You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
63 lines
2.9 KiB
Plaintext
63 lines
2.9 KiB
Plaintext
00:00 So wait, what is an integration database?
|
|
00:02 If you were just working on your own personal website
|
|
00:05 or some small project,
|
|
00:06 you don't have an integration database.
|
|
00:08 But if you work in a big enterprise,
|
|
00:10 a big corporation where there's many internal systems,
|
|
00:13 you may be working with an integration database.
|
|
00:15 Honestly, that's not a great fit
|
|
00:17 for NoSQL databases in general.
|
|
00:19 It also makes designing documents
|
|
00:21 for them more difficult.
|
|
00:22 In large corporations where you have many applications
|
|
00:26 that share the same data,
|
|
00:27 one way that we have built applications to share data
|
|
00:31 is to just share the same database.
|
|
00:33 We might have a bunch of different applications
|
|
00:36 and they're all going to talk to the same database
|
|
00:38 so they all have the same concept of a user,
|
|
00:40 they all have the same concept of an order,
|
|
00:42 things like that.
|
|
00:43 And this means the concept of the user
|
|
00:45 is as complicated as it can get.
|
|
00:47 Maybe the application on the top left
|
|
00:48 could have a real simple user,
|
|
00:49 the one on the top right actually needs something else,
|
|
00:52 the bottom one is something else still
|
|
00:54 until you've got a model across all these applications,
|
|
00:56 and that makes it super tricky.
|
|
00:58 Also in NoSQL databases and document databases,
|
|
01:01 the relationships are enforced in the application,
|
|
01:03 so that means all of these have to agree on
|
|
01:05 what the constraints are, what the relationships are,
|
|
01:08 and that can actually cause data integrity issues.
|
|
01:10 There's a lot of reasons
|
|
01:11 that an integration database isn't a great idea
|
|
01:13 for relational databases.
|
|
01:14 In fact, it's not a great idea at all, but it has been used
|
|
01:17 and because there's different applications
|
|
01:19 with different query patterns,
|
|
01:20 it makes designing your documents more difficult.
|
|
01:23 So instead what do we do?
|
|
01:24 We build application databases.
|
|
01:26 Maybe we have a bunch of different applications
|
|
01:28 just like before, but they all have their own data store
|
|
01:30 and they all talk to their own databases.
|
|
01:33 Of course they need to exchange data before
|
|
01:35 so maybe we do some sort of microservice thing
|
|
01:37 where they talk to a service bus
|
|
01:38 or they just talk to each other, things like that.
|
|
01:40 That means each individual database
|
|
01:43 and interaction with its own application
|
|
01:45 is super, super focused and limited.
|
|
01:48 Here in these cases,
|
|
01:49 MongoDB document databases make a lot more sense,
|
|
01:52 and it's easier to design the documents
|
|
01:54 because the range of queries is extremely focused
|
|
01:58 so you can target those particular questions
|
|
02:00 against a small set of queries,
|
|
02:03 the guidelines we just talked about.
|
|
02:04 So this is the kind of model you want to have
|
|
02:06 if you're doing data exchange within your organization,
|
|
02:10 and you're working with a document database.
|