1
1

00:00:00,590  -->  00:00:03,560
<v ->Why do we need quality of service or QoS?</v>
2

2

00:00:03,560  -->  00:00:06,530
Well, nowadays we operate converge networks,
3

3

00:00:06,530  -->  00:00:09,380
which means all of our networks are carrying voice, data
4

4

00:00:09,380  -->  00:00:11,970
and video content over the same wire.
5

5

00:00:11,970  -->  00:00:14,260
We don't have them all separate out like we used to.
6

6

00:00:14,260  -->  00:00:16,900
We used to have networks for phones and ones for data
7

7

00:00:16,900  -->  00:00:18,040
and ones for video,
8

8

00:00:18,040  -->  00:00:21,840
but now everything's riding over the same IP networks.
9

9

00:00:21,840  -->  00:00:24,650
So, because of this convergence of mediums,
10

10

00:00:24,650  -->  00:00:25,670
we have these networks
11

11

00:00:25,670  -->  00:00:27,970
that now have a high level availability
12

12

00:00:27,970  -->  00:00:29,340
to ensure proper delivery
13

13

00:00:29,340  -->  00:00:31,190
over all of these different medians,
14

14

00:00:31,190  -->  00:00:32,460
because we want a phone to work
15

15

00:00:32,460  -->  00:00:34,430
every time we pick up the phone, right?
16

16

00:00:34,430  -->  00:00:37,460
Well, by using QoS, we can optimize our network
17

17

00:00:37,460  -->  00:00:40,520
to efficiently utilize all the bandwidth at the right time
18

18

00:00:40,520  -->  00:00:42,400
to deliver the right service to our users
19

19

00:00:42,400  -->  00:00:45,040
and give a success and cost savings.
20

20

00:00:45,040  -->  00:00:47,260
Now, we want to have an excellent quality of service,
21

21

00:00:47,260  -->  00:00:49,090
an excellent service for our customers,
22

22

00:00:49,090  -->  00:00:52,160
and that's what we're going to start doing by using QoS.
23

23

00:00:52,160  -->  00:00:54,640
So what exactly is QoS?
24

24

00:00:54,640  -->  00:00:56,440
Well, quality of service enables us
25

25

00:00:56,440  -->  00:00:59,050
to strategically optimize our network performance
26

26

00:00:59,050  -->  00:01:01,090
based on different types of traffic.
27

27

00:01:01,090  -->  00:01:03,020
Previously, we talked about the fact
28

28

00:01:03,020  -->  00:01:05,570
that we want to categorize our different traffic types.
29

29

00:01:05,570  -->  00:01:08,890
I might have web traffic and voice traffic and video traffic
30

30

00:01:08,890  -->  00:01:10,060
and email traffic.
31

31

00:01:10,060  -->  00:01:11,380
And by categorizing it
32

32

00:01:11,380  -->  00:01:13,660
and identifying these different types of traffic,
33

33

00:01:13,660  -->  00:01:17,120
I can then prioritize that traffic and route it differently.
34

34

00:01:17,120  -->  00:01:19,540
So I might determine how much bandwidth is required
35

35

00:01:19,540  -->  00:01:21,120
for each of those types of traffic.
36

36

00:01:21,120  -->  00:01:23,830
And I can efficiently use my wide area network links
37

37

00:01:23,830  -->  00:01:27,070
and all that bandwidth available, for maximum utilization,
38

38

00:01:27,070  -->  00:01:29,660
and save me bandwidth costs over time.
39

39

00:01:29,660  -->  00:01:31,560
This can help me identify
40

40

00:01:31,560  -->  00:01:32,490
the types of traffic that I should drop
41

41

00:01:32,490  -->  00:01:34,390
whenever there's going to be some kind of congestion,
42

42

00:01:34,390  -->  00:01:35,960
because if you look at the average load,
43

43

00:01:35,960  -->  00:01:38,480
there's always going to be some peaks and some valleys.
44

44

00:01:38,480  -->  00:01:40,440
And so we want to be able to figure that out.
45

45

00:01:40,440  -->  00:01:41,770
So for example, here on the screen,
46

46

00:01:41,770  -->  00:01:44,130
you can see the peaks and the valleys.
47

47

00:01:44,130  -->  00:01:45,370
The peaks over time,
48

48

00:01:45,370  -->  00:01:46,950
and we need to be able to categorize things
49

49

00:01:46,950  -->  00:01:49,150
to fit within our bandwidth limitations.
50

50

00:01:49,150  -->  00:01:51,320
So for example, if we have things like VoIP,
51

51

00:01:51,320  -->  00:01:53,800
or voice over IP, or video service,
52

52

00:01:53,800  -->  00:01:55,400
they need to have a higher priority,
53

53

00:01:55,400  -->  00:01:57,150
because if I'm talking to you on a phone,
54

54

00:01:57,150  -->  00:01:59,150
I don't want a high amount of latency.
55

55

00:01:59,150  -->  00:02:01,320
From checking my bank balance, for instance, though,
56

56

00:02:01,320  -->  00:02:04,190
I can wait another half a second for the web page to load.
57

57

00:02:04,190  -->  00:02:06,770
From listening to you talk, that half a second delay
58

58

00:02:06,770  -->  00:02:08,140
starts sending like an echo,
59

59

00:02:08,140  -->  00:02:10,090
and it gives me a horrible service level.
60

60

00:02:10,090  -->  00:02:11,810
So we want to be able to solve that,
61

61

00:02:11,810  -->  00:02:14,610
and to do that, we use quality of service.
62

62

00:02:14,610  -->  00:02:17,330
Now there are different categories of quality of service.
63

63

00:02:17,330  -->  00:02:21,190
There are three big ones known as delay, jitter and drops.
64

64

00:02:21,190  -->  00:02:22,600
When I talk about delay,
65

65

00:02:22,600  -->  00:02:24,430
this happens when you look at the time
66

66

00:02:24,430  -->  00:02:27,510
that a packet travels from the source to the destination,
67

67

00:02:27,510  -->  00:02:29,300
this is measured in milliseconds,
68

68

00:02:29,300  -->  00:02:31,760
and it's not a big deal if you're dealing with data traffic,
69

69

00:02:31,760  -->  00:02:33,860
but if you're dealing with voice or video,
70

70

00:02:33,860  -->  00:02:35,870
delay is an especially big thing,
71

71

00:02:35,870  -->  00:02:38,110
especially if you're doing things live,
72

72

00:02:38,110  -->  00:02:40,330
like talking on the phone or doing a live stream,
73

73

00:02:40,330  -->  00:02:42,000
or something like that.
74

74

00:02:42,000  -->  00:02:44,970
Now, jitter is an uneven arrival of packets,
75

75

00:02:44,970  -->  00:02:47,800
and this is especially bad in voiceover IP traffic,
76

76

00:02:47,800  -->  00:02:49,640
because you're using something like UDP.
77

77

00:02:49,640  -->  00:02:52,800
And so if I sing something to you, like, "my name is Jason,"
78

78

00:02:52,800  -->  00:02:54,840
and you got "Jason my name is,"
79

79

00:02:54,840  -->  00:02:56,530
it sounds kind of weird, right?
80

80

00:02:56,530  -->  00:02:58,490
Now, usually it's not big chunks like that,
81

81

00:02:58,490  -->  00:02:59,890
but instead it's little bits
82

82

00:02:59,890  -->  00:03:01,990
and you'll hear these glick and glock sounds
83

83

00:03:01,990  -->  00:03:04,610
that make it jumble up because of that jitter.
84

84

00:03:04,610  -->  00:03:07,350
And this really sounds bad, and it's a bad user experience
85

85

00:03:07,350  -->  00:03:09,130
if you're using voiceover IP.
86

86

00:03:09,130  -->  00:03:10,980
And so jitter is a really bad thing
87

87

00:03:10,980  -->  00:03:12,940
when you're dealing with voice and video.
88

88

00:03:12,940  -->  00:03:15,630
Now, the third thing we have is what's known as a drop.
89

89

00:03:15,630  -->  00:03:18,350
Drops are going to occur during network congestion.
90

90

00:03:18,350  -->  00:03:20,180
When the network becomes too congested,
91

91

00:03:20,180  -->  00:03:22,260
the router simply can't keep up with demand,
92

92

00:03:22,260  -->  00:03:23,860
and the queue starts overflowing,
93

93

00:03:23,860  -->  00:03:25,710
and it'll start dropping packets.
94

94

00:03:25,710  -->  00:03:27,630
This is the way it deals with packet loss,
95

95

00:03:27,630  -->  00:03:30,530
and if you're using TCP, it'll just send it again.
96

96

00:03:30,530  -->  00:03:33,950
But again, if I'm dealing with VoIP, VoIP is usually UDP.
97

97

00:03:33,950  -->  00:03:34,890
And so if we're talking
98

98

00:03:34,890  -->  00:03:36,760
and all of a sudden my voice cuts out like that,
99

99

00:03:36,760  -->  00:03:38,100
that would be bad too.
100

100

00:03:38,100  -->  00:03:41,140
That's why we don't want to have packet drop on a VoIP call.
101

101

00:03:41,140  -->  00:03:43,400
And so we want to make sure that that doesn't happen.
102

102

00:03:43,400  -->  00:03:45,720
These network drops are something that can be avoided
103

103

00:03:45,720  -->  00:03:48,470
by doing the proper quality of service as well.
104

104

00:03:48,470  -->  00:03:49,620
So when we deal with this,
105

105

00:03:49,620  -->  00:03:52,170
we have to think about effective bandwidth.
106

106

00:03:52,170  -->  00:03:53,630
What is our effective bandwidth?
107

107

00:03:53,630  -->  00:03:55,260
This is an important concept.
108

108

00:03:55,260  -->  00:03:57,360
So let's look at this client and this server.
109

109

00:03:57,360  -->  00:03:59,100
There's probably a lot more to this network
110

110

00:03:59,100  -->  00:04:00,410
than what I'm showing you here on the screen,
111

111

00:04:00,410  -->  00:04:02,960
but I've simplified it down for this example.
112

112

00:04:02,960  -->  00:04:04,880
Here, you can see I have my client on the left,
113

113

00:04:04,880  -->  00:04:06,580
and he wants to talk to the server.
114

114

00:04:06,580  -->  00:04:07,940
So he goes up through the switch,
115

115

00:04:07,940  -->  00:04:10,850
which uses 100 megabit per second Cat-5 cable.
116

116

00:04:10,850  -->  00:04:12,280
Then he goes through a WAN link
117

117

00:04:12,280  -->  00:04:15,050
over a 256 kilobit per second connection
118

118

00:04:15,050  -->  00:04:17,240
because he's using an old DSL line.
119

119

00:04:17,240  -->  00:04:20,830
Then that connects from that ISP over a T1 connection
120

120

00:04:20,830  -->  00:04:22,130
to another router.
121

121

00:04:22,130  -->  00:04:25,580
That router connects to an E1 connection to another router.
122

122

00:04:25,580  -->  00:04:27,900
And from that router, it goes down a WAN link
123

123

00:04:27,900  -->  00:04:30,510
over a 512 kilobit per second connection,
124

124

00:04:30,510  -->  00:04:33,490
and then down to a switch with a gigabit connection,
125

125

00:04:33,490  -->  00:04:34,810
down to the server.
126

126

00:04:34,810  -->  00:04:37,460
Now, what is my effective bandwidth?
127

127

00:04:37,460  -->  00:04:40,370
Well, it's 256 kilobits per second,
128

128

00:04:40,370  -->  00:04:42,930
because no matter how fast any of the other links are,
129

129

00:04:42,930  -->  00:04:45,600
whatever the lowest link is inside of this connection,
130

130

00:04:45,600  -->  00:04:47,720
that is going to be your effective bandwidth.
131

131

00:04:47,720  -->  00:04:49,680
So we talked about quality of service categories,
132

132

00:04:49,680  -->  00:04:51,560
in our next lesson, we're going to be talking about
133

133

00:04:51,560  -->  00:04:53,280
how we can alleviate this problem
134

134

00:04:53,280  -->  00:04:56,280
of this effective bandwidth, and try to get more out of it,
135

135

00:04:56,280  -->  00:04:57,113
because we need to be able
136

136

00:04:57,113  -->  00:05:00,370
to increase our available bandwidth, but in this example,
137

137

00:05:00,370  -->  00:05:02,620
we're limited to 256 kilobits,
138

138

00:05:02,620  -->  00:05:05,100
which is going to be really, really slow for us.
139

139

00:05:05,100  -->  00:05:07,070
Now, I like to think about effective bandwidth
140

140

00:05:07,070  -->  00:05:08,810
like water flowing through pipes.
141

141

00:05:08,810  -->  00:05:11,090
I can have big pipes and I can have little pipes.
142

142

00:05:11,090  -->  00:05:12,170
And if I have little pipes,
143

143

00:05:12,170  -->  00:05:14,250
I'm going to get less water per second through it
144

144

00:05:14,250  -->  00:05:15,950
than if I have a really big pipe.
145

145

00:05:15,950  -->  00:05:18,210
And so this is the idea, if you think about a big funnel,
146

146

00:05:18,210  -->  00:05:20,430
it can start to back up on us, right?
147

147

00:05:20,430  -->  00:05:21,370
That's the concept,
148

148

00:05:21,370  -->  00:05:23,140
And we have to figure out how we can fix that
149

149

00:05:23,140  -->  00:05:25,140
by using quality of service effectively,
150

150

00:05:25,140  -->  00:05:27,590
which we're going to discuss more in the next video.
