1
00:00:07,020 --> 00:00:08,439
- Now let's review an introduction

2
00:00:08,439 --> 00:00:12,123
to Amazon Relational
Database Service or RDS.

3
00:00:13,010 --> 00:00:17,320
Now before we get into
the details of RDS itself,

4
00:00:17,320 --> 00:00:20,554
let's first take a look at what it means

5
00:00:20,554 --> 00:00:23,923
to manage a relational database service.

6
00:00:25,509 --> 00:00:29,300
So as you can see, managing
a Relational Database

7
00:00:29,300 --> 00:00:31,410
on Premises requires a lot.

8
00:00:31,410 --> 00:00:34,457
There is a lot involved
and of course your DBAs

9
00:00:34,457 --> 00:00:37,530
are probably not doing
all of these things.

10
00:00:37,530 --> 00:00:39,880
Uh, you're going to have different people

11
00:00:39,880 --> 00:00:41,130
at different levels.

12
00:00:41,130 --> 00:00:43,600
You know different levels
of hardware and software.

13
00:00:43,600 --> 00:00:46,320
In fact you probably have
different teams of people,

14
00:00:46,320 --> 00:00:49,310
right? But you can see that on Premises

15
00:00:49,310 --> 00:00:52,500
you are responsible. Your organization

16
00:00:52,500 --> 00:00:55,300
is responsible for
providing of course the data

17
00:00:55,300 --> 00:00:57,710
center in the Power and Cooling, uh

18
00:00:57,710 --> 00:00:58,830
that goes into that.

19
00:00:58,830 --> 00:01:01,520
You are responsible for building out,

20
00:01:01,520 --> 00:01:04,060
maintaining, and securing the network

21
00:01:04,060 --> 00:01:05,540
that allows, you know,

22
00:01:05,540 --> 00:01:08,660
all of these systems to communicate.

23
00:01:08,660 --> 00:01:10,670
You know, of course you
have a rack and a chassis

24
00:01:10,670 --> 00:01:11,950
that need to be put in place.

25
00:01:11,950 --> 00:01:14,560
You have CPU, memory, hard disk

26
00:01:14,560 --> 00:01:16,810
that need to be purchased and maintained

27
00:01:16,810 --> 00:01:18,740
and upgraded over time.

28
00:01:18,740 --> 00:01:20,740
And then of course you
have an operating system

29
00:01:20,740 --> 00:01:23,500
that needs to be installed
and patched and maintained.

30
00:01:23,500 --> 00:01:26,580
You have database engine
that needs to be installed

31
00:01:26,580 --> 00:01:28,190
and maintained. Right?

32
00:01:28,190 --> 00:01:31,070
There is a fair amount of
work that goes into that

33
00:01:31,070 --> 00:01:33,440
to ensure that you are always operating

34
00:01:33,440 --> 00:01:35,630
in a secure environment.

35
00:01:35,630 --> 00:01:36,910
And then of course,

36
00:01:36,910 --> 00:01:40,010
not only do you you
need to perform backups,

37
00:01:40,010 --> 00:01:42,620
it's also your responsibility to ensure

38
00:01:42,620 --> 00:01:44,840
that those backups are

39
00:01:44,840 --> 00:01:45,673
are being written to a system

40
00:01:45,673 --> 00:01:49,560
that itself is highly durable.

41
00:01:49,560 --> 00:01:52,292
And that you have multiple
copies of those backups

42
00:01:52,292 --> 00:01:53,330
(emphasizes) and

43
00:01:53,330 --> 00:01:55,810
that you have a

44
00:01:55,810 --> 00:01:59,650
reliable way of restoring
from those backups.

45
00:01:59,650 --> 00:02:01,690
And then of course you can

46
00:02:01,690 --> 00:02:02,860
write your SQL Queries.

47
00:02:02,860 --> 00:02:04,490
So if your DBA

48
00:02:04,490 --> 00:02:06,590
is burdened by backups

49
00:02:06,590 --> 00:02:08,120
and database engine updates

50
00:02:08,120 --> 00:02:09,894
and operating system

51
00:02:09,894 --> 00:02:10,800
(stutters)

52
00:02:10,800 --> 00:02:15,020
details, then they have a
lot more on their plate.

53
00:02:15,020 --> 00:02:18,950
Now, if we were to
compare this with running

54
00:02:18,950 --> 00:02:21,360
a database on EC2,

55
00:02:21,360 --> 00:02:22,560
then of course

56
00:02:22,560 --> 00:02:27,240
AWS is managing the
physical infrastructure.

57
00:02:27,240 --> 00:02:29,710
We offload the physical
infrastructure burdens

58
00:02:29,710 --> 00:02:33,100
to AWS so that they
are now managing power,

59
00:02:33,100 --> 00:02:36,490
cooling, network, chassis, CPU and so on.

60
00:02:36,490 --> 00:02:39,150
It is still our responsibility,
and this is true

61
00:02:39,150 --> 00:02:41,510
for other things that'll
we'll talk about later.

62
00:02:41,510 --> 00:02:43,280
As far as the operating system goes,

63
00:02:43,280 --> 00:02:46,830
it is still your
responsibility to ensure that

64
00:02:46,830 --> 00:02:49,980
that operating system remains patched

65
00:02:49,980 --> 00:02:51,960
as new security updates come out,

66
00:02:51,960 --> 00:02:54,960
as new versions of an
operating system come out

67
00:02:54,960 --> 00:02:56,690
even though its running on EC2

68
00:02:56,690 --> 00:02:58,730
and even though its
Amazon's responsibility

69
00:02:58,730 --> 00:03:00,770
to install the operating system,

70
00:03:00,770 --> 00:03:03,630
it is your responsibility
to keep it up-to-date.

71
00:03:03,630 --> 00:03:05,240
If you need to go from, you know,

72
00:03:05,240 --> 00:03:07,260
one version to a later version,

73
00:03:07,260 --> 00:03:09,030
then it's up to you to do that

74
00:03:09,030 --> 00:03:10,755
and because

75
00:03:10,755 --> 00:03:11,588
(pauses)

76
00:03:11,588 --> 00:03:15,630
EC2 is based on virtual machines

77
00:03:15,630 --> 00:03:17,870
and we have machine images,

78
00:03:17,870 --> 00:03:21,410
it becomes very easy for
us to automate the process

79
00:03:21,410 --> 00:03:23,690
of updating an operating system.

80
00:03:23,690 --> 00:03:28,010
And considering that EC2
instances are ultimately

81
00:03:28,010 --> 00:03:32,100
disposable, we simply
throw the instance away

82
00:03:32,100 --> 00:03:34,730
and replace it with a new instance, so

83
00:03:34,730 --> 00:03:38,880
the process of maintaining
an operating system

84
00:03:38,880 --> 00:03:41,330
actually becomes much more efficient,

85
00:03:41,330 --> 00:03:45,250
much more fluid than it was on premises.

86
00:03:45,250 --> 00:03:48,710
We're not locked into a
particular set of hardware.

87
00:03:48,710 --> 00:03:51,160
We have the ability to replace things

88
00:03:51,160 --> 00:03:53,650
much easier within EC2.

89
00:03:53,650 --> 00:03:54,810
And the same thing could apply

90
00:03:54,810 --> 00:03:56,350
to the database engine as well

91
00:03:56,350 --> 00:03:59,590
If you want to go from
one engine to another,

92
00:03:59,590 --> 00:04:03,090
you could create, you
could do all of that work

93
00:04:03,090 --> 00:04:05,499
on some other machine create and

94
00:04:05,499 --> 00:04:06,332
(stutters)

95
00:04:06,332 --> 00:04:10,070
once all of that... you're
satisfied with that arrangement

96
00:04:10,070 --> 00:04:12,320
then you could create a machine image

97
00:04:12,320 --> 00:04:15,773
and replace the existing
machine very easily.

98
00:04:17,100 --> 00:04:19,460
Backups are also still your responsibility

99
00:04:19,460 --> 00:04:21,690
but the benefit of having them on EC2,

100
00:04:21,690 --> 00:04:24,090
especially if you're using something like

101
00:04:24,090 --> 00:04:28,540
the Elastic Block Store, is that we can

102
00:04:28,540 --> 00:04:31,270
we can very easily take sanapshots

103
00:04:31,270 --> 00:04:32,730
of those block store volumes

104
00:04:32,730 --> 00:04:35,910
and know that those snapshots are stored

105
00:04:35,910 --> 00:04:38,130
in S3 that we already have

106
00:04:38,130 --> 00:04:42,240
a very durable place
to write those backups

107
00:04:42,240 --> 00:04:45,740
and that place is not
managed by us. Right?

108
00:04:45,740 --> 00:04:50,740
And so we have far fewer
responsibilities by running

109
00:04:50,760 --> 00:04:53,640
a relational database on EC2.

110
00:04:53,640 --> 00:04:58,640
Now let's compare all of
this to running on RDFs.

111
00:04:58,960 --> 00:05:03,960
With RDFs, as you can
see we are offloading

112
00:05:04,010 --> 00:05:08,700
all of these operational
burdens to Amazon.

113
00:05:08,700 --> 00:05:12,240
Amazon now takes over not only installing

114
00:05:12,240 --> 00:05:13,190
the operating system

115
00:05:13,190 --> 00:05:15,560
but also patching and updating

116
00:05:15,560 --> 00:05:19,000
the operating system of that instance.

117
00:05:19,000 --> 00:05:22,210
Amazon also manages the upgrading

118
00:05:22,210 --> 00:05:23,870
of the database engine itself

119
00:05:23,870 --> 00:05:25,960
and performing backups.

120
00:05:25,960 --> 00:05:28,550
And so as an organization

121
00:05:28,550 --> 00:05:31,850
we get to focus on our data.

122
00:05:31,850 --> 00:05:35,590
We get to focus on optimizing our schemas

123
00:05:35,590 --> 00:05:37,880
and optimizing our query statements

124
00:05:39,047 --> 00:05:42,800
and offloading the
operational burdens to AWS.

125
00:05:42,800 --> 00:05:45,660
Okay so the benefits of RDS would be

126
00:05:45,660 --> 00:05:48,550
that we have a reduced operational burden.

127
00:05:48,550 --> 00:05:51,960
As we've seen we can
offload a lot of those

128
00:05:51,960 --> 00:05:55,757
underlying physical and
software burdens to AWS.

129
00:05:57,330 --> 00:06:00,060
We also get our choice of

130
00:06:00,060 --> 00:06:03,023
a number of popular engines such as MySQL,

131
00:06:03,870 --> 00:06:07,130
SQL Server, Oracle, PostgreSQL, MariaDB

132
00:06:07,130 --> 00:06:11,160
and of course Amazon Aurora
is a drop-in replacement

133
00:06:11,160 --> 00:06:14,583
for later versions of
MySQL and PostgreSQL.

134
00:06:16,060 --> 00:06:19,940
We also have a very easy
ability to create Read replicas.

135
00:06:19,940 --> 00:06:22,190
There are plenty of situations were

136
00:06:22,190 --> 00:06:23,881
perhaps you are doing

137
00:06:23,881 --> 00:06:24,714
(stutters)

138
00:06:24,714 --> 00:06:26,930
a lot of reads and

139
00:06:26,930 --> 00:06:29,180
and those operations are

140
00:06:29,180 --> 00:06:31,880
maybe becoming a burden on the primary.

141
00:06:31,880 --> 00:06:35,810
And so by offloading reads to, you know,

142
00:06:35,810 --> 00:06:38,250
one or more servers and then allowing

143
00:06:38,250 --> 00:06:40,560
the primary to handle only writes,

144
00:06:40,560 --> 00:06:43,730
you can improve the performance
of the entire system.

145
00:06:43,730 --> 00:06:47,910
So again we have very easy
creation of read replicas

146
00:06:47,910 --> 00:06:51,090
where AWS handle the
creation, the maintenance

147
00:06:51,090 --> 00:06:54,470
of those actual servers, as well as

148
00:06:54,470 --> 00:06:56,513
ongoing synchronous replication.

149
00:06:57,770 --> 00:07:01,260
We also have a number of
things that are automated

150
00:07:01,260 --> 00:07:04,300
by RDS servers such as backups,

151
00:07:04,300 --> 00:07:05,983
operating system pa... um patches,

152
00:07:06,960 --> 00:07:08,760
database engine updates.

153
00:07:08,760 --> 00:07:12,130
So if you wanted to go
from, you know, for example

154
00:07:12,130 --> 00:07:14,970
MySQL five-five to five-six, you know,

155
00:07:14,970 --> 00:07:17,970
minor version updates, then those things

156
00:07:17,970 --> 00:07:20,260
can be done automatically for you

157
00:07:20,260 --> 00:07:22,540
and you get to control the window,

158
00:07:22,540 --> 00:07:24,930
the maintenance window in
which those things happen.

159
00:07:24,930 --> 00:07:27,960
Of course you may want
that to happen at one AM

160
00:07:27,960 --> 00:07:30,400
or five AM or even noon.

161
00:07:30,400 --> 00:07:33,800
Whenever...whenever you
determine that you have

162
00:07:33,800 --> 00:07:36,830
for example, lower levels of traffic.

163
00:07:36,830 --> 00:07:39,070
And then of course one
last thing to talk about

164
00:07:39,070 --> 00:07:40,800
with RDS is that we also have

165
00:07:40,800 --> 00:07:43,570
a very easy ability to failover,

166
00:07:43,570 --> 00:07:45,900
if you opt in to the right option

167
00:07:45,900 --> 00:07:48,780
such as a multi AZ deployment, then

168
00:07:49,798 --> 00:07:53,980
AWS can manage failover automatically.

169
00:07:53,980 --> 00:07:57,120
Whereas you can have
not only just a primary

170
00:07:57,120 --> 00:07:59,180
but you would also have a secondary

171
00:07:59,180 --> 00:08:03,620
that AWS is also managing
with synchronous replication.

172
00:08:03,620 --> 00:08:06,960
And if the primary fails,
RDS can automatically

173
00:08:06,960 --> 00:08:09,193
fail over to that secondary,

174
00:08:10,720 --> 00:08:12,450
right? So again,

175
00:08:12,450 --> 00:08:13,530
plenty of benefits

176
00:08:13,530 --> 00:08:16,383
by using the Relational Database Service.