A New Multi-Channel MAC Protocol With On-Demand Channel Assignment For Multi-Hop Mobile Ad Hoc Networks
1. 1
A New Multi-Channel MAC Protocol with On-
Demand Channel Assignment for Mobile Ad Hoc
Networks
S.-L. Wu, C.-Y. Lin, Y.-C. Tseng, and J.-P. Sheu
Int'l Symposium on Parallel Architectures, Algorithms and
Networks (I-SPAN), 2000
2
Abstract
Ė considers the access of multiple
channels in a MANET
Ė Features
Ė âon-demandâ style to assign channels
Ė Degree-independent
Ė It adapts to exchange few control
messages to achieve channel assignment
Ė No form of clock synchronization
2. 2
3
Outline
Ė Introduction
Ė Concerns with Using Multiple Channels
Ė Our Multi-Channel MAC Protocol
Ė Analysis and Simulation Results
Ė Conclusions
4
Introduction(1)
Ė Single-channel MAC protocols
Ė A common channel is shared by all mobile
hosts
Ė The standard of IEEE 802.11 has been
widely accepted
Ė The performance will degrade quickly as
the number of mobile hosts increases
3. 3
5
Introduction(2)
Ė Multi-channel MAC protocol
Ė Several advantages
Ė Higher throughput
Ė Less normalized propagation delay per channel
Ė It is easier to support QoS
6
Introduction(3)
Ė how to access the channel is
technology-dependent
Ė The categorization of mobile host
Ė Single-transceiver
Ė A mobile host can only access one channel at a
time
Ė Multi-transceiver
Ė A mobile host can access multiple channels
simultaneously
4. 4
7
Introduction(4)
Ė Related work
Ė Channel assignment in a traditional packet radio
network
Ė Two IEEE 802.11-like protocols separate control
traffic and data traffic into two distinct channels
Ė A scheme based on Latin square assumes a
TDMA-over-FDMA technology
Ė Channel assignment is static
Ė A clock synchronization is necessary
Ė A number of transceivers is equal to the number of
frequency bands
8
Introduction(5)
Ė Related work(cont.)
Ė The protocol [12] for Reconfigurable
wireless network assigns channels statically
and has a polling transceiver and a sending
transceiver
Ė The protocol [10] assigns channels
dynamically and mandates tow-hop
information
5. 5
9
Introduction(6)
Ė Related work(cont.)
Ė The protocol [24] is degree-independent
and each host has one transmitter and n
receivers concurrently listening on all n
channels
Ė A hop-reservation MAC protocol is also
degree-independent and requires clock
synchronization
10
Introduction(7)
Ė Our new MAC protocol
Ė Applies to FDMA and CDMA
Ė Requires two simplex transceivers per host
Ė Based on RTS/CTS-like reservation
mechanism and degree-independent
Ė The clock synchronization is un-necessary
Ė Dynamic assigns channels and in an âon-
demandâ fashion
6. 6
11
Introduction(8)
12
Concerns with Using Multiple Channels(1)
Ė SM:A Simple Multi-channel Protocol
Ė Static channel assignment
Ė The transmission follows IEEE 802.11
Ė The basic idea
Ė When a host X needs to send to a host Y,
X should tune to Yâs channel
Ė Two states
Ė RECEIVE
Ė SEND
7. 7
13
Concerns with Using Multiple Channels(2)
Ė Some Observations
Ė The traditional problem in a single-channel
system
Ė Hidden-terminal:(a)
Ė Exposed-terminal:(b)
14
Concerns with Using Multiple Channels(3)
Ė In SM protocol
Ė The hidden-terminal problem will become
more serious
Ė the exposed-terminal problem will become
less serious
Ė Some new problems may appear
Ė False Connectivity Detection
Ė Channel Deadlock Problem
10. 10
19
Our Multi-Channel MAC
Protocol(1)
Ė DCA(dynamic channel assignment)
Ė Our channel model
20
Our Multi-Channel MAC
Protocol(2)
Ė Each mobile host is equipped with tow
half-duplex transceivers
Ė Control transceiver
Ė Operates on the control channel to exchange
control packets with other mobile hosts
Ė Data transceiver
Ė Dynamically switches to one of the data
channels to transmit the data packet and ACK
11. 11
21
Our Multi-Channel MAC
Protocol(3)
Ė Each mobile host,say X,maintains the
following data structure
Ė CUL[]:channel usage list
Ė CUL[i] has three fields
Ė CUL[i].host
Ė CUL[i].ch
Ė CUL[i].rel_time
22
Our Multi-Channel MAC
Protocol(4)
Ė FCL:free channel list
Ė It is dynamically computed from CUL
Ė The complete protocol
Ė A wants to send to B,it first checks whether
the following two conditions are true:
Ė B is not equal to any CUL[i].host such that
CUL[i].rel_time>Tcurr+(TDIFS+ TRTS+TSIFS+TC
TS)
12. 12
23
Our Multi-Channel MAC
Protocol(5)
Ė There is at least a channel Dj such that for all i:
(CUL[i].ch=Dj) (CUL[i].rel_timeâĶ
Tcurr+(TDIFS+TRTS+TSIFS+TCTS))
24
Our Multi-Channel MAC
Protocol(6)
13. 13
25
Our Multi-Channel MAC
Protocol(7)
Ė Then A can send RTS(FCL,Ld) to B only if there is
no carrier on the control channel
Ė On a host B receiving the RTS(FCL,Ld) from A, it
has to check whether there is any
data channel Dj FCL such that for all i:
(CUL[i].ch=Dj) (CUL[i].rel_timeâĶTcurr + (TSIFS +
TCTS))
Ė If so,then B replies a CTS(Dj,NAVCTS) to A, where
NAVCTS = Ld/Bd+TACK+2Ï
Ė Otherwise,B replies a CTS(Test) to A
Ė Test=min{ ,CUL[i].rel_time} - Tcurr - TSIFS- TCTS
26
Our Multi-Channel MAC
Protocol(8)
Ė On an irrelevant host Câ B receiving
A's RTS(FCL,Ld), it has to inhibit itself
from using the control channel for a
period
NAVRTS = 2TSIFS+TCTS+TRES+2Ï
Ė Host A, after sending its RTS, will wait
for B's CTS with a timeout period of
TSIFS+TCTS+2Ï
14. 14
27
Our Multi-Channel MAC
Protocol(9)
Ė On host A receiving B's
CTS(Dj,NAVCTS) , it performs the
following steps:
Ė Append an entry CUL[k] to its CUL such
that
Ė CUL[k].host = B
Ė CUL[k].ch = Dj
Ė CUL[k].rel_time = Tcurr +NAVCTS
28
Our Multi-Channel MAC
Protocol(10)
Ė Broadcast RES(Dj,NAVRES) on the control
channel, where
Ė NAVRES=NAVCTS-TSIFS-TRES
Ė Send its DATA packet to B on the data
channel Dj. Note that this steps happens
in concurrent with step b)
Ė On an irrelevant host Câ A receiving
B's CTS(Dj,NAVCTS), C updates its CUL.
15. 15
29
Our Multi-Channel MAC
Protocol(11)
Ė On a host C receiving RES(Dj,NAVRES),
it appends an entry CUL[k] to its CUL
such that:
Ė CUL[k].host = A
Ė CUL[k].ch = Dj
Ė CUL[k].re_time = Tcurr +NAVRES
Ė On B completely receiving A's data
packet, B replies an ACK on Dj
30
Our Multi-Channel MAC
Protocol(12)
Ė To summarize our protocol
Ė Relies on the control channel to assign
data channels
Ė the deadlock problem can be avoided
Ė the missing RTS/CTS and the hidden-terminal
problems will be less serious.
16. 16
31
Analysis and Simulation
Results(1)
Ė Arrangement of Control and Data
Channels
Ė For simplicity, let's assume that the lengths
of all control packets(RTS, CTS, and RES)
are Lc,and those of all data packets Ld=9Lc.
32
Analysis and Simulation
Results(2)
Ė In this paper,we consider two bandwidth
models
Ė Fixed-channel-bandwidth:
Ė Each channel has a fixed bandwidth.
Ė Fixed-total-bandwidth:
Ė The total bandwidth offered to the network is
fixed.
17. 17
33
Analysis and Simulation
Results(3)
Ė Fixed-channel-bandwidth
Ė the maximum number of data channels
Ė n âĶLd/3* Lc (1)
Ė the utilization U of the total given
bandwidth.
Ė UâĶn/(n+1) (2)
Ė From (1) and (2),we derive that
Ė UâĶ Ld/3* Lc +Ld (3)
34
Analysis and Simulation
Results(4)
Ė Fixed-total-bandwidth
Ė The problems are:
Ė (i) how to assign the bandwidth to the control
and data channels
Ė (ii) how many data channels (n) are needed, to
achieve the best utilization.
Ė the maximum number of data channels
Ė n âĶ(Ld/Bd) / (3* Lc/Bc) (4)
18. 18
35
Analysis and Simulation
Results(5)
Ė the utilization U of the total given
bandwidth.
Ė UâĶ(n* Bd) / (n* Bd+Bc) (5)
Ė From (4) and (5),we derive that
Ė UâĶ Ld/3* Lc +Ld (6)
Ė Ld/(3* Lc +Ld)=(n* Bd) / (n* Bd+Bc)
Bc/(n* Bd)=3* Lc /Ld (7)
Ė Thus,the best ratio is
Ė 3* Lc /Ld
36
Analysis and Simulation
Results(6)
Ė Experimental Results
Ė Tow performance metrics
Ė Throughput =(Packet_Length*No_Successful_
Packets)/Total_Time
Ė Utilization =(Packet_Length*No_Successful_
Packets)/Total_Time*No_Channels
Ė Present results from 4 aspects
Ė Effect of the Number of Channels
20. 20
39
Analysis and Simulation
Results(9)
40
Analysis and Simulation
Results(10)
Ė Effect of Data Packet Length
Ė Fig11(a)(b)(c)(d) and Fig(12)
Ė Effect of the Bandwidth of the Control
Channel
Ė Fig(13)
Ė Effect of Host Density
Ė Fig(14)
24. 24
47
Analysis and Simulation
Results(17)
48
Conclusion
Ė The result for the fixed-channel-
bandwidth model is particularly
interesting for the currently favorable
CDMA technology
Ė We are currently working on extending
our access mechanism to a reservation
one