1
00:00:00,261 --> 00:00:01,990
- [Instructor] Next
up, we're going to talk

2
00:00:01,990 --> 00:00:05,415
a little bit about big data
databases, and specifically,

3
00:00:05,415 --> 00:00:08,320
two different categories
of those databases,

4
00:00:08,320 --> 00:00:12,820
NoSQL databases, and
NewSQL databases as well.

5
00:00:12,820 --> 00:00:16,110
Now, of course big data
is massive, and as such,

6
00:00:16,110 --> 00:00:18,600
requires massive amounts of storage.

7
00:00:18,600 --> 00:00:21,450
Typically, that storage is
going to be spread across

8
00:00:21,450 --> 00:00:24,840
massive data centers
world wide, and of course,

9
00:00:24,840 --> 00:00:27,780
those data centers are networked together.

10
00:00:27,780 --> 00:00:31,390
And, the concepts of
NoSQL and NewSQL databases

11
00:00:31,390 --> 00:00:35,970
were specifically designed to
deal with that type of data,

12
00:00:35,970 --> 00:00:38,300
where it's, just, so large,

13
00:00:38,300 --> 00:00:40,860
that it has to be stored worldwide.

14
00:00:40,860 --> 00:00:44,960
And those are challenges, and, sizes,

15
00:00:44,960 --> 00:00:48,740
that relational databases
really were not designed for,

16
00:00:48,740 --> 00:00:49,960
in the first place.

17
00:00:49,960 --> 00:00:53,730
So, these new databases had
to be created to handle,

18
00:00:53,730 --> 00:00:56,336
not only the massive amounts of data

19
00:00:56,336 --> 00:00:59,560
that are available to us
in big data, but also,

20
00:00:59,560 --> 00:01:03,770
the speed with which that
data gets created, as well.

21
00:01:03,770 --> 00:01:08,110
So, NoSQL itself originally
meant what its name implies,

22
00:01:08,110 --> 00:01:11,420
that is not using sequel to manipulate the

23
00:01:11,420 --> 00:01:12,980
data structured quarried language,

24
00:01:12,980 --> 00:01:16,140
like we demonstrated
earlier, in this lesson.

25
00:01:16,140 --> 00:01:19,900
But, nowadays, there's
actually more and more sequel

26
00:01:19,900 --> 00:01:22,390
being used with big data, so,

27
00:01:22,390 --> 00:01:24,547
it's now said to mean

28
00:01:24,547 --> 00:01:26,490
"Not Only SQL."

29
00:01:26,490 --> 00:01:28,950
and in fact, later on in this lesson,

30
00:01:28,950 --> 00:01:32,570
when we get to the Spark
application framework,

31
00:01:32,570 --> 00:01:34,045
one of the things that we'll use there

32
00:01:34,045 --> 00:01:39,025
is a concept called Spark SQL,
to quarry distributed data

33
00:01:39,025 --> 00:01:42,453
that resides on many nodes in a cluster.

34
00:01:44,760 --> 00:01:48,960
Now, NoSQL databases are
specifically meant for

35
00:01:48,960 --> 00:01:53,320
a combination of unstructured,
and/or semi-structured data.

36
00:01:53,320 --> 00:01:57,570
So unstructured data are
things like photos and videos

37
00:01:57,570 --> 00:02:00,620
that we're all posting on
our social media accounts,

38
00:02:00,620 --> 00:02:02,850
various concepts with natural language,

39
00:02:02,850 --> 00:02:05,240
such as the text we type in emails,

40
00:02:05,240 --> 00:02:07,096
text messages with one another,

41
00:02:07,096 --> 00:02:10,450
social media posts on
Twitter, and Facebook,

42
00:02:10,450 --> 00:02:12,070
and other sites as well.

43
00:02:12,070 --> 00:02:15,310
And semi-structured data is often

44
00:02:15,310 --> 00:02:17,680
JSON or XML documents

45
00:02:17,680 --> 00:02:22,090
that wrap unstructured data
with additional information,

46
00:02:22,090 --> 00:02:24,020
usually called metadata.

47
00:02:24,020 --> 00:02:26,300
So, for instance, back
when we were looking

48
00:02:26,300 --> 00:02:30,600
at tweets in Twitter, we saw
that with respect to a tweet

49
00:02:30,600 --> 00:02:33,040
that was either 140 characters maximum,

50
00:02:33,040 --> 00:02:35,280
or 280 characters maximum,

51
00:02:35,280 --> 00:02:38,800
what we actually got back
from the Twitter APIs

52
00:02:38,800 --> 00:02:41,730
was a much larger piece of data,

53
00:02:41,730 --> 00:02:43,610
in the form of a JSON object,

54
00:02:43,610 --> 00:02:47,650
that had tons and tons of
additional metadata in it as well.

55
00:02:47,650 --> 00:02:51,050
So, often, semi-structured
data is stuff like

56
00:02:51,050 --> 00:02:54,333
the tweets that we were
getting back from Twitter.

57
00:02:55,550 --> 00:02:58,010
Now, as we work our way through
the next couple of videos,

58
00:02:58,010 --> 00:03:01,020
we're going to talk about four major types

59
00:03:01,020 --> 00:03:06,020
of NoSQL databases, key-values
stores, document data stores,

60
00:03:06,090 --> 00:03:09,100
columnar databases,
which are kind of similar

61
00:03:09,100 --> 00:03:12,880
to SQL databases and relational databases.

62
00:03:12,880 --> 00:03:16,430
And, also a concept called
graph databases as well.

63
00:03:16,430 --> 00:03:17,880
And one of the things that you'll find

64
00:03:17,880 --> 00:03:20,910
as you start diving in to
these different database types,

65
00:03:20,910 --> 00:03:23,578
is that many of the
database vendors out there

66
00:03:23,578 --> 00:03:26,220
in the big-data space, of course provide

67
00:03:26,220 --> 00:03:29,610
free tiers or free
trials, that enable you to

68
00:03:29,610 --> 00:03:32,050
work with their tools,
get used to using them,

69
00:03:32,050 --> 00:03:33,683
and determine if they're the right ones

70
00:03:33,683 --> 00:03:35,820
for your applications.

71
00:03:35,820 --> 00:03:39,660
And, in fact, when we
get to our NoSQL example,

72
00:03:39,660 --> 00:03:41,960
we'll use the MongoDB database,

73
00:03:41,960 --> 00:03:46,500
and we'll specifically take
advantage of their free version

74
00:03:46,500 --> 00:03:48,923
of what they call an Atlas Cluster.