1
00:00:00,070 --> 00:00:01,660
Now, let's get into the second kind

2
00:00:01,660 --> 00:00:03,080
of load balancers we'll see

3
00:00:03,080 --> 00:00:05,230
is the application load balancer.

4
00:00:05,230 --> 00:00:07,700
So it's a layer seven only load balancer.

5
00:00:07,700 --> 00:00:09,450
So that means HTTP.

6
00:00:09,450 --> 00:00:10,810
And it allows you to route

7
00:00:10,810 --> 00:00:13,930
to multiple HTTP applications across machines.

8
00:00:13,930 --> 00:00:16,020
And these machines are going to be grouped

9
00:00:16,020 --> 00:00:17,660
in something called a target group.

10
00:00:17,660 --> 00:00:18,760
And it will make a lot of sense

11
00:00:18,760 --> 00:00:20,760
once we get into the hands-on.

12
00:00:20,760 --> 00:00:22,520
It allows you to load balance

13
00:00:22,520 --> 00:00:25,960
to multiple applications on the same EC2 instance.

14
00:00:25,960 --> 00:00:28,770
So using containers and using ECS as we'll see.

15
00:00:28,770 --> 00:00:32,880
And it has support for HTTP/2 and WebSockets.

16
00:00:32,880 --> 00:00:34,170
It supports redirect as well.

17
00:00:34,170 --> 00:00:35,460
So if you want it to redirect traffic

18
00:00:35,460 --> 00:00:37,560
from HTTP to HTTPS automatically,

19
00:00:37,560 --> 00:00:40,610
it could be done at the load balancer level.

20
00:00:40,610 --> 00:00:42,410
It also supports route routing.

21
00:00:42,410 --> 00:00:45,070
So there's routing based on different target groups.

22
00:00:45,070 --> 00:00:46,580
For example, you can route based

23
00:00:46,580 --> 00:00:49,670
on the target path of your URL.

24
00:00:49,670 --> 00:00:54,670
For example, example.com/users and example.com/posts.

25
00:00:54,740 --> 00:00:58,840
/users and /posts are different routes in your URL,

26
00:00:58,840 --> 00:01:00,770
different path, and so you can redirect

27
00:01:00,770 --> 00:01:02,310
these two things into different target groups.

28
00:01:02,310 --> 00:01:04,069
We'll see what that means in a second.

29
00:01:04,069 --> 00:01:06,650
You can also do routing based on the host name of the URL.

30
00:01:06,650 --> 00:01:09,380
So if your load balancer is accessed using

31
00:01:09,380 --> 00:01:13,170
one.example.com or other.example.com,

32
00:01:13,170 --> 00:01:15,970
it could be routed to different target groups,

33
00:01:15,970 --> 00:01:17,820
and you could also be routing based on

34
00:01:17,820 --> 00:01:19,610
the query strings and headers.

35
00:01:19,610 --> 00:01:21,920
So for example, example.com/reserves

36
00:01:21,920 --> 00:01:25,700
and id=123&order=false could be routed

37
00:01:25,700 --> 00:01:27,330
to a different target group.

38
00:01:27,330 --> 00:01:29,700
So ALB, and this is the short for

39
00:01:29,700 --> 00:01:31,170
application load balancers,

40
00:01:31,170 --> 00:01:33,440
they're great when you have micro services

41
00:01:33,440 --> 00:01:35,430
and container-based application.

42
00:01:35,430 --> 00:01:37,800
So as soon as we'll go into learning what docker is

43
00:01:37,800 --> 00:01:40,230
and Amazon ECS, ALB are going to be

44
00:01:40,230 --> 00:01:42,290
the go-to load balancers,

45
00:01:42,290 --> 00:01:44,400
because they have a port mapping features,

46
00:01:44,400 --> 00:01:46,630
which allows you to redirect to a dynamic port

47
00:01:46,630 --> 00:01:48,590
on the ECS instance, and again,

48
00:01:48,590 --> 00:01:51,350
more on that directly into the ECS section.

49
00:01:51,350 --> 00:01:53,290
And in comparison, if we want it to have

50
00:01:53,290 --> 00:01:55,949
multiple applications, okay,

51
00:01:55,949 --> 00:02:00,190
behind multiple, behind a classic load balancer,

52
00:02:00,190 --> 00:02:03,340
we would have to have multiple classic load balancer.

53
00:02:03,340 --> 00:02:06,320
We need to actually have one per application,

54
00:02:06,320 --> 00:02:07,870
whereas with load balancers, we're able to have

55
00:02:07,870 --> 00:02:11,340
one application load balancer in front of many applications.

56
00:02:11,340 --> 00:02:14,130
So maybe a graph will help.

57
00:02:14,130 --> 00:02:16,160
So we have our external application load balancer.

58
00:02:16,160 --> 00:02:18,530
It's public facing, and behind it,

59
00:02:18,530 --> 00:02:22,430
we have our first target group made of EC2 instances.

60
00:02:22,430 --> 00:02:26,560
And this one is going to be routing for the route /user.

61
00:02:26,560 --> 00:02:28,130
And we have a second target group

62
00:02:28,130 --> 00:02:30,262
made of EC2 instances again,

63
00:02:30,262 --> 00:02:32,360
and this one is going to be our search application,

64
00:02:32,360 --> 00:02:34,730
and there's going to be a health check as well with it.

65
00:02:34,730 --> 00:02:37,130
And it's going to be routed through rules

66
00:02:37,130 --> 00:02:39,650
for the /search routes.

67
00:02:39,650 --> 00:02:42,070
So as you can see here, we have two independent

68
00:02:42,070 --> 00:02:44,210
microservices that do different things.

69
00:02:44,210 --> 00:02:46,137
The first one is the user application.

70
00:02:46,137 --> 00:02:48,340
The second one is a search application.

71
00:02:48,340 --> 00:02:51,530
But they're behind the same application load balancer,

72
00:02:51,530 --> 00:02:53,960
which knows how to intelligently route

73
00:02:53,960 --> 00:02:56,240
to these target groups based on the route

74
00:02:56,240 --> 00:02:58,440
that is being used in the URL.

75
00:02:58,440 --> 00:03:00,790
So the target groups, what are they?

76
00:03:00,790 --> 00:03:03,380
The first one is that they can be EC2 instances,

77
00:03:03,380 --> 00:03:05,820
and they can be managed as we'll see very, very soon.

78
00:03:05,820 --> 00:03:08,494
They can be managed by Auto Scaling Group.

79
00:03:08,494 --> 00:03:09,820
It could be ECS tasks

80
00:03:09,820 --> 00:03:11,760
and we'll see this in the ECS section.

81
00:03:11,760 --> 00:03:13,210
It could be Lambda functions,

82
00:03:13,210 --> 00:03:14,980
and this is something that's not very known.

83
00:03:14,980 --> 00:03:17,990
So application load balancers can be in front

84
00:03:17,990 --> 00:03:20,220
of lambda functions, and we'll see what

85
00:03:20,220 --> 00:03:22,260
Lambda functions are in the future section,

86
00:03:22,260 --> 00:03:23,810
but they're the base of everything

87
00:03:23,810 --> 00:03:26,420
called serverless in AWS.

88
00:03:26,420 --> 00:03:30,480
And finally, it can be a front to IP addresses,

89
00:03:30,480 --> 00:03:32,520
and they must be private IP addresses.

90
00:03:32,520 --> 00:03:34,800
So ALB can route to multiple target groups,

91
00:03:34,800 --> 00:03:36,520
and the health checks are going to be done

92
00:03:36,520 --> 00:03:38,690
at the target group level.

93
00:03:38,690 --> 00:03:40,570
So let's take another example.

94
00:03:40,570 --> 00:03:43,040
And we have an ALB and we have two target groups.

95
00:03:43,040 --> 00:03:46,540
The first one is based on AWS with EC2 instances,

96
00:03:46,540 --> 00:03:48,410
and the second one is going to have

97
00:03:48,410 --> 00:03:52,250
private servers on premises, so in our own data center.

98
00:03:52,250 --> 00:03:55,180
And so for the target groups, you'd be able to exist.

99
00:03:55,180 --> 00:03:58,570
We need to specify the private IP of our servers

100
00:03:58,570 --> 00:04:01,460
into the target group, choose for them to be registered.

101
00:04:01,460 --> 00:04:04,090
And now say we have an application

102
00:04:04,090 --> 00:04:07,370
that is serving as request through the ALB.

103
00:04:07,370 --> 00:04:10,850
And what we want is to send the first

104
00:04:10,850 --> 00:04:13,400
to the first target group on the mobile-based traffic,

105
00:04:13,400 --> 00:04:16,660
and to the second target group on the desktop-based traffic.

106
00:04:16,660 --> 00:04:18,510
For this, we can use, for example,

107
00:04:18,510 --> 00:04:20,589
a query string or parameters routing.

108
00:04:20,589 --> 00:04:24,500
So if, in the URL that the clients are trying to use,

109
00:04:24,500 --> 00:04:27,500
there is a ?Platform=Mobile,

110
00:04:27,500 --> 00:04:30,850
we can write a rule in the ALB redirection rules,

111
00:04:30,850 --> 00:04:34,900
the routing rules, to redirect to the first target group.

112
00:04:34,900 --> 00:04:37,540
And if you have ?Platform=Desktop,

113
00:04:37,540 --> 00:04:40,160
so this is a query string or parameter,

114
00:04:40,160 --> 00:04:42,130
then we can say it to redirect

115
00:04:42,130 --> 00:04:43,730
to the target group number two.

116
00:04:43,730 --> 00:04:44,690
I don't know where you would do this,

117
00:04:44,690 --> 00:04:48,120
but this is just an example I'm providing you, okay?

118
00:04:48,120 --> 00:04:50,820
So good to know before we go into the hands-on,

119
00:04:50,820 --> 00:04:53,430
the first one is that you also get a fixed host name

120
00:04:53,430 --> 00:04:55,000
with your application load balancers,

121
00:04:55,000 --> 00:04:56,720
just like the classic one.

122
00:04:56,720 --> 00:04:58,870
The application servers don't see the IP

123
00:04:58,870 --> 00:05:00,230
of the client directly.

124
00:05:00,230 --> 00:05:02,600
The true IP of the clients is going to be inserted

125
00:05:02,600 --> 00:05:06,420
instead in the header called X-Forwarded-For.

126
00:05:06,420 --> 00:05:09,340
And so you can also get the port using X-Forwarded-Ports

127
00:05:09,340 --> 00:05:12,000
and the protocol used, being used,

128
00:05:12,000 --> 00:05:14,650
using X-Forwarded-Proto.

129
00:05:14,650 --> 00:05:17,810
And so what that means is that our client IP,

130
00:05:17,810 --> 00:05:21,260
which is 12.34.56.78 is directly talking

131
00:05:21,260 --> 00:05:23,930
to our load balancer, which performs something

132
00:05:23,930 --> 00:05:26,440
called a connection termination.

133
00:05:26,440 --> 00:05:29,270
And when your load balancer talks to your EC2 instance,

134
00:05:29,270 --> 00:05:31,240
it's going to use the load balancer IP,

135
00:05:31,240 --> 00:05:34,140
which is a private IP into your EC2 instance.

136
00:05:34,140 --> 00:05:37,310
And so for the EC2 instance to know the client IP,

137
00:05:37,310 --> 00:05:39,750
it will have to look at these extra headers

138
00:05:39,750 --> 00:05:41,880
in your HTTP request, which are called

139
00:05:41,880 --> 00:05:44,880
X-Forwarded-Port and Proto.

140
00:05:44,880 --> 00:05:45,770
All right, that's it.

141
00:05:45,770 --> 00:05:46,603
Now, let's go into the hands-on,

142
00:05:46,603 --> 00:05:48,900
and create our first application load balancer.