1
00:00:00,240 --> 00:00:01,280
Welcome to the section

2
00:00:01,280 --> 00:00:03,690
on AWS integration and messaging.

3
00:00:03,690 --> 00:00:06,170
So this section is kinda cool because we're going

4
00:00:06,170 --> 00:00:08,630
to see how we can orchestrate stuff

5
00:00:08,630 --> 00:00:11,880
between our different services using middleware.

6
00:00:11,880 --> 00:00:13,530
And so in this section basically,

7
00:00:13,530 --> 00:00:16,100
when we start to deploy multiple applications,

8
00:00:16,100 --> 00:00:18,920
they will inevitably have to communicate with one another.

9
00:00:18,920 --> 00:00:22,680
Okay, your services need to share information, share data.

10
00:00:22,680 --> 00:00:24,380
And so there will be two patterns

11
00:00:24,380 --> 00:00:26,460
of application communication out there.

12
00:00:26,460 --> 00:00:28,460
There's going to be a synchronous communication

13
00:00:28,460 --> 00:00:31,140
so your application will be directly connecting

14
00:00:31,140 --> 00:00:33,510
to another application of yours.

15
00:00:33,510 --> 00:00:36,050
And so for example we have, you know,

16
00:00:36,050 --> 00:00:38,700
we sell something online, and we have a buying service,

17
00:00:38,700 --> 00:00:40,450
and when something is bought, we need

18
00:00:40,450 --> 00:00:42,050
to talk to the shipping service

19
00:00:42,050 --> 00:00:44,440
to send that item that was just bought.

20
00:00:44,440 --> 00:00:45,740
As you can see here,

21
00:00:45,740 --> 00:00:48,020
my buying service and my shipping service,

22
00:00:48,020 --> 00:00:50,160
they're directly connected to one another,

23
00:00:50,160 --> 00:00:52,560
so there is some synchronous communication happening.

24
00:00:52,560 --> 00:00:55,230
My buying service says, hey something happened,

25
00:00:55,230 --> 00:00:57,520
shipping service, do it, okay?

26
00:00:57,520 --> 00:01:01,090
The other type of integration and pattern

27
00:01:01,090 --> 00:01:03,750
is going to be asynchronous or event based.

28
00:01:03,750 --> 00:01:06,090
And so there will be a middleware called

29
00:01:06,090 --> 00:01:07,930
a queue or called something else,

30
00:01:07,930 --> 00:01:10,970
basically that will connect your applications.

31
00:01:10,970 --> 00:01:12,740
So this time the buying service says,

32
00:01:12,740 --> 00:01:15,550
hey, someone bought something,

33
00:01:15,550 --> 00:01:19,130
and so I'm gonna put that into a queue, and that's it.

34
00:01:19,130 --> 00:01:21,430
And the shipping service says, hey queue,

35
00:01:21,430 --> 00:01:23,620
is there something that got bought recently?

36
00:01:23,620 --> 00:01:25,670
And the queue will return that element

37
00:01:25,670 --> 00:01:28,310
and the shipping service can do whatever it wants.

38
00:01:28,310 --> 00:01:30,820
So as you can see here, the buying service

39
00:01:30,820 --> 00:01:33,950
and the shipping service are not directly connected.

40
00:01:33,950 --> 00:01:35,350
There is a queue in between.

41
00:01:35,350 --> 00:01:36,810
And so because they don't directly talk

42
00:01:36,810 --> 00:01:39,930
to one another, this is asynchronous.

43
00:01:39,930 --> 00:01:43,300
Now, synchronous between applications can be

44
00:01:43,300 --> 00:01:44,360
a little bit problematic sometimes,

45
00:01:44,360 --> 00:01:47,500
because if one service overwhelms the other

46
00:01:47,500 --> 00:01:50,560
because there is a sudden spike of purchases or whatever,

47
00:01:50,560 --> 00:01:51,970
it could be a big problem right.

48
00:01:51,970 --> 00:01:53,710
So if you need to encode, for example

49
00:01:53,710 --> 00:01:54,980
we have a video encoding service,

50
00:01:54,980 --> 00:01:58,270
and we need to encode 1,000 videos but usually it's 10.

51
00:01:58,270 --> 00:02:01,120
Well, our encoding service is going to be overwhelmed

52
00:02:01,120 --> 00:02:03,040
and we're going to have outages.

53
00:02:03,040 --> 00:02:05,820
So, when you have these sudden spikes of traffic

54
00:02:05,820 --> 00:02:07,310
or you can't predict anything,

55
00:02:07,310 --> 00:02:10,240
then it's usually better to decouple your application

56
00:02:10,240 --> 00:02:13,150
and have the decoupling layer scale for you.

57
00:02:13,150 --> 00:02:16,870
So in that case that could be SQS for a queue model,

58
00:02:16,870 --> 00:02:19,280
that could be SNS for a pub/sub model,

59
00:02:19,280 --> 00:02:22,240
and that could be Kinesis, if you do real time streaming

60
00:02:22,240 --> 00:02:23,980
and you have like big data.

61
00:02:23,980 --> 00:02:26,540
So we'll learn all about those in this section.

62
00:02:26,540 --> 00:02:28,350
And what we'll learn is that now,

63
00:02:28,350 --> 00:02:30,700
using these three things, our services

64
00:02:30,700 --> 00:02:35,410
can scale independently from SQS, SNS, and Kinesis.

65
00:02:35,410 --> 00:02:38,670
And these three things scale as well really, really, well.

66
00:02:38,670 --> 00:02:40,050
So that's the whole paradigm.

67
00:02:40,050 --> 00:02:41,540
And so we're gonna get started with learning

68
00:02:41,540 --> 00:02:42,960
these three technologies in this lecture.

69
00:02:42,960 --> 00:02:44,560
So, see you in the next lecture.

