Chapter 2: Understanding Database Fundamentals

Before we start looking at specific AWS database technologies and services, it's important to understand the different types of databases that are available and what type of workloads you should consider putting into each database. We are doing this so that when we start learning about the various AWS services, you will understand how and why there are so many different types and options.

We will be studying how databases differ between running them on-premises and in the cloud in terms of access, administration, and maintenance. These topics will appear in the exam, so being able to define the differences is important.

If you already have a database background and are comfortable with the different database types and how they work, please feel free to skip this chapter and go straight to Chapter 3, Understanding AWS Infrastructure. But if you want to go ahead and learn about the differences, then stick around until the end...

On-premises versus cloud databases

An on-premises database could be defined as follows:

"A database that is owned, operated, and maintained by the customer within a location that they control and have full autonomy over the database and underlying servers and networking components. This can be a server room in their office or a rented cabinet within a shared data center that the customer has full and direct access to."

With an on-premises database, everything is done internally, from installation and implementation to running the database every day. Maintenance, security, and updates also need to be taken care of in-house. You will need to arrange for when the software will be purchased before it is installed on your servers. The customer will assume complete ownership and control, even if this is via a management company or service provider.

A cloud database could be defined as follows:

"A database that is owned, operated, and maintained by the customer...

SQL databases versus NoSQL

One of the largest decisions to make when planning a new database deployment is whether to use a Structured Query Language (SQL) or Not only SQL (NoSQL) database. These two types of databases differ greatly and making the wrong choice can compromise the performance and the ability of your application to function.

First, let's discuss the key features of both database types before doing a deep comparison of both so that you can decide between them.

SQL databases

SQL databases are designed to excel in storing structured data. They can carry out complex querying and they commonly store the minimum data possible by reducing any duplication of the data in a table in a process known as normalization. Normalized data means that accessing it often requires complex joins of different tables.

Normalized data would look similar to this:

Figure 2.1 – RDBMS table structure

These tables only contain the specific columns...

Relational database management systems

SQL or RDBMS databases have two main types, which describe the way data is stored on disk:

• Row-orientated
• Column-orientated

The different methods of storing the data and how it is arranged will offer very different performance patterns (that is, fast at some things but slow at others), and knowing about the right type to use can greatly improve the performance of your application. While both database types may appear very similar on the surface, they are quite different under the hood.

In the exam, there may be a question around a customer use case and asking which database would be the best fit.

First, let's look at row-orientated, which is the more common database system.

Row-orientated databases

In a row-orientated database, the data is stored in tables in normalized form (we discussed this in the SQL databases section) with links or keys between them.

Row-orientated databases store the data in continuous...

Key-value and document databases

We are now going to look at key-value and document databases. These are both NoSQL databases, but they have different use cases:

• Key-value databases have a unique key (such as a primary key) and then values stored next to the key. Key-value databases are useful when you have large amounts of data that needs to be queried quickly and when you have data that doesn't have clearly defined columns or data structures. Key-value databases let you store almost any data in the value component without the need to strictly define it, as you would with an RDBMS, where you would specify a string or integer, for example.
• Document databases store data in formats such as JavaScript Object Notation (JSON). JSON is used widely in programming languages, so having a database that uses the same format makes it an efficient choice for many developers. We will discuss JSON in more detail later.

First, let's look at key-value databases more closely...

Graph and ledger databases

For the final section of this chapter, we are going to study two more database types, one NoSQL and one a special type of database that can be made both with a SQL and NoSQL database, as we will see:

• Graph databases are used when you want to show the connections between items in your database. For example, consider Facebook and how they use friends of friends to help identify people you are likely to know and might want to connect with. This would be a good example of when a graph database could be a good option.
• Ledger databases are databases that keep track of every change that has ever been made. The database will never change any existing data but will add a new version next to the original. This can be very useful for systems such as bank transactions, where it is critical to have exceptional auditing controls.

Let's start by looking at graph databases.

Graph databases

Graph databases specialize in storing data in a manner...

Summary

In this first technical chapter, we have covered a lot of different databases, all of which will likely feature in your exam.

You should now know how to explain the benefits and compromises of cloud databases versus on-premises databases and understand the key motivations for moving to the cloud in terms of scalability, cost, security, compliance, and performance. You have also learned how all the database engines that will be covered in the exam work and how to identify key use cases for each based on their key features.

So far, we have looked at technologies that work both on-premises and in the cloud. Now, it's time to focus on the specific AWS technologies that you will need to know about for the exam. The next chapter will teach you the basics of AWS infrastructure, where you will deploy and maintain your databases. Without this knowledge, you will struggle in the exam as there will likely be questions on how to set up the infrastructure for your databases...

Further reading

Earlier, we briefly mentioned blockchain technology. For further information on this topic and how it can be used, please consider the following Packt book:

Blockchain by Example - https://www.packtpub.com/free-ebook/blockchain-by-example/9781788475686.