Professional Documents
Culture Documents
10 1 1 14 2757
10 1 1 14 2757
of IEEE 2003 Global Communications Conference (Globecom), San Francisco, USA, Dec 2003
Optical Fiber
Abstract— Wireless networks using radio over fiber (RoF)
BS BS BS
technology operating in millimeter-wave bands have been sug- CS
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)
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
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.
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.
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)
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.