The document describes a lecture on the Transport layer topic of TCP congestion control. It discusses how TCP determines network capacity, adjusts the number of packets in transit, and uses acknowledgments to pace transmission. It also describes congestion causes like buffer shortages, slow links, and processors. TCP congestion control uses additive increase multiplicative decrease and slow start to dynamically adjust the congestion window size based on perceived network congestion.
The document describes a lecture on the Transport layer topic of TCP congestion control. It discusses how TCP determines network capacity, adjusts the number of packets in transit, and uses acknowledgments to pace transmission. It also describes congestion causes like buffer shortages, slow links, and processors. TCP congestion control uses additive increase multiplicative decrease and slow start to dynamically adjust the congestion window size based on perceived network congestion.
The document describes a lecture on the Transport layer topic of TCP congestion control. It discusses how TCP determines network capacity, adjusts the number of packets in transit, and uses acknowledgments to pace transmission. It also describes congestion causes like buffer shortages, slow links, and processors. TCP congestion control uses additive increase multiplicative decrease and slow start to dynamically adjust the congestion window size based on perceived network congestion.
The document describes a lecture on the Transport layer topic of TCP congestion control. It discusses how TCP determines network capacity, adjusts the number of packets in transit, and uses acknowledgments to pace transmission. It also describes congestion causes like buffer shortages, slow links, and processors. TCP congestion control uses additive increase multiplicative decrease and slow start to dynamically adjust the congestion window size based on perceived network congestion.
Download as PPT, PDF, TXT or read online from Scribd
Download as ppt, pdf, or txt
You are on page 1/ 32
Computer networks
Name: K.SUDHA Designation: Lecturer Department: Electrical and Electronics Engineering
Subject code: CS2361
Year: III Unit: III Title: Introduction to Transport layer TCP Overview End to end issues Segment format Connection establishment TCP sliding window Stream control Transmission Protocol Simple demultiplexor TCP Congestion Control • Determines the network capacity • Adjust the number of packets that can have safely in transit • Acks to pace the transmission of packets • TCP is self clocking • Avoids congestion • Maxwindow=MIN(CongestionWindow,AdvertisedWindow) • EffectiveWindow=MaxWindow-(LastByteSent-LastByteAcked) Caused By • the shortage of buffer space. • slow links. • slow processors
• Possible solutions – End-to-end versus link-by-link control – Rate-Based versus Credit-Based control – The rate-based traffic-flow technique constantly – Integrated congestion control • Integrated congestion control Principles of Congestion Control
Congestion: • informally: “too many sources sending too much data too fast for network to handle” • different from flow control! • manifestations: – lost packets (buffer overflow at routers) – long delays (queueing in router buffers) • a top-10 problem! Scenario 1: Queuing Delays Host A in : original data out • two senders, two receivers Host B unlimited shared • one router, infinite output link buffers
buffers • no retransmission
• large delays when
congested • maximum achievable throughput Scenario 2: Retransmits • one router, finite buffers • sender retransmission of lost packet
Host A in : original data out
'in : original data, plus
retransmitted data
Host B finite shared output
link buffers Scenario 3: Congestion Near Receiver • four senders Q: what happens as in • multihop paths and increase ? • timeout/retransmit in Host A out in : original data 'in : original data, plus retransmitted data finite shared output link buffers
Host B Approaches towards congestion control Two broad approaches towards congestion control:
• no explicit feedback from control: network • routers provide feedback to end • congestion inferred from end- systems system observed loss, delay – single bit indicating • approach taken by TCP congestion (SNA, DECbit, TCP/IP ECN, ATM) – explicit rate sender should send at TCP Congestion Control • end-end control (no network How does sender perceive assistance) congestion? • sender limits transmission: • loss event = timeout or 3 LastByteSent-LastByteAcked CongWin duplicate acks • Roughly, • TCP sender reduces rate (CongWin) after loss CongWin event rate = Bytes/sec RTT function of • CongWin is dynamic, three mechanisms: perceived network congestion – AIMD – slow start – conservative after timeout events TCP AIMD multiplicative decrease: cut additive increase: increase CongWin in half after CongWin by 1 MSS every loss event RTT in the absence of loss events: probing c o n g e s tio n w in d o w
2 4 K b y te s
1 6 K b y te s
8 K b y te s
tim e
Long-lived TCP connection
TCP Slow Start • When connection begins, When connection begins, increase rate exponentially CongWin = 1 MSS – Example: MSS = 500 bytes & fast until first loss event RTT = 200 msec – initial rate = 20 kbps • available bandwidth may be >> MSS/RTT – desirable to quickly ramp up to respectable rate TCP Slow Start (more) • When connection begins, Host A Host B increase rate exponentially until first loss event: one segm ent
RTT – double CongWin every RTT – done by incrementing two segm ents CongWin for every ACK received • Summary: initial rate is slow but four segm ents ramps up exponentially fast
time Refinement (more)
Q: When should the
exponential increase switch to linear? A: When CongWin gets to 1/2 of its value before timeout.
Implementation: • Variable Threshold • At loss event, Threshold is set to 1/2 of CongWin just before loss event TCP sender congestion control Event State TCP Sender Action Commentary ACK receipt Slow Start CongWin = CongWin + Resulting in a doubling for (SS) MSS, of CongWin every RTT previously If (CongWin > Threshold) unacked set state to data “Congestion Avoidance” ACK receipt Congestio CongWin = CongWin+MSS Additive increase, for n * (MSS/CongWin) resulting in increase of previously Avoidance CongWin by 1 MSS unacked (CA) every RTT data Loss event SS or CA Threshold = CongWin/2, Fast recovery, detected by CongWin = Threshold, implementing triple Set state to “Congestion multiplicative decrease. duplicate Avoidance” CongWin will not drop ACK below 1 MSS. Timeout SS or CA Threshold = CongWin/2, Enter slow start CongWin = 1 MSS, Set state to “Slow Start” Duplicate SS or CA Increment duplicate ACK CongWin and Threshold ACK count for segment being not changed Congestion Avoidance Mechanisms • Helps to avoid congestion • Additional functionality into the router to assist in anticipation of congestion • to control congestion once it happens
• to repeatedly increase load in an effort to find the point at
which congestion occurs, and then back off Mechanisms • router-centric: DECbit and RED Gateways
• host-centric: TCP Vegas
DECbit
• DECbit
• Add binary congestion bit to each packet header
• Router – monitors average queue length over last busy+idle cycle
– set congestion bit if average queue length greater than 1
when packet arrives
– attempts to balance throughput against delay
• DECbit
• End Hosts • destination echos bit back to source
• source records how many packets resulted in set bit
• if less than 50% of last window's worth had bit set, then increase CongestionWindow by 1 packet
• if 50% or more of last window's worth had bit set, then decrease CongestionWindow by 0.875 times UDP—User Datagram Protocol • An unreliable, connectionless transport layer protocol • UDP format. See picture • Two additional functions beyond IP: – Demultiplexing: deliver to different upper layer entities such as DNS, RTP, SNMP based on the destination port # in the header. i.e., UDP can support multiple applications in the same end systems. – (Optionally) check the integrity of entire UDP. (recall IP only checks the integrity of IP header.) • If source does not want to compute checksum, fill checksum with all 0s. • If compute checksum and the checksum happens to be 0s, then fill all 1s. • UDP checksum computation is similar to IP checksum, with two more: – Add extra 0s to entire datagram if not multiple of 16 bits. – Add pseudoheader to the beginning of datagram. UDP pseudoheader
29 UDP datagram
0 16 31
Source Port Destination Port
UDP Length UDP Checksum
Data
Back to UDP—User Datagram Protocol
30 Figure 8.16 Back to UDP—User Datagram Protocol
UDP pseudoheader
0 8 16 31
Source IP Address
Destination IP Address
00000000 Protocol = 17 UDP Length
1.Pseudoheader is to ensure that the datagram has indeed
reached the correct destination host and port. 2. The padding of 0s and pseudoheader is only for the computation of checksum and not be transmitted.
