I am running the simulation for multi-hop network to observe the effect of the different number of hops on the CBR throughput. I ran the simulation for hops = 1, 2, 3, 4 and 5. After running the simulation, I have observed that CBR throughput remains almost same for the values 1, 2 and 3 but when I increased hops to 4, the CBR throughput decreased slightly. The most surprising aspect of the simulation results was that when I increased the hop count to 5, the CBR throughput jumped to 73 % from 67 %.
I fail to understand this upward trend as I thought that due to different aspect such as duty cycling, nodes synchronization, the throughput should decrease when increasing the hop count but the results for counts count 5 were surprising. Can someone explain me why is the reason for this behavior?
Throughput of network traffic decreases with an increase in the hops count because the packet error probability increases by increasing the number of hops, causing the network throughput to decrease.
With respect to the real scenario, at every node it is required to add buffer for handle the processing delay and packet loss, Note buffer size is crucial. In every link path, there should be propagation delay. At every node, the buffer size should be maintained (case 1 fixed for every node and case 2 different sizes for the different node), use some mechanism likes M/M/1/G/PS. Try to modify your algorithm and code.
In multi-hop network, the throughput of the network increase with the increase in number of hops and this increment is exponential. I would suggest you to read this paper, which will satisfy you for your question.
http://ieeexplore.ieee.org/document/6506536/
In your results, the number of packets sent in the case of 5 hops (26103 packets) is less than the four cases above it. Then, the number of hops is not the only factor in your algorithm. To verify your results, you have to run your algorithm many times with different values of packets sent and take the average of your results.
use Below AWK :
BEGIN {
recvdSize = 0
startTime = 400
stopTime = 0
}
{
event = $1
time = $2
node_id = $3
pkt_size = $8
level = $4
# Store start time
if (level == "AGT" && event == "s" && pkt_size >= 512) {
if (time < startTime) {
startTime = time
}
}
# Update total received packets' size and store packets arrival time
if (level == "AGT" && event == "r" && pkt_size >= 512) {
if (time > stopTime) {
stopTime = time
}
# Rip off the header
hdr_size = pkt_size % 512
pkt_size -= hdr_size
# Store received packet's size
recvdSize += pkt_size
}
}
END {
printf("Average Throughput[kbps] = %.2f\t\t StartTime=%.2f\tStopTime=%.2f\n",(recvdSize/(stopTime-startTime))*(8/1000),startTime,stopTime)
}
Any related work on Throughput utilization of WSN for worst case situations?
- Badges
- Science topic