Tccp mass replay

Answer 1

The main purpose of the protocol was to prevent network packets from being dropped. The triggering methods of the retransmissions were time-out retransmissions, fast retransmissions, Sack retransmissions, and Dual Sack (D-Sack).

** 1. Overtime retransmissions result in a large number of retransmissions **
1. ** Triggered scenario **
- When a data packet is lost or an ACK response is lost, a time-out re-transmission may be triggered. The sender sets a timer when sending data. If the sender does not receive an acknowledgment from the receiver after a specified time (RTO, Time Out Retransmission Time), the sender will resend the data packet.
2. ** Effect of RTO setting on mass retransmissions **
- The RTO could not be too long or too short. If the RTO was too long, it would take a long time to re-transmit if a packet was lost, which would be inefficient. If the RTO was too short, if the packet was blocked due to the network condition and the transmission was slow but not lost, it would also trigger re-transmission, which would cause the network to be more congested and trigger more re-transmissions. In a linux-based system, to calculate RTO, two samples were taken of the RTL (packet round-trip time). The RTL time was measured and then a smoothed RTL value was calculated by the weighed average. The fluctuation range of the RTL was also measured to avoid the situation where large fluctuations could not be found. The formula for calculating RTO is suggested in the Robuchelle Foundation's Robuchelle Foundation. RTO is calculated for the first time (let R1 be the first measurement of the RTL): RTL = R1, DevRTL = R1/2, RTO=[? * RTL +? * DevRTL =? * R1+? * (R1/2)]; RTO is calculated for the subsequent time (let R2 be the new measurement of the RTL): RTL = RTL +[* (RTL-RTL)=R1 +* (R2 - R1)], DevRTL =(1 -Beta) * DevRTL +[*(RTL-RTL)]| RTT - SRTT|)=(1 -Beta) * (R1/2)+ Beta * (R2 - R1), RTO =, RTO =,
- If the network jam triggered a time-out, the time-out period would be doubled to prevent frequent retransmissions from worsening the network jam. That is, the next time-out period would be set to twice the current value every time the network was retransmitted. However, if the network condition continued to be poor, it might still cause a large number of retransmissions.
3. ** The relationship between the drawback of overtime retransmissions and mass retransmissions **
- The drawback of a time-out re-transmission was that the time-out period might be relatively long, and the waiting time for the re-transmission might be too long. In the case of an unstable network, this could lead to a large number of data packets accumulating and waiting to be retransmitted, resulting in a large number of retransmissions.

** 2. Rapid retransmissions result in a large number of retransmissions **
1. ** Activation mechanism **
- Fast Retransmission no longer used time as the standard for retransmissions, but data as the standard for retransmissions. For example, when the receiver receives subsequent packets (such as Seq3, Seq4, and Seq5) but does not receive the Seq2 packet, the receiver will reply three times with an ACK2 packet. After the sender receives three packets of AK2, it will re-transmit the Seq2 packet before the timer times out.
2. ** A factor that may trigger a large number of retransmissions **
- Since the sender did not know which packets of the sender responded to the three ACK2 packets (namely Seq3, Seq4, Seq5), there might be situations where only Seq2 or all the packets were retransmitted depending on the implementation of the protocol. If all the data packets were to be retransmitted, it might cause a large number of retransmissions when there were many similar situations in the network.

** 3. The relationship between Sack and D-Sack and mass retransmissions **
1. **Sack Retransmission **
- Sack retransmissions were selective retransmissions. It was used to solve the problem of not knowing which packets needed to be retransmitted during fast retransmissions. By adding a SACK field to the options field of the protocol, the receiver will send the sequence number range of the received data packet to the sender. This way, the sender can find the lost data packet through the SACK message and resend the data packet. If the SACK mechanism fails, for example, if the sender or receiver does not support it (SACK must be supported by both the sender and receiver. It can be enabled through the net.ipv4.tcc_Sack parameters in Linux2.4 and later), or if the SACK message transmission is wrong, it may cause the accurate selective retransmissions to be unable, which may lead to unnecessary retransmissions or even a large number of retransmissions.
2. **D - SACK**
- D-Sack could let the sender know which data packets were received repeatedly. It could help determine whether the sent packet was lost or the receiver's response was lost. It could also know whether the sender's data packet was delayed or duplicated in the network. If there was a problem with the D-SACK mechanism, such as not being able to accurately determine the above situation, it would affect the sender's judgment of the network situation, which may lead to wrong retransmissions. In some cases, it may lead to a large number of retransmissions.

"Choose" was equally exciting. Everyone was welcome to read it!