Analysis of TCP variants
Download as PPTX, PDF6 likes3,230 views
This document discusses several TCP congestion control algorithms: TCP Tahoe, Reno, New Reno, SACK, and Vegas. It provides details on how each algorithm handles slow start, congestion avoidance, fast retransmit, and congestion detection. TCP Vegas is highlighted as being superior to the other algorithms because it can detect and retransmit lost packets faster, has fewer retransmissions, more efficiently measures bandwidth availability, and experiences less congestion overall through proactive congestion detection and modified slow start and congestion avoidance.
Analysis of TCP variants
- 1. By Name: Nirav Bhojani Roll No: 14MCEI05
- 2. TCP TAHOE RENO NEW RENO SACK TCP VEGAS
- 3. Congestion control algorithm Slow Start Phase: Sender sets cwnd to 1 and for each ACK received it increase cwnd by 1. It follows until packets are loosed and then we decrease sending rate and set cwnd to 1 and start again Detects packet loss by timeouts
- 4. Tahoe uses Additive Increase Multiplicative Decrease After timeouts, it sets cwnd to 1 and threshold value(ssthresh) to half of cwnd and starts slow start until it reaches threshold value. After that it increments linearly until it encounters packet loss
- 6. Takes a complete timeout interval to detect a packet loss every time a packet is lost it waits for a timeout and the pipeline is emptied.
- 7. Retains slow start and retransmit timer of Tahoe In Fast Retransmit, after receiving 3 duplicate acks, it denotes that segment was lost And enters into Fast-Recovery
- 8. 1. When third duplicate ack is received • Set ssthresh = cwnd/2 • Retransmit missing segment • Set cwnd = ssthresh +3 2. After getting duplicate acks, increment cwnd by 1 3. When next ack arrives which acknowledges new data then set cwnd = ssthresh
- 10. Performs very well when the packet losses are small When there are multiple packet losses then detects only one packet loss in one RTT If the window is very small and when the loss occurs then we would never receive enough duplicate acknowledgements for a fast retransmit and we would have to wait for a timeout
- 11. Detects multiple packet losses and thus is much more efficient that RENO in the event of multiple packet losses. it differs from RENO in that it doesn’t exit fast-recovery until all the data which was out standing at the time it entered fast recovery is acknowledged.
- 12. it notes the maximums segment which is outstanding and when a fresh ACK is received then there are two cases: 1. If it ACK’s all the segments which were outstanding when we entered fast recovery then it exits fast recovery and sets CWD to ssthresh and continues congestion avoidance like Tahoe. 2. If the ACK is a partial ACK then it deduces that the next segment in line was lost and it re-transmits that segment and sets the number of duplicate ACKS received to zero.
- 13. New-Reno suffers from the fact that it takes one RTT to detect each packet loss. When the ACK for the first retransmitted segment is received only then can we deduce which other segment was lost.
- 14. Extension of TCP Reno Retains slow-start and fast retransmit parts of RENO Requires that segments should be acknowledged selectively. Thus each ACK has a block which describes which segments are being acknowledged.
- 15. When the sender enters fast recovery, it initializes a variable pipe which is an estimate of how much data is outstanding in the network, and it also set CWND to half the current size. Every time it receives an ACK it reduces the pipe by 1 and every time it retransmits a segment it increments it by 1. Whenever the pipe goes smaller than the CWND window it checks which segments are un received and send them. If there are no such segments outstanding then it sends a new packet. Thus more than one lost segment can be sent in one RTT.
- 16. The biggest problem with SACK is that currently selective acknowledgements are not provided by the receiver. To implement SACK we will need to implement selective acknowledgment which is not a very easy task.
- 17. Proactive measure to encounter congestion It does not depend solely on packet loss as a sign of congestion. It detects congestion before the packet losses occur. Three major changes induced by Vegas are: 1. New Re-Transmission Mechanism 2. Congestion avoidance 3. Modified Slow-start
- 18. It keeps track of when each segment was sent and it also calculates an estimate of the RTT by keeping track of how long it takes for the acknowledgment to get back. Whenever a duplicate acknowledgement is received it checks to see if the (current time - segment transmission time) > RTT estimate; if it is then it immediately retransmits the segment without waiting for 3 duplicate acknowledgements or a coarse timeout.
- 19. To catch any other segments that may have been lost prior to the retransmission, when a non duplicate acknowledgment is received, if it is the first or second one after a fresh acknowledgement then it again checks the timeout values and if the segment time since it was sent exceeds the timeout value then it re-transmits the segment without waiting for a duplicate acknowledgment. Thus in this way Vegas can detect multiple packet losses.
- 20. It does not use the loss of segment to signal that there is congestion. It determines congestion by a decrease in sending rate as compared to the expected rate, as result of large queues building up in the routers. Thus whenever the calculated rate is too far away from the expected rate it increases transmissions to make use of the available bandwidth, whenever the calculated rate comes too close to the expected value it decreases its transmission to prevent over saturating the bandwidth. Thus Vegas doesn’t waste bandwidth by transmitting at too high a data rate and creating congestion.
- 21. Vegas increases exponentially only every other RTT, between that it calculates the actual sending through put to the expected and when the difference goes above a certain threshold it exits slow start and enters the congestion avoidance phase.
- 22. TCP Vegas is definitely better than 1. Tahoe: It can detect and retransmit lost packet much sooner than timeouts in Tahoe. It also has fewer re-transmissions since it doesn’t empty the whole pipe whenever it loses packets. It is better at congestion avoidance and its modified congestion avoidance and slow start algorithms measure incipient congestion and very accurately measures the available bandwidth and therefore use network resources efficiently and don’t contribute to congestion.
- 23. 2. Reno: More than half of the timeouts of Reno are prevented by Vegas as it detects and re-transmits more than one lost packet before timeout occurs. It doesn’t have to always wait for 3 duplicate packets so it can retransmit sooner.
- 24. 3. New-Reno: It prevents many of timeouts of New-Reno as it doesn’t need to wait for 3 duplicate ACK’s before it retransmits a lost packet. Its congestion avoidance mechanisms to detect incipient congestion are very efficient and utilize network resources much more efficiently. Because of its modified congestion avoidance and slow start algorithm there are fewer retransmits.
- 25. 4. SACK In its estimation of incipient congestion, and its efficient estimation of congestion by measuring change in throughput rather than packet loss. This would result in a better utilization of bandwidth and lesser congestion. Also it appears more stable than SACK. The reason for this being that SACK uses packet losses to denote congestion. So that the sender continually increase sending rate until there is congestion. Another advantage of TCP Vegas or rather the disadvantage of SACK is that it is not very easy to incorporate SACK in the current TCP.
- 26. https://www.hep.ucl.ac.uk/~ytl/tcpip/backgr ound/tahoe-reno.html https://sites.google.com/site/networksprojec twiki/bit10/compnetworks/tcp-performance-in- high-speed-long-distance-networks/ working-oftcp-tahoe-newreno-and- vegas https://inst.eecs.berkeley.edu/~ee122/fa05/ projects/Project2/SACKRENEVEGAS.pdf Chapter 12 from High Speed Networks and Intranets by William Stallings
Editor's Notes
- #19: Thus it gets around the problem faced by Reno of not being able to detect lost packets when it had a small window and it didn’t receive enough duplicate ACK’s.