Evaluation with different Duty Cycle

In the first evaluation on the random distributed network, we take duty cycle as the control parameter to measure the performance of each MAC protocol. The value of duty cycle varies from 3% to 16%. When duty cycle is 3%, sensor nodes only work 0.09 second every 3 second in S-MAC and 0.12 second every 3.9 second in LAMAC.

When duty cycle is 16%, sensor nodes work relatively more time. The parameter is used to adjust the percentage of active and sleeping period and also affects the lifetime of the whole network. The system is under medium traffic load to simulate an average case. The packet generation interval is one packet every 10 second at each sensor node.

Figure 5-8 shows the average transmission latency with different duty cycle. The result shows that LAMAC and 802.11 CSMA can achieve quite good transmission latency with different duty cycle. With 802.11 CSMA, the duty cycle can not affect its performance at all because it has no sleeping schedule and remain active all the time.

With LAMAC, we can use active-sleep cycle more flexibly by using adaptive sleeping scheme. Whenever a sensor node gets packet to send, it can inform its parent node to reserve another extra adaptive active period to transmit this packet instead of

waiting for the next frame. And when parent node receives the FRP packet, it also sends a FRP packet to its next hop node to reserve adaptive active period. The extra adaptive active period packet will propagate along the path and provides packets to be transmitted efficiently. Thus the duty cycle also doesn’t affect LAMAC much.

However, the result shows that S-MAC is affected by the duty cycle very seriously. Its performance gets worse with the increasing of duty cycle. The reason is that although S-MAC also use adaptive active scheme, it can not reserve the extra active period along the path. Packets only can be forwarded two hops away in a frame. If packets need to be forwarded more than two hops, they have to wait until the beginning of next frame. This means S-MAC can not use sleeping period efficiently.

0

Figure 5-8： Average transmission latency with different duty cycle

Figure 5-9 shows the average energy consumption per packet with different duty cycle. The result shows that 802.11 CSMA uses much more power to transmit a

packet than LAMAC and S-MAC. This extra energy consumption is caused by its full active schedule and idle listening problem. S-MAC uses about three times the energy than LAMAC. This is because S-MAC needs all nodes to wake up to transmit a little amount of packets. For example, there may be only one packet in the network.

However, in order to put this packet one hop away, the whole sensor nodes need to wake up one time. Even with the adaptive active scheme, it still only can move this packet two hops away in a frame. Besides this, the adaptive active technique in S-MAC has an important drawback. It requires every node within the communication range of the sender and receiver to awake in the end of the current transmission.

However, only one node will be the next active node. This means most nodes waste energy to perform the adaptive active procedure. Figure 5-9 shows LAMAC achieves the best energy efficiency. There is only a slight increasing of energy consumption when duty cycle is low. This is because our experiment is running under medium traffic load. And the low duty cycle is used for heavy traffic load. Thus, sensor nodes waste some energy in idle listening when there are no packets to send.

0 50 100 150 200 250 300

1/6 1/9 1/12 1/15 1/18 1/21 1/24 1/27 1/30 1/33 Duty Cycle

Energy Consumption Per Packet (mJ)

LAMAC S-MAC 802.11 CSMA

Figure 5-9： Energy consumption per packet with different duty cycle

Figure 5-10 shows the throughput with different duty cycle. The definition of throughput is the amount of data received by base station in a certain interval. In out experiment, the throughput means the amount of data which base station can receive every second. Figure 5-10 shows 802.11 CSMA has the best performance. This is because its full active schedule. LAMAC has the second throughput next to 802.11 CSMA and only has small difference with 802.11 CSMA. The result also shows the throughput of S-MAC is much worse than LAMAC. This is because every node has many interference nodes by using S-MAC. With LAMAC, sensor nodes only contend with nodes which are the same level. However, by using S-MAC, sensor nodes not only contend with the same level nodes but also the nodes which are the higher level and the lower level. Thus, its throughput is much worse than LAMAC.

0

Figure 5-10： Throughput with different duty cycle

In the last evaluation with different duty cycle, we will measure the protocol work efficiency. Figure 5-11 shows the protocol work efficiency with different duty cycle.

The result shows LAMAC uses energy much more effectively than others. Although S-MAC can extend the lifetime of the whole system by using periodical sleeping, it actually uses energy in the similar way with 802.11 CSMA when sensor nodes are in active period. In other words, S-MAC just extends the lifetime of system but not improve the usage of energy. Even though S-MAC uses the adaptive active technique, it can only improve the work efficiency slightly. On the contrary, LAMAC does improve the energy usage. We reduce the medium competition of each sensor nodes and increase the transmission probability. This makes the performance of LAMAC much better than others.

0 20000 40000 60000 80000 100000 120000 140000 160000

1/6 1/9 1/12 1/15 1/18 1/21 1/24 1/27 1/30 1/33 Duty Cycle

Data Get before System Shut Down (Bytes)

LAMAC S-MAC 802.11 CSMA

Figure 5-11： Data get with different duty before system shut down