Module 6

Module 6
Introduction to databases
Introduction to Modern Application Development Dr Gaurav Raina (IIT Madras), Tanmai Gopal (Hasura)
Contents
Objective:
1. Understand the need for a database

2. What a DBMS is and what problems it solves
3. The backend architecture of a web server and a DBMS
Topics:
1. Separating code from data

2. DBMS features
3. Backend architecture
When do we need to handle data?
Serving & Saving “content”
1. We need to use data to create HTML that fits a common structure:

a. Different article data to dynamically create HTML with a common structure (templating)
b. E-commerce sites have a common product page with different data
2. The data is just content that needs to be served to the user.
a. Eg: Files, videos, images, pdf
3. We might want to store information that is sent via client requests that is important for processing
later on
a. Eg: username/password or delivery address in an e-commerce experience
4. Client requests might submit files (like audio/video)
Separating code from data
Analyse a few examples, and see what bits of the application we would put in code and what bits we
classify as data
- Flipkart
- Gmail
- Facebook
Our current understanding of data
We covered 3 examples of using data:
1. Article templating
2. Counter incrementing
3. Adding to a names array
In all 3 examples, we stored the data as a data-structure in our code. However, some problems:
1. Article data was a part of the source code. What if we had 10,000 articles? What if it was more than
a few GB? We can’t load that much as a part of our program source code! (RAM would be uselessly
consumed)
2. Code maintenance and data maintenance are 2 separate problems which would get coupled
3. The counter and the names array is not persisted. Every single time the server restarts, we lose that
information.
So. How do we solve this?
Extremely simple ideas!
1. Separate the data files from the code files and allow storage of large varying volumes of data
2. Load data on demand to reduce pressure on resources
3. Save data back to the data files as required
Implementation
We can implement a database ourselves.
Just store our data in separate files. Split them across multiple files if the files get too large. We can read
the file to load data and save data.
But this will become a NIGHTMARE.
Think about writing code that does all these things:
1) Find data according to a particular property. Eg: All users with age greater than 21
2) Writing/updating data concurrently on the same file
3) Multiple people updating the same data. Eg: Wikipedia article being updated in parallel
4) Your server/hardware/OS crashes while you are writing information to the file
5) How would you secure access to these files?
Common problems should be solved once
In a high-performance and scalable way:
1) Storing data (knowledge of harddisks even!)

2) Querying data
3) Writing/updating data
4) Managing failure & corruption
This is why software to manage databases exist. One should almost NEVER EVER implement one’s own
database.
Terminology
A database is a term that is used to refer to a collection of data. Excel sheets, a folder full of files, a CSV
files are all databases.
In app development a database is commonly used to refer to a database managed by a specific software
like SQLite, MySQL, Postgres or MongoDB. Databases can be created by server-side or client-side
software.
These software that manage one or more databases for you are called DBMSs (DataBase Management
Systems).
Using a database
1. We already understood that if we have data in the right data structures in our server side code, we
can achieve whatever we want
a. Articles object
b. Counter integer
c. Names array
2. We now just need to get the right data into our data structures
3. Process:
a. Connect to the database
b. Whenever required, query from the database
c. Load the results into a data-structure that your server-side code can manipulate
Web Frameworks & ORMs
1. This process is also very common within the same web framework. So web-frameworks often come
with tools to help you do this querying database, loading results business
2. For many frameworks, these tools fall in a category called ORMs (Object-Relational Mappings)
Backend architecture (before database)
Web server
Computer (server machine)
Backend architecture (with database)
query processing
response request
Web server DBMS
Files
Disk
Computer (server machine)
Key takeaways
● A database is a vital component of a modern app.
○ It is easy to separate portions of the app that should be a part of the application source code and portions of
the app that should be maintained as data.
● A database management system (the software that manages the database) should be used, as far
as possible, instead of trying to handle data operations yourself
● Modern DBMSs solve common problems across apps in a performant and efficient way
○ Storing high volumes of data, or rapidly growing data
○ Querying data
○ Managing concurrent access
○ Managing access control
○ Managing storage & recovery from failure and corruption

Module 6 - Introduction To Databases

Uploaded by

Module 6 - Introduction To Databases

Uploaded by

1. Understand the need for a database

1. Separating code from data

1. We need to use data to create HTML that fits a common structure:

But this will become a NIGHTMARE.

Think about writing code that does all these things:

1) Storing data (knowledge of harddisks even!)

Computer (server machine)

You might also like