1
00:00:06,520 --> 00:00:08,200
- Now let's review an introduction

2
00:00:08,200 --> 00:00:12,150
to Amazon ECS for Kubernetes or EKS.

3
00:00:12,150 --> 00:00:15,560
With Amazon ECS for Kubernetes, we gain

4
00:00:15,560 --> 00:00:19,250
a fully managed control
plane for Kubernetes

5
00:00:19,250 --> 00:00:24,230
and so AWS will create and
manage the master nodes

6
00:00:24,230 --> 00:00:27,780
that make up that control
plane for Kubernetes.

7
00:00:27,780 --> 00:00:31,010
And so keep in mind that
with Kubernetes, you have all

8
00:00:31,010 --> 00:00:33,400
of the Kubernetes specific software

9
00:00:33,400 --> 00:00:35,940
and you also have etcd.

10
00:00:35,940 --> 00:00:39,530
So the Kubernetes control
plane relies on etcd

11
00:00:39,530 --> 00:00:42,930
which is a distributed key value store

12
00:00:42,930 --> 00:00:47,140
and so it's really important
to maintain high availability

13
00:00:47,140 --> 00:00:50,570
of that service and durability in the data

14
00:00:50,570 --> 00:00:54,620
because the the values
in etcd are back to disk.

15
00:00:54,620 --> 00:00:59,080
And so, while AWS creates and
manages those master nodes,

16
00:00:59,080 --> 00:01:02,880
ensuring constant durability
and high availability

17
00:01:02,880 --> 00:01:05,600
for the Kubernetes control plane,

18
00:01:05,600 --> 00:01:09,660
the customer, we the customer
will manage the worker nodes.

19
00:01:09,660 --> 00:01:11,600
And we will take a look at a diagram

20
00:01:11,600 --> 00:01:13,560
on this here in just a moment.

21
00:01:13,560 --> 00:01:17,560
And so, ECS Kubernetes is certified

22
00:01:17,560 --> 00:01:19,780
to be Kubernetes conformant.

23
00:01:19,780 --> 00:01:22,500
It is compatible with
applications designed

24
00:01:22,500 --> 00:01:24,200
for standard Kubernetes.

25
00:01:24,200 --> 00:01:26,600
It essentially is
standard Kubernetes except

26
00:01:26,600 --> 00:01:30,310
that AWS is managing the master nodes

27
00:01:30,310 --> 00:01:31,460
for us.

28
00:01:31,460 --> 00:01:33,290
So, let's take a look here

29
00:01:33,290 --> 00:01:36,400
at a diagram showing the architecture.

30
00:01:36,400 --> 00:01:41,240
So what you can notice
here is that as a customer,

31
00:01:41,240 --> 00:01:46,010
we will have our own VPC,
right so we would create

32
00:01:46,010 --> 00:01:49,230
our own VPC and in that VPC,

33
00:01:49,230 --> 00:01:54,033
we are then responsible
for the worker nodes.

34
00:01:54,903 --> 00:01:55,736
AWS,

35
00:01:57,410 --> 00:01:59,040
Amazon is creating

36
00:01:59,040 --> 00:02:02,880
another VPC specifically for

37
00:02:02,880 --> 00:02:04,420
the master nodes,

38
00:02:04,420 --> 00:02:07,570
right so AWS is managing
those master nodes,

39
00:02:07,570 --> 00:02:09,885
we know that they are there,

40
00:02:09,885 --> 00:02:12,500
but they are

41
00:02:12,500 --> 00:02:17,370
pretty well opaque to us, we
don't really get to see exactly

42
00:02:17,370 --> 00:02:18,830
what's going on behind the scenes

43
00:02:18,830 --> 00:02:21,990
we're just trusting that AWS is managing

44
00:02:21,990 --> 00:02:26,190
the Kubernetes worker
nodes in a highly available

45
00:02:26,190 --> 00:02:28,870
fault tolerant, secure and durable way.

46
00:02:28,870 --> 00:02:30,751
And so as far as the

47
00:02:30,751 --> 00:02:33,260
worker nodes go,

48
00:02:33,260 --> 00:02:35,390
these instances here,

49
00:02:35,390 --> 00:02:39,905
these are up to us to
launch and put in place.

50
00:02:39,905 --> 00:02:42,107
AWS does provide

51
00:02:42,107 --> 00:02:45,053
a machine image that includes

52
00:02:45,053 --> 00:02:48,711
the Kubernetes software
and allows us to connect

53
00:02:48,711 --> 00:02:51,860
the worker nodes to the master nodes

54
00:02:51,860 --> 00:02:56,497
and as you can see here,
these worker nodes will all

55
00:02:56,497 --> 00:03:00,890
essentially communicate
with the master nodes

56
00:03:02,410 --> 00:03:06,070
once they are up and running
but it is our responsibility

57
00:03:06,070 --> 00:03:08,500
to ensure that we're using
the right machine image

58
00:03:08,500 --> 00:03:11,237
for the worker nodes, that we are using

59
00:03:11,237 --> 00:03:13,563
the right security groups.

60
00:03:14,460 --> 00:03:16,562
And then of course, we
do have complete freedom

61
00:03:16,562 --> 00:03:18,104
to use

62
00:03:18,104 --> 00:03:20,620
just about any EC2 instance

63
00:03:20,620 --> 00:03:21,518
we feel is appropriate.

64
00:03:21,518 --> 00:03:24,800
We could use, you know, c5.larges,

65
00:03:24,800 --> 00:03:28,200
we could use m58.xlarges we could use,

66
00:03:28,200 --> 00:03:30,990
you know, our five machines,

67
00:03:30,990 --> 00:03:32,784
we could put a blend of different machines

68
00:03:32,784 --> 00:03:34,318
in as worker nodes,

69
00:03:34,318 --> 00:03:36,450
if our applications you know,

70
00:03:36,450 --> 00:03:39,489
have the ability to run on
different types of machines,

71
00:03:39,489 --> 00:03:40,322
right,

72
00:03:40,322 --> 00:03:42,620
So what we do with their worker nodes

73
00:03:42,620 --> 00:03:43,760
is totally up to us,

74
00:03:43,760 --> 00:03:45,706
but they are our responsibility.

75
00:03:45,706 --> 00:03:49,540
And then of course, as we
deploy containers to Kubernetes

76
00:03:50,400 --> 00:03:51,740
those containers

77
00:03:52,900 --> 00:03:56,880
will be launched in run
here on the worker nodes

78
00:03:56,880 --> 00:04:00,232
and then our users, if we
are running any kind of

79
00:04:00,232 --> 00:04:03,153
application that's
accessible over the internet,

80
00:04:03,153 --> 00:04:06,134
such as API's or web socket applications.

81
00:04:06,134 --> 00:04:08,226
Then of course, our users will come in

82
00:04:08,226 --> 00:04:10,061
through the internet gateway

83
00:04:10,061 --> 00:04:13,062
through a load balancer
and then the load balancer

84
00:04:13,062 --> 00:04:16,900
could load balance to the containers

85
00:04:16,900 --> 00:04:18,140
running on

86
00:04:19,070 --> 00:04:20,417
these worker nodes.

87
00:04:20,417 --> 00:04:23,700
Right, so that's EKS
Elastic Container Service

88
00:04:23,700 --> 00:04:25,010
for Kubernetes,

89
00:04:25,010 --> 00:04:28,803
giving us a fully managed
Kubernetes control plane.