the whole network (13 communities) into two subnetworks, so we introduce two gateways
(GW1 and GW2). Because not all nodes have a direct link with theirs gateways or with the other nodes, then some nodes
need to relay data through a digipeater in order to be avoid not being disconnected or to be part of the MWN. So, the target
coverage area or service area can be completed by means of digipeaters, if this is needed.
From this study we can derive the following conclusions: It is possible design rural area networks with 2.4GHz ISM
technology utilizing Multihop Wireless Ad hoc networks with reasonable cost infrastructure. In short to do that, we proceed
as follows: a) we utilized a real community sites in the rural areas of the northern central region of Nicaragua with complex
geographical environments, b) we derive the G matrix utilizing a free share program for radio propagation entitled Radio Movil
[?], c) We assume a very simple routing based on the Dijkstra’s Algorithm, d) We reckon all the traffic load matrices involved
(see the methodology), and e) Finally, we compute the (upper bound) capacity of the networks by convex optimization [?].
The network of 13 nodes is split in two parts, namely: subnetworks 1 with 7 nodes (San Luis, La Fraternidad, San Lorenzo,
La Grecia, El Carrizo, El Angel1, y El Mojon) with an Internet gateway at the head town of the municipality San Juan de
Limay; and subnetwork 2 with 6 nodes (El Ojoche, San Marcanda, El Varillal, San Miguel, Cristo Rey, El Jobo) and we
consider as internet gateway (taking into account the national infrastructure available),San Juan del Rio Coco and ENITEL in
San Juan de Limay.
Assuming an ADSL connection to Internet in the main town with an asymmetrical traffic for the uplink (user to the network
headend) we are able to estimate these capacities (in terms of throughputs): Min(Uplink/user), Min(Downlink/user), Total
end-to-end rate, and Total max-min end-to-end rate.
To study the systems we have considered three possibilities: a) omnidirectional antennas with gain of 6 dBI, b) beam-steering
antenna systems with half power beamwidth de 30 degrees and antenna gain of 12 dB with interference margin (IM) IM=2
dB c) beam steering with half power beamwidth de 30 degrees with antenna gain of 12 dB with IM =5.
The data rates for the air interface (in the downlink) are usually higher than the typical connection rate given by the current
DSL service, and they are the followings: 1.15 Mbps per nodes (using omni-directional antennas with 6 dBi of antenna gain),
and in the range of 2.33-3.9 Mbps per node with beam steering antennas with 12 dB of gain (IM=2dB).