Chpater 3 Proposed MAC Protocol
3.6 Contention resolution
When nodes want to transmit data, they have a probability Pr to send RTS. Pr is propotion to maximum allowable power (LPRTS).
2) (1 A *
= L
Pr PRTS r (19)
LPRTS is the maximum allowable power for a node to transmit RTS without disrupting nearby transmissions and A is a constant to optimize the probability Pr. r is the number of retransmission value, and its initial value is zero. When retransmission times increase, Pr decrease. When collisions happen, nodes retransmit RTS with probality Pr. For example, as shown in Fig. 3-11. , node A is transmitting data to node B, and node C is transmitting data to node D. Node G and node E wants to transmit RTS. Since the LPRTS of node E is larger than the LPRTS of node G, The probability Pr of node E is also larger than probability Pr of node G. Therefore node E has a large probability to transmit RTS message than node G.
Figure 3 - 11 Example of nodes use probability Pr to transmit RTS
28
Now we define the probability Pc to avoid collisons from relaying CTS messages.
Probability Pc direct proportion to the number of neighbors that can are with the transmission range of power Pt .
2) (1 B*
= N
Pc r (20)
where N is the number of neighbor that can reach by power Pt, and B is a constant to optimize the probability Pc. r is the number of retransmission times, and its initial value is zero. When collisions happen, nodes relay CTS with probability Pc. For example, as shown in Fig. 3-12. , node A is transmitting data to B and node C is transmitting data to node D, when node G, F and H overhear node B’s CTS, they need to relay CTS to their neighbors. Since F’s N=3(E,I,J) G’s N=2(K,J) H’s N=1(M), thus node F has a larger probability among all to relay CTS than node G and H.
Figure 3 - 12 An illustrative example of proposed contention resolution
29
3.7 Performance analysis
In this section, we introduce the model analysis with our protocol. We will describe the preliminaries of our protocol, and then we use the preliminaries to calculate the throughputs of our protocol.
A. Preliminaries
We assume that RTS, CTS, DATA, and ACK packets are with fixed lengths of LRTS , LCTS, Ldata and LACK. R is the data rate. Tr, Tc, Tdata, and TACK are transmission times of RTS, CTS, DATA, and ACK packets. We can calculate them by equation (21), (22), (23), and (24).
R
Tr = LRTS (21)
R
Tc = LCTS (22)
R
Tdata = Ldata (23)
R
TACK = LACK (24)
Let τ be the probability that a node will transmit data in a time period
τ=α*β (25)
α is a probability of the packet arrival probability, and we assume the process follow a poisson distribution. β is a probability that a node will be selected to transmit data, and it depends on the simulation environment.
30
In the equation 26, where λ is the frame arrival rate and T is the expected duration of a successful transmission, a collision, and slots being idle.
Let Pi be a probability that a sender has no frame to send. Pi can be expressed as the equation 27. In the equation 27, τ is a probability that a node will transmit data in a time period. 1-τ is a probability that a sender has no frame to send.
τ -1
=
Pi (27)
Let Pc be a probability that a transmitted frame experiences a collision. Pc can be expressed as
Pc = τ ( 1 - ( 1 - τ )Nc – Nr -1 ) (28) where Nc is a number of the nodes in the collision area, and Nr is the number of nodes that the receiver’s neighbor can relay.
Nc = π ( Rmax - Rmin )2 ρ (29) Nr = π ( 1.5Rmin - Rmin )2 ρ (30)
ρ is a the density distribution of nodes
Rmin is the distance between a sender and a receiver.
Rmax is the max distance that a sender can send a message
In the equation 28, Since Nc is a number of the nodes in the collision area and Nr is
31
the number of nodes that the receiver’s neighbor can relay, 1-(1 - τ )Nc-Nr-1 means the probability of all nodes in the collision area will at least one node try to transmit data.
τ(1− (1 − τ )Nc-1) means the probability that a node will cause collisions.
Let Pb be a probability that a sender sense the channel busy, and Pb can be expressed as means the probability of all nodes in the busy area will at least one node try to transmit data. τ (1 - (1 - τ )Nb-1) means the probability that a node will cause collisions.
Let Pt be a probability of a sender’s successful transmission. Pt can be expressed as the equation 33.
(successful send packets have packets to send P
( P successful send packets P fail send packets
packets
32
B. Performance analysis
The throughput S for each transmitter is calculated as the amount of successful Transmission (bits) per unit time slot (second). S can be expressed as
busy because of a successful transmission. Tc is an average time the sender is sensed busy by the stations during a collision. Tt is an average time a sender is successfully transmission. Ti, Tb, Tc, and Tt can be expressed as
Fig.3-13 shows a time diagram of packet transmission time.
33
Figure 3 - 13 A time diagram of a packet transmission time
34
Chpater 4 Performance Evaluation
In this chapter, we evaluate the performance of our protocol and compare it with POWMAC and the IEEE 802.11 scheme. We use NS2 (version 2.33) [11] to evaluate the proposed protocol. We compare our protocol with POWMAC because the latter one is also a transmission power control MAC protocol based on a single channel, single transceiver design. In this chapter, we will introduce the environment setting and the simulation result will be shown.
4.1 Environment setting
In this section, we will show the parameters used in our protocol. Some parameters of IEEE 802.11 is shown in table 4-1.
Table 4 - 1 Simulation and analysis parameters Simulation times 60 second
A 0.28
B 8
Data packet size 2 KB
Data rate 1 Mbps
SINR threshold 6 dB Max Transmission range 250 meters
Carrier sensing threshold 3.44283e-09 joule Propagation Model Two-Ray Ground CWMin 31 CWMax 1023
35
Preamble Length 144 PLCP Header Length 48 PLCP Data Rate 1.0e6 Short Retry Limit 7 Long Retry Limit 4
We consider four scenarios according to different topologies and mobility constrains. In the first scenario, we first assume a linear topology and a node will move toward with one node. In the second scenario, we construct 25 nodes separate in a 500*500 square area. In the third scenario, nodes in second scenario will random move toward any direction. In the fourth scenario, we set particular position of six nodes to see the performance of relaying CTS. We compare our protocol in relaying CTS and without relaying CTS method.
4.2 Scenario one
We first simulate a linear topology for the purpose of highlighting the advantages and operational details of our protocol. The distances between the terminals are also shown in the Figure. 4-1. Terminal A is first transmitting to node B, and then node C is transmitting to node D. In this scenerio, node B starts moving to node C at a speed of 1.5 m/s.
36
Figure 4 - 1 A Line topology for scenario one
Fig. 4-2 depicts the throughput of the network. We can see the performance of our protocol is better than POWMAC and IEEE 802.11 MAC. In POWMAC, when nodes want to send data, they need to wait the channel clear and the AW window times. The nodes also need to wait the AW time counter ended even the network is not very crowded. In IEEE 802.11 MAC, nodes can send data only when the channel is clear.
In the first scenario, only one transmission can proceed at a time since all terminals are within the carrier-sense range of each other. However, according to our protocol, in first 10s, the two transmissions A->B and C->D can proceed simultaneously. For the next 40s, when node C gets closer to node B, noises increase between node B and C, and throughputs decrease. After 50s, the interferences become larger than the threshold, node C will not try to transmit data. Therefore, only node A exchanges RTS/CTS with node B, node C can’t transmit to node D.
Figure 4 - 2 Throughputs in the scenario one
37
4.3 Scenario two
We now study the performance under more realistic network topologies. We distribute 25 terminals in a 500m*500m square area. The square is split into 25 smaller square areas, one for each terminal. The location of a terminal within the small square is selected randomly. For each sender, the destination terminal is selected randomly from the one-hop neighbors. We randomly pick ten terminals to send data.
Figure 4 - 3 The topology of scenario two
The performance is shown in Fig. 4-4. It is easy to obvious that the performance of our protocol is better than POWMAC and IEEE 802.11 protocol in different data rates.
In IEEE 802.11 only few pairs can transmit simultaneously. In the simulation, we observe than only 3 nodes can transmit data simultaneously in 802.11. In POWMAC, 5 nodes can transmit data simultaneously. In our protocol, 7 nodes can transmit data simultaneously. Since in IEEE 802.11, nodes overhearing RTS/CTS message, they
38
will set NAV and they can’t send any data. In POWMAC, it needs to use the AW windows to synchronize the transmission. It wastes the AW windows time and data transmission need to synchronize. It can’t make sure that all data is sent to the destination on time. Since POWMAC needs to transmit extra packet “DTS” for synchronization, a extra overhead also reduce throughputs. In our protocol, we can let multiple transmissions exist without synchronization, and our protocol doesn’t need to waste the AW windows time. Although our protocol adjust the power of control messages and relay CTS messages, the hidden terminal problem still can’t completely be avoided. The performance of our protocol can’t complete match with the analysis performance.
Figure 4 - 4 Throughputs in the scenario two
4.4 Scenario three
In scenario three, we add the mobility to nodes in the second scenario. The terminals are randomly selected the direction and their speeds are 0.3m/s. The performance is shown in Fig. 4-5. The performance of our protocol is still better than
39
POWMAC and IEEE 802.11. We adjust the proper power to send data. If nodes move, SNR of a channel will decrease, and we will increase the power until it reaches its upper bounded. The upper bounded power can’t disturb ongoing pairs. Similarly, since we can’t complete avoid hidden terminals problem, we can’t complete match with the analysis performance.
In Fig. 4-6., we set nodes with different speeds between 0.3m/s to 0.7m/s. We can observe that when speeds increase, the throughput of our protocol will decrease. At the beginning, 7 nodes can transmit data simultaneously in our protocol, and 5 nodes can transmit data simultaneously. Throughput of our protocol is 25% up to POWMAC. As time goes by, the distance between senders and receivers increase.
Since senders need to use more power to transmit data, number of nodes which can transmit data simultaneously decreases, and throughputs also decrease.
Fig. 4-7 shows energy consumption in different scenarios. We can observe that energy consumption of our protocol is 18% low to IEEE 802.11 and 15% up to POWMAC. Since in IEEE 802.11, nodes always use a maximum power to send data, nodes wastes a lot of power in data transmission. Our protocol needs to relay CTS to avoid hidden terminal problems, power consumptions are more than POWMAC.
Although our protocol wastes more power than POWMAC, throughputs in our protocol is better than POWMAC. In Fig. 4-8., we can observe that power consumptions per packet. Power consumption per packet in our protocol is 50% low to IEEE 802.11 and 14% low to POWMAC.
In Fig. 4-9, we compare our protocol with relay CTS and without relay CTS method. We observe that when nodes in speed of 3 m/s, throughputs of relay CTS
40
method is higher than without relay method by upper to 15%. When the speed increases, relay CTS method throughputs gradually decrease. When nodes in the speed of 0.7 m/s, throughputs of relay CTS method is close to without relay CTS method. Since nodes moving fast, the relaying CTS method can’t completely relay CTS information to the hidden terminals.
In Fig.4-10 shows energy consumption in different scenarios. We can observe that energy consumption of without-relay-CTS protocol is 17% low to with-relay-CTS protocol. Since nodes may waste some power relay CTS, the energy consumption is more than nodes without relay CTS. In Fig. 4-11, we observe that with-relay-CTS method consumes less energy per Kbps than without-relay-CTS method. Although with-relay-CTS method need to waste some energy to relay CTS, it throughputs is better than without-relay-CTS method. The energy consumption per Kbps is also less than without relay CTS method.
Figure 4 - 5 Throughputs in the scenario three
41
Figure 4 - 6 Throughputs of different speeds
Figure 4 - 7 Energy consumption in different scenarios
42
Figure 4 - 8 Energy consumption per packet in different scenarios
Figure 4 - 9 Throughputs of relay CTS method and without relay CTS method in different speeds
43
Figure 4 - 10 Energy consumption in different scenario
Figure 4 - 11 Energy consumption per Kbps in different scenario
4.5 Scenario four
In the section 3.4, we relay the CTS message to avoid the hidden terminals problems. In order to testing the method of relaying CTS message, we build the topology in the scenario four. The distances between terminals are also shown in the figure. 4-12. We set three flows in the scenario four, A->B, D->C and E->F.
44
Figure 4 - 12 The topology to show the effect of relaying CTS
The Fig. 4-13 shows the performance of relay CTS and without CTS. In Fig. 4-13, we observe node E doesn’t know the ongoing pair D->C in without relay CTS protocol. Node E will use a maximum power to send control messages, so the ongoing pair D->C will be disturbed. Only A->B pair can successful transmission data. In relay-CTS protocol, node F will relay CTS to node E. Node E can adjust the transmission power. Hence A->B, D->C and E->F can transmit simultaneously.
Figure 4 - 13 Performance of relay CTS and without relay CTS in scenario four
45
Chpater 5 Conclusion and Future Work
Since in IEEE 802.11 MAC protocol, when nodes want to transmit data, they need to wait channels clear, and expose problems will happen. However, in IEEE 802.11 physical layer (PHY), nodes can correct decode packets when signal to noise ratio (SNR) reaches a threshold. We combine the packets decode threshold in PHY and sensing channel clear data transmission in MAC layer. We modify IEEE 802.11 MAC protocol based on SNR threshold in PHY layer.
In this essay, we propose a new adaptive transmission power control MAC protocol, which can significantly improve the network throughputs using a single channel and single transceiver environment. Dissimilar with IEEE 802.11, our protocol doesn’t need to wait the channel clear. We adjust the transmission power of control and data packets based on SNR information. When the channel quality is good enough, nodes will start to transmit data. Dissimilar with POWMAC, we adjust of transmission power of control packets instead of data synchronization. The reason is that data synchronization will increase packets delay time, and nodes need to wait some times to transmit data. Since our protocol doesn’t use extra packets, the packets overhead are less than POWMAC. We also consider hidden terminals problems by relaying CTS to receivers’ neighbors. In order to verify our protocol, we analyze throughputs of our protocol and we use ns2 to run simulations.
In simulations, we compare throughputs and power consumption of our protocol, POWMAC, and IEEE 802.11 in four scenarios of different topologies. The simulation has shown that throughput in our protocol is better than POWMAC and IEEE 802.11
46
protocol, and the power consumption in our protocol is more than POWMAC but less than IEEE 802.11.
Some limitations are still on our protocol. First, if nodes speed increase fast, the neighbor table cannot update immediately. Second, since we relay CTS, the power consumptions increase. We will discuss these issues in the future.
47
References
[1] IEEE 802.11 working group. Wireless LAN medium access control (MAC) and physical layer (PHY) specifications.(1999)
[2] IEEE 802.11 working group. Wireless LAN medium access control(MAC) and physical layer (PHY) specifications: high-speed physical layer in the 5GHz Band, (1999)
[3] IEEE 802.11 working group. Wireless LAN medium access control (MAC) and physical layer (PHY) specifications: higher-speed physical layer extension in the 2.4GHz Band, (1999).
[4] Jeffrey P. Monks, Vaduvur Bharghavan, and Wen-mei W. Hwu, “A Power Controlled Multiple Access Protocol for Wireless Packet Networks,” in Proc. IEEE International Conference on Computer Communications (INFOCOM’01), San Francisco, California, USA, Mar. 2001
[5]A. Muqattash and M. Krunz, “Power controlled dual channel (PCDC) medium access protocol for wireless ad hoc networks,” in Proc. IEEE International Conference on Computer Communications (INFOCOM’03), San Francisco, California, USA, Mar. 2003.
[6] A. Muqattash and M. Krunz, “POWMAC: a single-channel power-control protocol for throughput enhancement in wireless ad hoc networks,” IEEE J. Select.
Areas Communications , vol. 23, no. 5, pp. 1067–1084, May 2005.
[7] Bassel Alawieh, Chadi Assi and Wessam Ajib, “A Power Control Scheme for Directional MAC Protocols in MANET,” in Proc. IEEE Wireless Communications and Networking Conference, Hong Kong, March 2007.
[8] Swades De, Komlan Egoh and Gaurav Dosi, “receiver initiated power control multi-access protocol in wireless ad hoc networks,” in Proc. IEEE Sarnoff Symposium,
48
New Jersey, USA , May 2007.
[9] Kuei-Ping Shih, Chau-Chieh Chang and Yen-Da Chen, “A Fragmentation-based Data Collision Free MAC Protocol with Power Control for Wireless Ad Hoc Networks,” in Proc. I IEEE Wireless Communications and Networking Conference, Las Vegas, USA, April 2008.
[10] Minghao Cui and Violet R. Syrotiuk, “Fair Variable Transmission Power Control” in Proc. IEEE Global Telecommunications Conference, Washington, DC, USA, November 2007.
[11] NS-2 simulator, http://www.isi.edu/nsnam/ns/
[12] C. Ware, J. Chicharo, and T. Wysocki, “Simulation of capture behavior in IEEE 802.11 radio modems,” in Proc. IEEE Veh. Technol. Conf., vol.3, Fall 2001, pp. 1393–
1397.
[13] R. Wattenhofer, L. Li, P. Bahl, and Y.-M. Wang, “Distributed topology control for power efficient operation in multihop wireless ad hoc networks,” in Proc. IEEE INFOCOM Conf., 2001, pp. 1388–1397.
[14] S.-L. Wu, Y.-C. Tseng, and J.-P. Sheu, “Intelligent medium access for mobile ad hoc networks with busy tones and power control,” IEEE J. Sel. Areas Commun., vol.
18, no. 9, pp. 1647–1657, Sep. 2000.
[15] K. Xu, M. Gerla, and S. Bae, “How effective is the IEEE 802.11 RTS/CTS handshake in ad hoc networks,” in Proc. IEEE GLOBECOM
Conf., vol. 1, Nov. 2002, pp. 72–76.
[16] C.-H. Yeh, “ ROAD: A variable-radiusMAC protocol for ad hoc wireless networks,” in Proc. IEEE Veh. Technol. Conf., vol. 1, 2002, pp. 399–403.
[17] H. Zhou, C.-H. Yeh, and H. Mouftah, “ A solution for multiple access in variable-radius mobile ad hoc networks,” in Proc.IEEE Int. Conf. Commun., Circuits, Syst. , vol. 1, 2002, pp. 150–154.