1
00:00:00,000 --> 00:00:05,000
In its basic form, TCP may have a window size of 1

2
00:00:06,000 --> 00:00:10,000
which means that for every segment transmitted by the sender

3
00:00:11,000 --> 00:00:16,000
the receiver sends an acknowledgement for that individual segment.

4
00:00:17,000 --> 00:00:19,000
This however slows down the throughput dramatically

5
00:00:20,000 --> 00:00:23,000
because the sender cannot transmit anymore data

6
00:00:24,000 --> 00:00:29,000
until it to receives acknowledgement of that  single segment transmitted.

7
00:00:30,000 --> 00:00:32,000
The throughput would be very low

8
00:00:33,000 --> 00:00:34,000
depending on the round trip timer

9
00:00:35,000 --> 00:00:38,000
between sending data and receiving the acknowledgement.

10
00:00:39,000 --> 00:00:42,000
TCP however thus allow for greater window sizes

11
00:00:43,000 --> 00:00:44,000
to allow for more segments

12
00:00:45,000 --> 00:00:48,000
to be transmitted before receiving an acknowledgement. 
  
12

13
00:00:49,000 --> 00:00:50,000
The window is the number of data segments

14
00:00:51,000 --> 00:00:52,000
the sender is allowed to send

15
00:00:53,000 --> 00:00:56,000
without getting an acknowledgement from the receiver.

16
00:00:57,000 --> 00:00:58,000
In this case we’ve set it to 1

17
00:00:59,000 --> 00:01:02,000
that means when host A wants to send traffic to  host B

18
00:01:03,000 --> 00:01:07,000
it can send 1 segment because the window size is set to 1

19
00:01:03,000 --> 00:01:07,000
host B, once it receives that segment, sends an acknowledgement.

20
00:01:14,000 --> 00:01:18,000
In this example, let’s assume that host A, send the segment with the sequence number of 1

21
00:01:19,000 --> 00:01:21,000
host B would acknowledge for segment 2.

22
00:01:22,000 --> 00:01:24,000
Host A will then send segment 2

23
00:01:25,000 --> 00:01:29,000
and host B, once successfully receiving that segment

24
00:01:30,000 --> 00:01:32,000
will acknowledge or ACK for segment 3.

25
00:01:33,000 --> 00:01:36,000
Host A will then send segment 3 to host B.

26
00:01:37,000 --> 00:01:41,000
This process will then continue for the duration of the session

27
00:01:42,000 --> 00:01:46,000
this is obviously very reliable however, the throughput is very low.

28
00:01:47,000 --> 00:01:50,000
A would need to buffer outgoing segments

29
00:01:51,000 --> 00:01:54,000
until it receive an acknowledgement for the segment transmitted.

30
00:01:55,000 --> 00:01:57,000
now for argument sake

31
00:01:58,000 --> 00:02:00,000
let’s assume that it takes 1 second for traffic

32
00:04:04,000 --> 00:04:06,000
with the start with the small window size

33
00:04:07,000 --> 00:04:10,000
and then exponentially increase the window size to gauge

34
00:04:11,000 --> 00:04:15,000
the amount of data  that the receiver can receive and what the network can handle.

35
00:04:16,000 --> 00:04:20,000
You’ve probably notice this, when downloading a file from the internet

36
00:04:21,000 --> 00:04:23,000
initially the download speed is slow

37
00:04:24,000 --> 00:04:27,000
but then increases to a certain point over time.

38
00:04:28,000 --> 00:04:31,000
This is because the window size initially are small

39
00:04:32,000 --> 00:04:35,000
but then increases exponentially until a packet is dropped

40
00:04:36,000 --> 00:04:40,000
or the receiving host cannot handle the amount of data it's receiving.

41
00:04:41,000 --> 00:04:43,000
So you’ll notice initially, that the download speed is very slow

42
00:04:44,000 --> 00:04:47,000
increases very quickly and then  gets to a point

43
00:04:48,000 --> 00:04:49,000
and then stays around that speed.

44
00:04:50,000 --> 00:04:51,000
So once again, let’s assume

45
00:04:52,000 --> 00:04:55,000
that the hosts in this example have a fixed window size  of 3.

46
00:04:56,000 --> 00:05:01,000
That means that A can send 3 segments before receiving an acknowledgement.

47
00:05:02,000 --> 00:05:05,000
So in this case host A  sends segment 1,2 and 3

48
00:05:06,000 --> 00:05:10,000
host B acknowledges for segment 4

49
00:05:11,000 --> 00:05:15,000
thus letting A know that it successfully received segment  1,2 and 3

50
00:05:16,000 --> 00:05:21,000
host A then sends segment 4,5 and 6 because it has a fix window size of 3

51
00:05:22,000 --> 00:05:25,000
and host B acknowledges for segment 7

52
00:05:26,000 --> 00:05:31,000
thus letting A know, that it had received segments 4,5 and 6.

53
00:05:32,000 --> 00:05:37,000
Remember with acknowledgements, acknowledge the next segment that you want to received

54
00:05:38,000 --> 00:05:40,000
not the segment that you have already received. 
     
80

55
00:05:41,000 --> 00:05:44,000
So host B does not acknowledge for segment 6

56
00:05:45,000 --> 00:05:46,000
but acknowledges segment 7.

57
00:05:47,000 --> 00:05:52,000
As I've mentioned, a sliding window allows hosts to gauge

58
00:05:53,000 --> 00:05:57,000
the amount of data the receiver can receive and what the network can handle.

59
00:05:58,000 --> 00:06:01,000
So in this example let’s assume that the host are using a sliding window

60
00:06:02,000 --> 00:06:05,000
the way that they determine what the network can handle

61
00:06:06,000 --> 00:06:10,000
is when a packet is dropped by the network, the host will slow down.

62
00:06:11,000 --> 00:06:13,000
This information is covered in a lot more detail

63
00:06:14,000 --> 00:06:17,000
in courses that contain quality of service.

64
00:06:18,000 --> 00:06:20,000
With this course just assume that when a packet gets drop

65
00:06:21,000 --> 00:06:23,000
they reduced their window size dramatically

66
00:06:24,000 --> 00:06:32,000
in brief, the window size is either the window granted to the sender by the receiver

67
00:06:33,000 --> 00:06:38,000
or a calculated window called the congestion window or CWND

68
00:06:39,000 --> 00:06:45,000
the congestion window or CWND is initially set to very low value

69
00:06:46,000 --> 00:06:49,000
at connection establishment and then increases at an exponential rate.

70
00:06:50,000 --> 00:06:53,000
For every lost segment, the congestion window is half

71
00:06:54,000 --> 00:06:57,000
after lost segment has been successfully re-transmitted

72
00:06:58,000 --> 00:07:01,000
the congestion window grows again until it reaches

73
00:07:02,000 --> 00:07:05,000
a value half of the original congestion window

74
00:07:06,000 --> 00:07:11,000
it then slows down its growth, using an algorithm called congestion avoidance.

75
00:07:12,000 --> 00:07:17,000
It does exponentially grows up to half the original congestion window size

76
00:07:18,000 --> 00:07:21,000
and then increases slowly at the linear rate.

77
00:07:22,000 --> 00:07:27,000
In quality of service Weighted Random Early Detection or WRED

78
00:07:28,000 --> 00:07:33,000
can be used improve efficiency of TCP transmissions across the link

79
00:07:34,000 --> 00:07:39,000
as packets are randomly dropped from various flows or various sessions

80
00:07:40,000 --> 00:07:41,000
going across an individual interface

81
00:07:42,000 --> 00:07:45,000
rather than packets from multiple senders being drop at the same time.

82
00:07:46,000 --> 00:07:49,000
this avoid an issue called global synchronization

83
00:07:50,000 --> 00:07:54,000
where packets from multiple TCP sessions are drop at the same time

84
00:07:55,000 --> 00:08:01,000
and therefore multiple host reduce their window size and slowdown at the same time
     
110

85
00:08:02,000 --> 00:08:04,000
and then gradually increase their window size

86
00:08:05,000 --> 00:08:08,000
and therefore their throughput at the same time.

87
00:08:09,000 --> 00:08:14,000
So you have a many hosts slowing down and speeding up at the same time

88
00:08:15,000 --> 00:08:17,000
with WRED some hosts will be slowing down

89
00:08:18,000 --> 00:08:24,000
and other hosts will be increasing their throughput because or random drops.

90
00:08:25,000 --> 00:08:30,000
Please refer to quality of service documentation for more detail about WRED

91
00:08:34,000 --> 00:08:39,000
In this example let’s assume that we start with the initial window size of 3.

92
00:08:40,000 --> 00:08:49,000
So A transmits 3 segments to B, however only segment  1 and 2 arrive at host B

93
00:08:50,000 --> 00:08:52,000
segment 3 goes missing.

94
00:08:53,000 --> 00:09:00,000
Host B may reduce its window size in this example to 2 and acknowledge segment 3.

95
00:09:01,000 --> 00:09:08,000
Remember the acknowledgement is for the next packet that that hosts expects to receive

96
00:09:10,000 --> 00:09:12,000
and because host B did not receive segment 3

97
00:09:13,000 --> 00:09:17,000
it’s acknowledging successful receipt of segment 1 and 2.

98
00:09:18,000 --> 00:09:25,000
Host A will re-transmit segment 3 and in this example send segment 4

99
00:09:26,000 --> 00:09:27,000
but also request a window size of 3.

100
00:09:28,000 --> 00:09:32,000
Both those segment are successfully received by host B

101
00:09:33,000 --> 00:09:35,000
so host B will acknowledge for segment 5

102
00:09:36,000 --> 00:09:39,000
but in this example, still wants a window size of 2.

103
00:09:40,000 --> 00:09:45,000
Host A will only send 2 segments because the negotiated window size is 2

104
00:09:46,000 --> 00:09:51,000
but A may still request that the window size be increase to 3.

105
00:09:52,000 --> 00:09:57,000
With the sliding window, there is dynamic negotiation of a window size

106
00:09:58,000 --> 00:10:02,000
and that window size may change dramatically during a session

107
00:10:03,000 --> 00:10:07,000
depending on what the receiver can process and what the network can handle.