congestion-window-tcp-comparison

TcpNewReno :
packetLossCount TcpNewReno 38
this is project made for comparing different tcp congestion control algorithm using gns3

The basic idea of these extensions to the fast retransmit and fast recovery algorithms are as The TCP
sender can infer, from the arrival of duplicate acknowledgments, whether multiple losses in the same
window of data have most likely occurred, and avoid taking a retransmit timeout or
making various congestion window reductions due to such an event.
● Looking at the plot, we observe that the slow start phase is linear in the collision avoidance
phase. We get curved output at the end; hence as soon as we experience the congestion window
again drop minimum value.
● We also observed that at once congestion detected slope of the congestion window is high when
it is low but as it gets high gradient also get lesser anticipating congestion.
TcpHighSpeed :
packetLossCount TcpHighSpeed 38
Tip highspeed Is a modification of the TCP-Reno congestion control mechanism with TCP
connections with large congestion windows. H-TCP is a loss-based algorithm, using
additive-increase/multiplicative- decrease to
control the TCP congestion window

● In the graph, we don’t see much difference with New Reno.
● But at the top at TCP acknowledgment loss or timeout, we see there are sharper edges than
usual.
TcpVeno :
Packet Loss count TCP Veno is 38
TCP Veno module is a congestion control module to improve TCP performance over wireless networks.
It improves over the TCP Reno congestion control algorithm by using the estimated state of a connection
based on TCP Vegas.TCP Veno reduces the "blind" reduction of TCP window regardless of the
cause of packet loss. This TCP version distinguishes between accidental loss(non- congestion state)
and congestion loss (congestion state). Also, depending on this difference, it refines the congestion
window adjustment. In the wireless environment, the packet loss is because of the noise and link error.

Packet Loss count TCP Veno is 38
TCP Veno module is a congestion control module to improve TCP performance over wireless networks.
It improves over the TCP Reno congestion control algorithm by using the estimated state of a connection
based on TCP Vegas.TCP Veno reduces the "blind" reduction of TCP window regardless of the
cause of packet loss. This TCP version distinguishes between accidental loss(non- congestion state)
and congestion loss (congestion state). Also, depending on this difference, it refines the congestion
window adjustment. In the wireless environment, the packet loss is because of the noise and link error.
If we compare TCP new reno, there is not much difference in the graph, but we can see that
curves and lines are much smoother for TCP Veno.
TcpVegas:
Packet loss Count 38
TCP Vegas emphasizes packet delay, rather than packet loss, as a signal to determine the
rate at which to send packets. Unlike TCP-Reno, which detects congestion only after it
has happened via packet drops, TCP-Vegas detects congestion at a nascent stage
based on increasing RTT values of the packets in the connection.

looking at the graph, we can see that the congestion window not dropping below a certain
threshold throughout minimum congestion window base is higher as compared to other
algorithms
Channel Bandwidth 2Mbps Application Datarate 2Mbps

● When channel bandwidth is almost the same with app datarate 2 Mbps, we see a more
random variable drop after collision base return.
● 2 Mbps bandwidth we see reached highest congestion window close to 16000 that is
● The highest congestion window throughout the transfer is almost the same for all
configurations.
● According to graphs, we can deduce that after some threshold value of bandwidth of
channel congestion doesn’t get affected by that much margin.
App Datarate 0.5 Mbps Channel Bandwidth 6Mbps