1 00:00:02,810 --> 00:00:04,820 So we learn about SQL and NoSQL, 2 00:00:04,820 --> 00:00:09,530 let's now compare them and let's start with horizontal and vertical scaling because as I mentioned at 3 00:00:09,530 --> 00:00:15,260 the end of the last lecture, we often need to scale our database to keep up with our growing application 4 00:00:15,380 --> 00:00:18,490 with more and more users. Horizontal and vertical scaling 5 00:00:18,500 --> 00:00:21,640 are the two approaches we can use to scale our database. 6 00:00:21,680 --> 00:00:27,540 Now what do they mean? Well in horizontal scaling, we simply add more servers. 7 00:00:27,620 --> 00:00:31,750 and the advantage here of course is that we can do this infinitely. 8 00:00:31,820 --> 00:00:38,420 We can always buy new servers, be that on a cloud provider or in our own data center and connect 9 00:00:38,420 --> 00:00:42,650 them to our database and split our data across all these servers, 10 00:00:42,650 --> 00:00:47,630 of course this means that we also need some process that runs queries on all of them and merges them together 11 00:00:47,630 --> 00:00:48,580 intelligently, 12 00:00:48,590 --> 00:00:54,050 so this is generally something which is not that easy to do but this is of course a good way of scaling. 13 00:00:54,800 --> 00:01:03,020 Vertical scaling simply means that we make our existing server stronger by adding more CPU or memory 14 00:01:03,020 --> 00:01:05,890 or with something like that, especially with cloud providers, 15 00:01:05,890 --> 00:01:09,800 this is typically very easy, you simply choose another option from the dropdown, 16 00:01:09,830 --> 00:01:11,960 you pay more and you're done, 17 00:01:11,960 --> 00:01:18,260 the problem here is that you have some limit, you can't fit infinitely much CPU power into a single 18 00:01:18,260 --> 00:01:20,150 machine. 19 00:01:20,170 --> 00:01:22,720 So these are the two ways we can scale, 20 00:01:22,720 --> 00:01:24,790 how let's compare a SQL and NoSQL 21 00:01:24,790 --> 00:01:33,310 regarding that and in general. Now in general in SQL we use schemas, we also have relations, 22 00:01:33,310 --> 00:01:39,250 these are two core characteristics and data is typically distributed across many many tables which 23 00:01:39,250 --> 00:01:41,850 are then connected through relations. 24 00:01:41,860 --> 00:01:48,580 Now regarding the scaling, it's important that horizontal scaling often is very difficult or even impossible 25 00:01:48,790 --> 00:01:50,810 due to the way SQL works, 26 00:01:50,830 --> 00:01:57,030 you can of course add more servers but running them all on one shared data cloud 27 00:01:57,040 --> 00:02:01,160 so to say, one shared database is pretty difficult. 28 00:02:01,210 --> 00:02:03,400 Vertical scaling is easily possible, 29 00:02:03,400 --> 00:02:09,500 you can simply make your server stronger but adding more servers can be very hard or even impossible, 30 00:02:09,500 --> 00:02:11,960 so definitely not trivial. 31 00:02:12,100 --> 00:02:18,790 So this is a problem possibly if we have multiple or thousands of read and write queries per second, 32 00:02:18,850 --> 00:02:26,920 then maybe our SQL database especially if we do very complex joins between related tables can reach 33 00:02:27,070 --> 00:02:30,770 limits or can not be the best choice. 34 00:02:30,880 --> 00:02:39,250 NoSQL is schemaless and has only a few relations if at all, the data is typically not distributed 35 00:02:39,250 --> 00:02:45,800 across multiple collections but instead we work with merged or nested documents in an existing document, 36 00:02:45,800 --> 00:02:51,610 though we of course also have a couple of collections for the different features of our application typically. 37 00:02:52,240 --> 00:02:55,410 With NoSQL, horizontal scaling is easier, 38 00:02:55,570 --> 00:03:01,000 still something where you have to know what you do but there are cloud providers which do that for us 39 00:03:01,180 --> 00:03:07,090 so we don't have to know the ins and outs of that and in general, due to the way it works with less connections 40 00:03:07,090 --> 00:03:13,630 and so on, this is possible. And therefore we also get great performance for mass read and write requests and 41 00:03:13,630 --> 00:03:19,190 NoSQL can be very performant in an application with high throughput. 42 00:03:19,240 --> 00:03:25,420 Now this makes SQL look very bad but the full truth is that it always depends on the kind of data 43 00:03:25,420 --> 00:03:26,740 you are storing, 44 00:03:26,770 --> 00:03:31,810 if you are storing where the relations are really important and where you want to have a split up across 45 00:03:31,810 --> 00:03:36,310 tables and where you want to have strong schemas, SQL can be perfect, 46 00:03:36,310 --> 00:03:41,610 also not every part of your data is accessed multiple times per second. 47 00:03:41,620 --> 00:03:45,130 You can have parts of your application where you manage general data, 48 00:03:45,130 --> 00:03:50,600 let's say user data which does not change that often and therefore, SQL might be very good 49 00:03:50,600 --> 00:03:59,260 there. Other parts of the application, let's say orders or shopping carts that do change frequently could 50 00:03:59,260 --> 00:04:05,170 be stored with NoSQL and there, the relations might also not be that important because you can always put 51 00:04:05,230 --> 00:04:10,420 all the information that belongs to a shopping cart or to an order in one single document 52 00:04:10,510 --> 00:04:15,310 and even if you do for example store some user data there, you might not need to touch that document 53 00:04:15,400 --> 00:04:21,360 just because the user change his photo because you probably didn't store that along with the order anyways. 54 00:04:21,550 --> 00:04:23,870 Now in this course, we will build both, 55 00:04:23,890 --> 00:04:26,430 we will use both because I want to teach you both 56 00:04:26,590 --> 00:04:32,190 and it's not so much about this course application but you should know how to use SQL with nodejs 57 00:04:32,320 --> 00:04:37,900 because maybe you need to add in your application or you're working on a project where you don't decide which 58 00:04:37,900 --> 00:04:40,460 database to use but you simply have to use it. 59 00:04:40,480 --> 00:04:46,010 So we will use SQL first but then I will also show you how to use NoSQL of mongodb, 60 00:04:46,120 --> 00:04:52,300 so we'll basically implement the course project or the current state of the course project with both 61 00:04:52,510 --> 00:04:55,910 databases and you will see how to interact with both of them. 62 00:04:55,960 --> 00:05:01,630 And let's start with SQL in this module, let's install a MySQL database and learn how to interact 63 00:05:01,630 --> 00:05:04,170 with it from inside our nodejs code.