1
00:00:06,651 --> 00:00:08,130
- Let's take a look at some

2
00:00:08,130 --> 00:00:11,410
of the details of the Athena service.

3
00:00:11,410 --> 00:00:15,320
And the first thing that
we need to understand is

4
00:00:15,320 --> 00:00:17,293
going to be the pricing model,

5
00:00:18,140 --> 00:00:20,210
because this is gonna help us figure

6
00:00:20,210 --> 00:00:24,540
out how we are going to optimize
the placement of our data

7
00:00:24,540 --> 00:00:27,660
as well as its basic organization.

8
00:00:27,660 --> 00:00:32,060
We are going to pay per
query to use the service.

9
00:00:32,060 --> 00:00:33,910
There's no actual storage charges

10
00:00:33,910 --> 00:00:38,910
because the data actually
exists in S3 or other services,

11
00:00:39,120 --> 00:00:41,532
but we are going to be paying

12
00:00:41,532 --> 00:00:45,925
based on the amount of data
that is returned by the query

13
00:00:45,925 --> 00:00:50,240
but also on the amount
of data that was queried

14
00:00:50,240 --> 00:00:54,586
in the first place to get that response.

15
00:00:54,586 --> 00:00:56,132
And because of this,

16
00:00:56,132 --> 00:01:00,010
we may want to think a
little bit more carefully

17
00:01:00,010 --> 00:01:03,730
about how we organize
data, especially in S3

18
00:01:03,730 --> 00:01:08,010
into prefixes that would
match the different types

19
00:01:08,010 --> 00:01:10,150
of queries that we would be issuing

20
00:01:10,150 --> 00:01:13,528
against the data to minimize the size

21
00:01:13,528 --> 00:01:16,193
of the overall data set that gets queried.

22
00:01:17,700 --> 00:01:22,700
Next, we have the option of
enforcing asset compliance.

23
00:01:23,390 --> 00:01:27,603
So if we need to issue
strongly consistent queries

24
00:01:27,603 --> 00:01:30,493
that's a parameter that
we can provide as well.

25
00:01:32,590 --> 00:01:34,610
We have the ability to use more

26
00:01:34,610 --> 00:01:39,070
than just the AWS CLI or SDK.

27
00:01:39,070 --> 00:01:42,740
You can treat your Athena service

28
00:01:42,740 --> 00:01:45,920
as a JDBC or ODBC endpoint,

29
00:01:45,920 --> 00:01:48,703
so your applications can
issue queries directly.

30
00:01:51,100 --> 00:01:53,750
You can also partition the data

31
00:01:53,750 --> 00:01:58,750
beyond just the individual
prefixes in the S3 buckets.

32
00:02:01,140 --> 00:02:05,890
And finally, you can use a
feature called Federated Query.

33
00:02:05,890 --> 00:02:07,310
And this is another one

34
00:02:07,310 --> 00:02:10,410
of those elements that helps
to separate this service

35
00:02:10,410 --> 00:02:14,160
from anything that you could
conceive of implementing

36
00:02:14,160 --> 00:02:17,610
in an on-prem data center and reminding us

37
00:02:17,610 --> 00:02:22,050
of the power that comes
with being cloud native.

38
00:02:22,050 --> 00:02:25,170
And so let's take a
look at Federated Query,

39
00:02:25,170 --> 00:02:29,090
assuming that you're
using a glue data catalog

40
00:02:29,090 --> 00:02:31,480
which is the current feature set

41
00:02:31,480 --> 00:02:35,250
of Athena and not the old Athena catalog.

42
00:02:35,250 --> 00:02:36,420
You have the ability

43
00:02:36,420 --> 00:02:41,420
to query services that
are not just S3 buckets

44
00:02:41,600 --> 00:02:44,340
and prefixes within those buckets.

45
00:02:44,340 --> 00:02:46,240
So what else can you query?

46
00:02:46,240 --> 00:02:49,450
Well, it could be a cloud watch log group

47
00:02:49,450 --> 00:02:51,950
with logs that are either space delimited

48
00:02:51,950 --> 00:02:55,943
or JSON objects or whatever
happens to be in there.

49
00:02:57,570 --> 00:03:01,223
You can federate your query
against a DynamoDB table.

50
00:03:02,850 --> 00:03:05,160
You can federate against DocumentDB.

51
00:03:05,160 --> 00:03:06,660
So you have a couple of different

52
00:03:06,660 --> 00:03:09,063
NoSQL options on the backend.

53
00:03:10,380 --> 00:03:13,070
And since Athena takes data

54
00:03:13,070 --> 00:03:16,870
and makes it available
for SQL based queries

55
00:03:16,870 --> 00:03:18,920
you're basically treating
your backend data

56
00:03:18,920 --> 00:03:20,580
as if it were relational.

57
00:03:20,580 --> 00:03:23,670
Well, it can actually be relational.

58
00:03:23,670 --> 00:03:26,680
It could be an RDS database

59
00:03:26,680 --> 00:03:30,040
or even an EC two instance that is running

60
00:03:30,040 --> 00:03:33,343
either MySQL or a Postgres engine.

61
00:03:35,140 --> 00:03:39,020
And finally, you can even
federate your Athena queries

62
00:03:39,020 --> 00:03:42,383
against custom external endpoints.