Computer networks have experienced an explosive growth over the past few years and with that growth have come severe congestion problems. For example, it is now common to see internet gateways drop 10% of the incoming packets because of local buffer overflows. Our investigation of some of these problems has shown that much of the cause lies in transport protocol implementations ( not in the protocols themselves): The ‘obvious’ ways to implement a window-based transport protocol can result in exactly the wrong behavior in response to network congestion. We give examples of ‘wrong’ behavior and describe some simple algorithms that can be used to make right things happen. The algorithms are rooted in the idea of achieving network stability by forcing the transport connection to obey a ‘packet conservation’ principle. We show how the algorithms derive from this principle and what effect they have on traffic over congested networks. In October of ’86, the Internet had the first of what became a series of ‘congestion col- lapses’. During this period, the data throughput from LBL to UC Berkeley (sites separated by 400 yards and two IMP hops) dropped from 32 Kbps to 40 bps. We were fascinated by this sudden factor-of-thousand drop in bandwidth and embarked on an investigation of why things had gotten so bad. In particular, we wondered if the 4.3 BSD (Berkeley U NIX ) TCP was mis-behaving or if it could be tuned to work better under abysmal network conditions. The answer to both of these questions was “yes”.

1. Congestion Control avoidance
Seminar Presented
by
Mr. Anthony-Claret Onwutalobi
On the 10th of March (Winter Session)
University of Helsinki, Finland

2. Introduction
 Definition of Terms:
 Congestion - is a condition of severe
delay caused by an overload of datagram
at one or more switching point (router).
[2]
 Congestion Control: is a distributed
algorithm use by TCP to share network
resources among competing users. [3]

3. Area to Cover
 Overview of Congestion Control
 The use of Old TCP and its problem
 The Introduction of New TCP and its
Benefits
 The limitation of the four Algorithm
 Summary
 Suggestions and questions

4. Introduction Continue..
 The Old TCP and How it operates:
 UDP -
 The Short Coming
 Packet transmitted via Connectionless network
 Packet can be lost or destroyed when network
hardware fails
 Packet loss due to heavy load.
 Packet arrived out of order
 Packet delivered after a long delay.

5. The New TCP
(Positive Acknowledgement Transmission Protocol)
 In principle, a reliable delivery with no
duplicate or data loss as oppose to UDP
protocol.
 This mechanism requires both endpoints to be
on a constant communication whereby the
sender sends application packages (datagram)
and wait for an acknowledgement from the
receiver.
 The sender stores the history of the packages
sent and also set a timer when it sends a
packet. It waits for a reply for a set periodic
and will retransmit the package if no reply ACK
received.

6. Short Coming of New TCP
 solves the problem of Connectionless UDP
protocol by ensuring the reliability of data
transmitted over the network,
 however it is too mechanical and it wastes
a lot of network bandwidth.
 It also has poor utilization of network
resources because the sender must delay
sending a new package until it receives an
acknowledgement for the previous packet

7. Introducing Sliding Window
 This limitation propels a research and introduction of
sliding window.
 Sliding window keeps tracks of transmitted and
acknowledged packets.
 Since the timer is set for each datagram sent over
the network, the lost packets are easily identified
and retransmitted
 Although this sliding window protocol is connection-oriented
which guarantee data connection.

8. Source http://www.tutorialsweb.com/networking/tcpip/images/Fig10_SlidingWindow.jpg

9. Slow Start
 Slow-start is run when the pipe is empty (i.e.,
when first starting or re-starting after a timeout).
Its goal is to get the "ack clock" started so
packets will be metered into the network at a
reasonable rate.
 Algorithm:
 Add a congestion window, cwnd, to the per-connection
state.
 When starting or restarting after a loss, set cwnd to
one packet
 On each Ack for new data, increase cwnd by one
packet.
 When sending, send the minimum of the receiver’s
advertised window and cwnd.

10. Assumption: The maximum congestion window size is 8. Therefore, the congestion window size will not
increase after it reaches the size 8. The sender starts with the congestion window size 1. Upon each ACK
from the receiver, the sender increases the congestion window size by 1. For example, when the sender
receives "ACK 1", its congestion window
size is increased to be 2. Thus, in the next window, the sender can send 2 packets consecutively without
receiving any ACKs from the receiver.

11. Congestion Avoidance
 congestion avoidance, is run any
time *except* when (re-)starting
and is responsible for estimating the
(dynamically varying) pipe size[4]


12. When to Enter Congestion Avoidance
 Congestion avoidance is a TCP
restriction technique of regulating
slow start exponential duplication
method to avoid flooding the
network with segment which could
cause congestion.
 Sender receives Ack over a lost
packet.
 Sender timeout Expires

13. Slow Start and Congestion Avoidance
Algorithm
 Slow Start and Congestion Avoidance Algorithm: [3]
 Initialization for a given connection
 Sets cwnd to one segment and ssthresh to 65535 bytes
 The TCP output routine never sends more than minimum of cwnd and
the receiver’s advertised window
 When congestion occurs (indicated by a timeout or the reception of
duplicate Ack),
 One-half of the current window size (the minimum of cwnd and the
receiver's
 advertised window, but at least two segments) is saved in ssthresh.
 Additionally, if the congestion is indicated by a timeout, cwnd is set to
one segment
 (slow start)
 When new data is acknowledged by the other end, increase cwnd, but
the way it
 increases depends on whether TCP is performing slow start or
congestion avoidance.

14. Fast Retransmit
 Since TCP does not know whether a duplicate ACK
is caused by a lost segment or just a reordering
of segments, it waits for a small number of
duplicate ACKs to be received.
 It is assumed that if there is just a reordering of
the segments, there will be only one or two
duplicate ACKs before the reordered segment is
processed, which will then generate a new ACK.
 If three or more duplicate ACKs are received in a
row, it is a strong indication that a segment has
been lost. TCP then performs a retransmission of
what appears to be the missing segment, without
waiting for a retransmission timer to expire.

15. Fast Recovery
 After fast retransmit sends what
appears to be the missing segment,
congestion avoidance, but not slow
start is performed.
 This is the fast recovery algorithm.
It is an improvement that allows
high throughput under moderate
congestion, especially for large
windows. [2]

16. References
 [1] The Great Idea Finder “Fascinating facts about the
invention of the Traffic Light William L. Potts
“http://www.ideafinder.com/history/inventions/trafficlight.ht
m Date Accessed: 10th February 2008
 [2] D. E. Commer, “Internetworking with TCP/IP, Volume 1:
Principles, protocols, and Architecture, 2006
 [3] W. R. Stevens, “TCP Slow Start, Congestion Avoidance,
Fast Retransmit, and Fast Recovery Algorithms, “RFC 2001,
Jan 1997
 [4] D. E. Commer, “Internetworking with TCP/IP, Volume 1:
Principles, protocols, and Architecture, 2006
 [5] http://www.tutorialsweb.com/networking/tcp-ip/
images/Fig10_SlidingWindow.jpg

17. References - Continue
 [6] V. Jacobson, “Modified TCP congestion
Avoidance Algorithm, “end2end –internet April
30, 1990 ftp://ftp.isi.edu/end2end/end2end-interest-
1990.mail
 [7] D. X. Wei C. Jin, S. H. Low, S. Hedge, “FAST
TCP” Motivation, Architecture, Algorithm,
Performance IEEE Network, 2005
 [8]http://www.isi.edu/nsnam/DIRECTED_RESEAR
CH/DR_WANIDA/DR/JavisInActionSlowStartFrame
.html Date Accessed 15/02/08
 [9]http://www.cs.rice.edu/~amsaha/Papers/Cexa
m/notes/node109.html 15/02/08

18. Thank You for Paying Attention
 Suggestions! and Questions???