To appear in Proc.

of IEEE 2003 Global Communications Conference (Globecom), San Francisco, USA, Dec 2003

Performance Evaluation of a MAC Protocol for

Radio over Fiber Wireless LAN operating in the
60-GHz band
Hong Bong Kim, Adam Wolisz
Telecommunication Networks Group
Department of Electrical and Computer Engineering
Technical University of Berlin
Sekr FT5-2 Einsteinufer 25
10587 Berlin Germany
Email: hbkim, wolisz
@ee.tu-berlin.de

Optical Fiber
Abstract— Wireless networks using radio over fiber (RoF)
BS BS BS
technology operating in millimeter-wave bands have been sug- CS

gested as promising solutions to meet increasing user bandwidth

and mobility demands. A system based on this technology has Picocell 1 Picocell 2 Picocell 3
properties quite different from those of conventional WLAN (a)
BWtotal
systems in that every room in a building hosts at least one BW ch BWg

picocell having its own base station (BS). Thus a challenging f1 ... fm fm+1 ... f2m
problem lies in the medium access control (MAC) protocol design t1 t2 t3 t4 t5 t 6 t 7 t8 t9 t10
f1 (b)
so that it can support QoS requirements as well as a fast and f2
Down-link f3 ...
easy handover. A MAC protocol (Chess Board Protocol) based f4
on frequency switching (FS) codes has been proposed by the f5
(c)
authors considering the situation [7]. In this paper performance t1 t2 t3 t4 t5 t 6 t7 t8 t9 t10
evaluation results for simple six variants of it are described and fm+1
fm+2
discussed. Up-link fm+3 ...
fm+4
fm+5
I. I NTRODUCTION frame (t f ) slot (t s )
(d)

In order to meet modern ever increasing user bandwidth and

mobility demands, a wireless network based on radio over Fig. 1. System description. (a) A RoF WLAN system operating in the
millimeter-wave band, (b) The total system bandwidth is subdivided into 
fiber (RoF) technology operating in millimeter-wave (mm- channels, where  

 ,  and  are the total system bandwidth,
wave) bands has been suggested as a promising solution. In the channel bandwidth and the guard bandwidth, respectively, (c) and (d) show
this network mm-wave signal is transmitted over optical fiber frequency switching (FS) patterns for down- and up-link when the number of
channels is five.
between the control station (CS) and base stations (BSs), and
the BSs serve as access points for mobile hosts (MHs). Since
the quite different properties of mm-wave from those in the
mechanism allows a MH to stay tuned to its frequency during
usual wireless LAN (WLAN) bands (2.4 or 5 GHz) the size
handover, which is a major characteristic of the MAC protocol.
of the picocells is limited to at most a room in an indoor
In this paper simulation study for simple six variants of the
environment, thus leading to a very large number of BSs
MAC protocol is described. The paper starts with a brief de-
and frequent handovers of MHs between picocells. Therefore
scription of the Chess Board protocol. In section III simulation
simple and cost-effective BS will be a key to the success of the
results are shown and discussed. Section IV summarizes the
system. Receltly much of the research in this field has been
work.
focused on such components operating in mm-wave bands [2]–
[6]. Of mm-wave bands especially the 60 GHz band is of II. C HESS B OARD P ROTOCOL D ESCRIPTION
much interest since a massive amount of license-free spectrum
has been allocated with a worldwide overlab of 3 GHz (59– A. Basic Operations
62 GHz) [1]. A brief description of the Chess Board protocol is given in
A MAC protocol (Chess Board protocol) for the system this section, please refer to [7] for more detail. The simple
featuring fast and easy handover and QoS support has been structure of the BS and 60-GHz wave characteristics leads to
proposed by the authors [7], which is based on frequency a centralized network architecture with many picocells, where
switching (FS) codes. Adjacent picocells employ orthogonal most of the BS functions of conventional WLANs are shifted
FS codes to avoid possible co-channel interference. This to the CS (Fig. 1 (a)).

By subdividing the total system bandwidth,  (fre-
quency) channels are obtained (Fig. 1 (b)), where  channels
!!"#%$
are&used for down-link transmission and the
#(')#('*"!"#%$
other  channels for up-link transmis-
+,#('*+$
sion. A pair of frequency channels are used together
to support
 +  #('*+ $
frequency division
 !"!1
duplex (FDD) operation, where
(-/. 0   ) are for down- and up-link
transmission, respectively. In addition, the time axis is also
subdivided into time slots of equal length and  time slots
are grouped into a frame.
Every BS supports all channels, each of them, however,
being supported in each proper time slot. Fig. 1 (c) and (d)
show FS patterns for down- and up-link, respectively, when
 is five. During every frame time, each of the  time slots
and  channels is utilized once and only once. Adjacent
picocells must not use identical FS patterns to avoid possible
co-channel interference. And one FS pattern can be used in
several picocells if they are spaced far enough that there is no
co-channel interference between them. For proper operations
using FS patterns, we assume that the system is synchronous.
Fig. 2 shows down- and up-link slot formats and essentially,
they have the same format. The down-link slot begins with a
MAC address indicating the destination of the slot. This MAC
address is followed by a permission field, which authorizes
transmission of the MH specified in the MAC Address in the
following up-link slot. The next field is for down-link payload,
destined to the MH specified by the MAC address. The last
field consists of another MAC address and reservation result
which indicate if the request for bandwidth from the addressed
MH is successfully confirmed or not.
The up-link slot consists of two parts. The first one is used
for uplink data transmission and it consists of MH’s MAC
address, piggyback field, and payload field. The second part
is for reservation and it must not be used by the MH currently
using the the first part but may be used by any other MHs
having data to send, and not being assigned bandwidth in this
channel so far. If the piggyback field is set it means that MH
still has more data to send and requests assignment of one
more slot, and non-setting this field indicates that its transmit
buffer is empty. Each payload field (both up-link and down-
link) allows packing several small packets or fragmentation
for a large size packet.
To request a permit for up-link transmission on some chan-
nel, the MH sends a request to the CS, using the reservation
field in any slot in this channel (Fig. 2).
B. Mobility Support
As long as a MH remains tuned to a certain channel
(frequency pair) and time slot pair, i.e., listening on the down-
Next
ch
Permit
Piggyback
MAC
DOWN addr.
P Data Reserv.
UP
MAC
P Data Reserv.
addr.
time
guard
time
Packet ... Packet
Fig. 2. Slot formats for the down- and up-link data transmission.
another picocell. Furthermore, using the following slot for the
uplink transmission the MH can and should make a request
for transmit permission/reservation in the new pair of time
slots. Note that the CS knows that the MH with this MAC
address has been transmitting previously in a different picocell,
so it is possible a) to treat such request with higher priority
than requests for new reservations and b) release any possible
reservation in the old picocell.
C. Six Variants of Chess Board Protocol
Six variants of the Chess Board protocol, classified into two
groups, are considered in the paper as shown in Table I. In
the first group (group A), MHs are assumed not to have a
capability to change channels during operation, whereas MHs
in group B are assumed to be able to change channels.
MAC A1 is the simplest in which the CS allows a single
channel to be used by only one MH, and it has no queue for
requests. When the channel is being used, further requests are
blocked. In particular, if a MH moves into another picocell in
which “its” channel is already used, there is no way to continue
transmission. MAC A2 is similar with the only difference that
the CS does enqueue requests for “busy” channels rather than
rejecting the request, and when the channel is released the CS
assigns it to the MH at the head of the queue. In MAC A3
interleaved usage of slots in a single channel by several MHs
is performed. Slots are assigned using a simple round-robin
fashion thereby assuring equal portion of the capacity to each
candidate MH.
Essentially group B MAC protocols are similar to those
in group A, with the difference, that in the case of group
B MAC protocols the CS attempts to find capacity in other
channels, not only in the channel on which the request has
been issued. For instance in MAC B1 when a request from a
MH arrives at the BS, the CS investigates each channel if it is
reserved or not. If a channel is free it assigns the channel to
the MH by sending the channel number. When all channels are
reserved the request is rejected. A queue for requesting MHs
link channel, the MH knows precisely3245$
that a slot from the CS is maintained in MAC B2, and as soon as a channel is released
should appear every frame time so that it can expect it is allocated to the MH at the head of the queue. MAC B3
the time instances at which down-link slots arrive. However, is similar to B2 with an exception that when the number of
when it moves into an adjacent picocell using a different requesting MHs is greater than the number of channels, all the
FS pattern, the MH will receive the down-link slot in an channels are shared by the MHs in a round-robin fashion. In
unexpected time instance. Thus the MH 2 4
can easily realize – this case whenever the CS transmits a permit to a MH it must
within at most a single frame time ( ) – that it moved into also inform the MH of the next channel number in the permit
 10 m 10 m TABLE I
S IX VARIANTS OF THE C HESS B OARD P ROTOCOL FOR R O F WLAN
O PERATING IN 60 GH Z B AND
10 m

Channel change Request when the

MAC Group
during operation channel is used
A1 blocked
10 m

BS
A2 NO queued A
MH A3 shared
assigned if any,
B1
blocked if not
assigned if any,
Fig. 3. Indoor environment for simulation. B2 YES B
queued if not
assigned if any,
B3
shared if not
field of downlink slot (Fig. 2). It can be expected that group
B MAC protocols will outperform group A MAC protocols at
TABLE II
the expense of increased complexity.
S UMMARY OF THE S IMULATION PARAMETERS
III. N UMERICAL R ESULTS
Picocells 4 Guard time 32 bytes
A. Simulation Scenario and Assumptions
MHs 40 FCS 4 bytes
In the simulation study emphasis is placed on the perfor- BS Capacity Buffer
mance comparison of six variants of the Chess Board protocol 155 Mbps 10 Mbytes
(LK;M? ) Length
and the effects of two parameter (the number of channels and MAC adrs 6 bytes Mobility Model Random Waypoint
slot length) on system performance. Because of its centralized Permit field 1 byte Speed of MH 1.5 m/s
nature of the system downlink transmission is so simple that Reservation slot 16 bytes Simulated Time 3000 sec
only uplink performance is considered. Statistics
The indoor environments for simulation includes four pico- Flag 8 bytes after 300 sec
Collection
cells and 40 MHs (Fig. 3). MHs are assumed to be freely mov-
ing across boundaries between picocells according to random
waypoint mobility model. Some assumptions for simulations
are as follows: B. Delay Performance
6 After having sent a request a MH receives the reservation
The mean packet delay is the average time spent by a packet
result in one frame time from the down-link channel. from the instant it is generated till its transmission is complete.
6 The capacity of a BS ( 78*9;: ) is 155 Mbps. The channel
Fig. 4 and 5 show the mean packet delay of group A and B
data rate (7<8*=?> ) is equal to 7<8@9A:CBDFEHG , where  MAC protocols when the number of channels is five and the
is the number of channels and G is fixed size overhead slot size is 1000 bytes, respectively. In Fig. 4 as the traffic
including MAC address, permit field, reservation field and load increases, A2 and A3 perform better than A1. Similarly
guard time. in Fig. 5 B2 and B3 are better than B1 as the traffic load
6 Message traffic consists of three packets: 41 bytes (45%),
grows. We can also see that group B MAC protocols highly
576 bytes (35%), and 1500 bytes (20%) [8]. ourperform group A MAC protocols because of the group B’s
6 Interarrival time between messages is exponentially dis-
trunking ability. Since the similar trends are observed with
tributed. different number of channels and slot size, from now on we
6 Channel is perfect (error-free).
will only consider A3 and B3 protocols.
6JI -persistent algorithm is used for sending a request
The impact of the slot size in MAC A3 is shown in Fig. 6
packet. indicating smaller slot size is better than larger one. However
6 MHs use harmonic backoff algorithm (attempting prob-
"! due to the fixed size overhead too small size may cause
abilities 1, 1/2, 1/3, ) in computing their probabilities larger delay under heavy traffic load. In order to investigate
to transmit a request. the effects of the number of channels on delay performance,
6 Initial channel distribution by MHs is uniform.
simulations were run with a fixed-size slot and various number
Other assumptions and parameters for simulation are sum- of channels. Fig. 7 shows that smaller number of channels
marized in Table II. is beneficial in terms of the mean packet delay, which is
Each simulation was run for 3000 sec (simulated time) because the frame
N2 4
time2SR1is$ directly proportional to the number
including warm-up phase of 300 sec. 95 % confidence intervals of channels .OQP .
were calculated for the mean packet delay and normalized An interesting fact is observed when the traffic load is 3
throughput whose variations from the sample mean were less Mbps. When  is 19, each of the 17 channels is shared by
than 2% for all results. two MHs and the other two channels are shared by three MHs
 Mean packet delay of A3 Mean packet delay of A3
12 15
A1 1 Mbps
A2 14 2 Mbps
A3 3 Mbps
13
10 4 Mbps
12

Mean packet delay (msec)

Mean packet delay (msec)

11

8 10
9
8
6
A3 7
4 Mbps
A2 6
5 3 Mbps
4
A1 4 2 Mbps
3 1 Mbps
2
2
1

0 0
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Traffic load per MH (Mbps) Slot size (bytes)

Fig. 4. The mean packet delay of group A MACs when  is five and slot Fig. 6. The mean packet delay of A3 when the number of channels are five
size is 1000 bytes. and the slot size is 200 – 5000 bytes.

Mean packet delay of A3

Mean packet delay of B3
5 1 Mbps
B1 2 Mbps
B2 3 Mbps
4.5 B3 4 Mbps
2
10

Mean packet delay (msec)

4
Mean packet delay (msec)

3.5
4 Mbps
3

B1 3 Mbps
2.5 1
10 2 Mbps
B2
2 1 Mbps

1.5

1
B3 0
10
0.5

0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 Number of channels
Traffic load per MH (Mbps)

Fig. 7. The mean packet delay of A3 when the slot size is 1000 bytes and
Fig. 5. The mean packet delay of group B MACs when  is five and slot the number of channels is 5 – 25.
size is 1000 bytes.

in rare collision and minimum reservation delay. If  is over

respectively since channel distribution of MHs is assumed ten the system has more channels than MHs on the average,
uniform and total number of MHs is 40. If the three MHs using thereby increasing frame time and wasting of bandwidth.
the same channel come together in the same picocell their total
traffic load (9 Mbps) becomes greater than the channel date C. Throughput Performance
rate (8.1 Mbps T 155/19), thus resulting in high mean packet Throughput (normalized) is defined in the paper as a fraction
delay. Whereas when  is 20 each channel is shared by only of time during which user data is transmitted. Fig. 10 and 11
two MHs. Even if two MHs using the same channel stay in the show throughputs of MAC A3 and B3, respectively. Since
same picocell their total traffic load (6 Mbps) is smaller than the channel data rate is determined by the number of channels
the channel data rate (7.75 Mbps T 155/20). That is the reason smaller number of channels performs better than larger number
why the mean packet delay is lower when  is 20 than when of channels. Furthermore, larger slot size is favorable in terms
 is 19. Therefore, we can see that the mean packet delay of the throughput performance. From the point the line is no
in MAC A3 depends not only on the number of channels and longer linear packet loss begins to occur. It can be seen that
slot size but also on the channel distribution of MHs. MAC B3 highly outperforms MAC A3.
The impact of slot size of MAC B3 is shown in Fig. 8. Just
as in A3 MAC protocol smaller slot size is advantageous in IV. C ONCLUSION
delay performance but because of fixed-size overhead small The wireless LAN environment using radio over fiber tech-
slot is not always beneficial. Fig. 9 shows how the number of nology operating in the millimeter-wave band imposes quite
channels influences the delay performance. With light traffic different requirements on the system design as compared to
delay grows along with the number of channels since frame the conventional WLANs. Since the high penetration loss of
time is proportional to it. As traffic load increases the delay has millimeter-wave signal many BSs should be employed to cover
a minimum when  is around ten. Note that in our simulation indoor areas. In such network with high number of small
scenario the average number of MHs in a picocell is ten. So cells, the issue of mobility management has a very special
when  is 10 each MH is assigned its own channel resulting significance. A MAC protocol, called Chess Board protocol,
 Throughput of B3
1
s = 1000
s = 3000
0.9 s = 5000 ch = 5
Mean packet delay of B3
15
0.8
14 1 Mbps

Normalized throughput
3 Mbps
13 0.7
5 Mbps
12 7 Mbps
Mean packet delay (msec)

0.6
11
10 0.5
9 ch = 15
0.4
8
7 7 Mbps 0.3
6
5 Mbps 0.2
5
3 Mbps
4 0.1
1 Mbps
3
0
2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1
Traffic load per MH (Mbps)
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Slot size (bytes) Fig. 11. Throughput of B3 MAC protocol.

Fig. 8. Mean packet delay of B3 when the number of channels are the slot
size is 200 – 5000 bytes.
featuring fast and easy handover and QoS support has been
proposed by the authors [7]. In this paper six variants of
the protocol were considered and their performance has been
evaluated by simulation study.
Simulation results have shown that group B MAC protocols
Mean packet delay of B3
which are assumed to have a capacity to change channels
1 Mbps
3 Mbps
5 Mbps
during operation highly outperform group A MAC protocols
7 Mbps
for which a fixed channel is assumed for each MH. Delay
Mean packet delay (msec)

performance of both of them depends on the slot size and the

1
10
number of channels; moreover, in group A MAC protocols
it relies also on channel distribution of MHs. Smaller slot
7 Mbps
5 Mbps
size and smaller number of channels are beneficial in delay
3 Mbps performance. On the other hand larger slot size and smaller
1 Mbps
number of channels are in favor of throughput performance.
10
0
To exploit bandwidth resources effectively, MH’s capability
to change channels during operation is required in a highly
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Number of channels
densed mobile environment.
R EFERENCES
Fig. 9. Mean packet delay of B3 when the slot size is 1000 bytes and the
number of channels is 5 – 25. [1] P. Smulders, “Exploiting the 60 GHz Band for Local Wireless Multimedia
Access: Prospects and Future Directions,” IEEE Commun. Mag. pp.140-
147, Jan. 2002.
[2] K. Kitayama, et. al, “An Approach to Single Optical Component Antenna
Base Stations for Broad-Band Millimeter-Wave Fiber-Radio Access Sys-
tems,” IEEE Trans. Microwave Theory Tech., vol. 48, pp.2588-2594, Dec.
2000.
Throughput of A3
1 [3] T. Kuri, et. al, “60-GHz-Band Full-Duplex Radio-On-Fiber System Us-
s = 1000
s = 3000 ing Two-RF-Port Electroabsorption Transceiver,” IEEE Photon. Technol.
0.9
s = 5000 ch = 5
Lett., vol.12, no.4, pp.419-421, Apr. 2000.
0.8 [4] K. Kitayama, K. Ikeda, T. Kuri, A. Stöhr, and Y. Takahashi, “Full-duplex
Normalized throughput

0.7
demonstration of single electroabsorption transceiver basestation for mm-
wave fiber-radio systems,” Microwave Photon. MWP 2001, pp. 73-76,
0.6
2001.
0.5 [5] G. Grosskopf, D. Rohde, and R. Eggemann, “155 Mbit/s Data Trans-
mission at 62 GHz Using an Optically Steered Antenna,” ECOC 2000,
0.4
Muenchen, Germany, Sept. 2000, vol. 3, pp. 53-54.
0.3 ch = 15 [6] R.-P. Braun, G. Grosskopf, D. Rohde, and F. Schmidt, “Low-Phase-Noise
0.2
Millimeter-Wave Generation at 64 GHz and Data Transmission Using
Optical Sideband Injection Locking,” IEEE Photon. Technol. Lett., vol.
0.1 10, no. 5, pp. 728-730, May 1998.
0
[7] H. B. Kim, H. Woesner and A. Wolisz, “A Medium Access Control
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Traffic load per MH (Mbps) Protocol for Radio over Fiber Wireless LAN operating in the 60-GHz
Band,” in Proc. 5th European Personal Mobile Commun. Conf., Apr.
2003. http://www-tkn.ee.tu-berlin.de/publications/proc.html.
Fig. 10. Throughput of A3 MAC protocol.
[8] K. Thompson, G. J. Miller and R. Wilder, “Wide-Area Internet Traffic
Patterns and Characteristics,” IEEE Network, pp.10-23, Nov. 1997.