Master System Design with This Ultimate Course – Your Complete Guide to Building Scalable Systems

AGT Development
18 followers
2 Video Views·Mar 21, 2026
Master System Design with This Ultimate Course – Your Complete Guide to Building Scalable Systems
AI Generated·May contain errors
So let's start our learning journey.
This course is broken into three sections on the high level design section.
The videos dive into the large scale problems that distributed systems face
when it comes to request routing, managing their state, authentication,
API contracts, and so on.
So you're likely to use these learnings in your team
discussions and design reviews.
I would say this section is perfect for SD one, two, and three.
Very importantly, we also have a fundamental section,
which helps you understand the concepts,
which are explained in a level and low level design at a deep level.
We are talking about eventual consistency,
the different policies around caching,
the different algorithms around caching networks,
whether that is TCP versus UDP or even video streaming with WebRTC.
So that is one of the fundamental building blocks of software engineering and
then systems. Of course,
we also have a bonus section of meeting recordings here.
We have all the zoom sessions,
which have ever been recorded at interview ready.
These are held twice a month and we discuss different things from designing
train algorithms to what DNS and CDNs are.
This video is freely available and you can check it out right now.
As for the features of this course,
you can rate any video that you like good or bad.
We closely monitor feedback and update videos based on the ratings.
You also have the about section in the videos,
which describes what this video is about.
The discussions tab is where you can post your comments and your thoughts.
And interestingly, you can take down your notes over here,
which will be autosaved once you are done typing.
That means before an interview,
all of your notes are stored in the lessons itself.
Also, if there are resources that we want to direct you to,
you can see them in the resources section of a video.
The videos have timestamps so that it's easy for you to navigate around them
with the appropriate titles.
We also have subtitles for all the videos so that if you don't understand my
accent, you can at least understand the text.
We have quizzes to test your knowledge.
Getting it right is kind of proof that you understand the concept.
You can of course retry the quizzes as many times as you like and review answers
where an explanation of the right answer is given.
If you have delicate eyes like mine, you can always go to dark mode.
If you want to search for something,
you can click on the search bar and go to the result.
And finally, if you like the content, please feel free to share it.
We would love to see your support on Twitter,
LinkedIn and amongst your friends. Thank you so much for listening.
I look forward to seeing you in the next class. All the best.
In this series,
we'll be talking about distributed systems and how they're designed to meet
business requirements. When I say design,
I mean the placement of different components in a distributed system and how
they interact with each other.
We'll be demonstrating this with an example of an e-commerce website.
This is a website where users can view and purchase listed products.
The example will help us understand how real world systems use design patterns
and tools to serve customers. If you're working in a startup or a side project,
you'll see that the course reflects real life scenarios.
And if you're a software engineer or technical manager,
these learnings will help you make good technical decisions.
By the end of the series, you'll know how to design a distributed system.
You'll also know the common design patterns,
algorithms and tools that help power a distributed system. Finally,
you'll be able to effectively communicate your ideas in a team meeting or in a
design interview,
which will help your team make better decisions and will help you cover more
ground in design interviews.
All system design is evaluated using three metrics. One,
the simplicity of the design. Is the design understandable?
Can we change it if necessary? Two, fidelity.
Fidelity to the original requirements are all the requirements being met by the
given design. And three, cost effectiveness. Is this design affordable?
Is it feasible?
Here we look at the implementation costs of the design and any tooling that we
have to purchase. Throughout the series,
I'll encourage you to evaluate our design decisions based on these three
metrics. All the best. Let's begin.
Let's take an example of an e-commerce website and I'll be building a small
background story so that things make sense.
We will be tying back to the story from time to time.
Whenever we are introducing any new product requirement,
any new business requirement,
there may be times when we want to cut down on costs or make the design more
extensible.
But at all times we'll be using the same example of an e-commerce website.
So let's begin. Let's say you meet a friend who sells t-shirts.
Currently they're selling t-shirts either through social media or offline.
And what they're really looking to do is build an online presence.
They should be able to sell t-shirts online.
So they come to you as a technical expert and ask you what they should do to
sell t-shirts online. Remember the three metrics.
You want this to be simple.
You want it to cover all requirements and you want it to be cost effective.
So the best solution is an existing solution.
Maybe you can sell your t-shirts on Amazon.
The person tells you that Amazon is not the best choice because they're not able
to customize the look and feel of their shop.
These are sold as products on Amazon and the recommendations may be completely
different from what they want to recommend if a person is buying a t-shirt.
So at this point you tell them,
why don't you host your t-shirt shop over Shopify?
It gives you reasonable control.
You have your own webpage and there are lots of integrations that Shopify gives
you for payment providers and delivery providers.
Your friend checks out Shopify, looks at all the features and says,
this is wonderful. I'll go ahead with it.
You're looking at the core principle of a software architect is to solve the
business problem. It's not to create a technical solution,
but to actually solve the customer's problem.
If the solution already exists, do what is already out there.
Use the existing solution. In three months time, your friend comes back.
They say that Shopify is doing very well.
Their business is making around a hundred sales per day,
but they have some serious issues. Number one,
whenever they're looking for order details on Shopify,
they're not able to get a detailed log of how the t-shirt was taken from order
to delivery. Often they get customer calls saying,
can you tell me where the t-shirt is right now? I mean, is it delivered?
And what you have to do is you have to log in to a portal.
You have to check the delivery status and then you come back to the customer
and you have to manually tell them right now it's in processing or right now
it's out for delivery or right now it has been delivered.
I'm surprised you haven't got the support issues are expensive because any kind
of delivery tracking is difficult.
The second issue that customers are saying is that when they want international
orders, when they are looking to pay to an international card,
your payment gateway is not working.
So let's say they're delivering the t-shirt to a person in your country,
but they are ordering the t-shirt from outside your country.
So they have a different credit card or debit card that they want to use.
Your payment gateway is refusing to accept this currency.
So it would be really nice if the payment gateway could be either changed and
that's stopping, let's say 10 to 20% of all sales. At this point, again,
you use your basic system design principles and say,
let me not design the system.
Is it possible for Shopify or any of its partners who are integrated with it to
solve this problem?
Can you create a support ticket and talk to the payment gateway provider,
have them accept different payments. The business person now says,
I've already done that.
They refuse to accept international cards unless you have your own website.
Any integration with Shopify is not going to work here.
It's the same problem with the delivery provider.
Both of these problems are going to last forever unless you're able to move out
of these providers, the payment provider and the delivery provider.
Since Shopify is not satisfying your requirements and has no clear deadline when
they will, you create a technical wrapper,
a solution to solve these two problems. This is where system design begins.
This is where most businesses start. You have a set of integrations.
You have an existing solution which you're using.
Now you want to extract functionalities away from it,
bring in more control and lesser bugs in your new implementation and meet
necessary business requirements in your new implementation. At a high level,
what do you really need?
You need an integration with Shopify such that when a user wants to make a
payment, they come to your custom page. In this page,
you may have a different payment provider who allows international cards.
For example, you may give the options of Razorpay,
PayPal and Stripe in this payment page.
The other thing you want is a page for delivery tracking.
So once a payment has been completed, once an order has completed,
once an order has been booked,
you want a page where users can enter their user ID.
Maybe they can enter the order ID and they can see the delivery status.
So let's draw this out as a bunch of block diagrams.
Most of these are black boxes. You don't know how Shopify works.
You don't know how PayPal or Razorpay works,
but you do need to know how your system is going to interact with these
existing systems to satisfy customer needs.
So when a customer is looking to pay, you have this entry in Shopify,
which tells you where to redirect the customer.
So that's going to be your payment page.
And when the customer is looking to track the status of an order,
you have the link entered in Shopify.
If they click that and your delivery tracking page opens up, okay, that's it.
That's all you need. Now, if you're going to manage payments and delivery,
then you need to store some data in your system for a particular order.
Did the payment really go through this information will need to be stored in
your system. The second thing is for an order, what is the delivery state?
You can only display this on the page if you have the information stored
somewhere in your server.
So the page is going to be asking your server for information and your server
has to give the page that information.
And the way in which this interaction happens between a mobile device,
which has a browser running or a desktop,
which has a browser running and a server is called an API application
programmable interface.
This basically means that if you send the server information in a particular way,
you can expect data to be returned to you in a particular way.
And that's why this is called a contract.
An API contract defines exactly what kind of information needs to be sent to the
server to get a response in a particular format. In your case,
you're looking for delivery information and payment information.
If you give the server the order ID, right,
the right spelling of order underscore ID,
and then you have the order ID given in double quotes.
This is a JSON format.
You're going to get a response telling you what is the payment status for that
ordering. Now, if you display this on the webpage,
the customer is happy because they can see if their order passed or failed.
If they failed, then they can connect with customer support.
If it passed, things are good.
So this is how the interaction between a server and a client happens.
Similarly, delivery status can be managed by a different API.
If you send a JSON request to delivery status for now,
just imagine that this is some sort of a port running on the server,
which is taking electronic messages and designalizing them.
The server will read this message,
run a function and finally return the response.
So this is wonderful because an API basically for the server is a function.
It takes some parameters as inputs and gives you an output,
which is what any standard function does.
So you can write the API in any language you'd like.
You can write it in Python. You can go in Java.
And as long as the function executes successfully,
the client gets a response.
Now all you need to do as a developer is just write this code,
write these functions, flesh them out. Once you have them written locally,
this can be notepad. This can be PyCharm, any ID that you like.
You take this code, copy paste it in a cloud server, and then you execute it.
You just run that file and your server is running.
You have incoming requests. Your code will execute and give back responses.
That's all you need to do.
And a lot of this is actually managed now by cloud solution providers.
So you don't even have to do the copy pasting so much.
You can write from GitHub and GitHub can have action runners,
which are going to automatically deploy.
We won't get too much into the CICD pipeline.
We are looking at the high level design,
but code execution is a thing that is very, very important.
If you want to make your life easy as a software engineer.
In this series,
we'll be talking about distributed systems and how they're designed to meet
business requirements. When I say design,
I mean the placement of different components in a distributed system and how
they interact with each other.
We'll be demonstrating this with an example of an e-commerce website.
This is a website where users can view and purchase listed products.
The example will help us understand how real world systems use design patterns
and tools to serve customers.
If you're working in a startup or a side project,
you will see that the course reflects real life scenarios.
And if you're a software engineer or technical manager,
these learnings will help you make good technical decisions.
By the end of the series, you'll know how to design a distributed system.
You'll also know the common design patterns,
algorithms and tools that help power a distributed system. Finally,
you'll be able to effectively communicate your ideas in a team meeting or in a
design interview,
which will help your team make better decisions and will help you cover more
ground in design interviews.
All system design is evaluated using three metrics. One,
the simplicity of the design. Is the design understandable?
Can we change it if necessary? Two, fidelity.
Fidelity to the original requirements are all the requirements being met by the
given design. And three, cost effectiveness. Is this design affordable?
Is it feasible?
Here we look at the implementation costs of the design and any tooling that we
have to purchase. Throughout the series,
I'll encourage you to evaluate our design decisions based on these three
metrics. All the best. Let's begin.
Let's take an example of an e-commerce website and I'll be building a small
background story so that things make sense.
We will be tying back to the story from time to time.
Whenever we are introducing any new product requirement,
any new business requirement,
there may be times when we want to cut down on costs or make the design more
extensible. But at all times,
we'll be using the same example of an e-commerce website. So let's begin.
Let's say you meet a friend who sells t-shirts.
Currently they're selling t-shirts either through social media or offline.
And what they're really looking to do is build an online presence.
They should be able to sell t-shirts online.
So they come to you as a technical expert and ask you what they should do to
sell t-shirts online. Remember the three metrics.
You want this to be simple.
You want it to cover all requirements and you want it to be cost effective.
So the best solution is an existing solution.
Maybe you can sell your t-shirts on Amazon.
The person tells you that Amazon is not the best choice because they're not able
to customize the look and feel of their shop.
Things are sold as products on Amazon and the recommendations may be completely
different from what they want to recommend if a person is buying a t-shirt.
So at this point you tell them,
why don't you host your t-shirt shop over Shopify?
It gives you reasonable control.
You have your own webpage and there are lots of integrations that Shopify gives
you for payment providers and delivery providers.
Your friend checks out Shopify, looks at all the features and says,
this is wonderful. I'll go ahead with it.
You're looking at the core principle of a software architect is to solve the
business problem. It's not to create a technical solution,
but to actually solve the customer's problem.
If the solution already exists, do what is already out there.
Use the existing solution. In three months time, your friend comes back.
They say that Shopify is doing very well.
Their business is making around a hundred sales per day,
but they have some serious issues. Number one,
whenever they're looking for order details on Shopify,
they're not able to get a detailed log of how the t-shirt was taken from order
to delivery. Often they get customer calls saying,
can you tell me where the t-shirt is right now? I mean, is it delivered?
And what you have to do is you have to log in to a portal.
You have to check the delivery status and then you come back to the customer and
you have to manually tell them right now it's in processing or right now it's out
for delivery or right now it has been delivered.
I'm surprised you haven't got the support issues are expensive because any kind
of delivery tracking is difficult.
The second issue that customers are saying is that when they want international
orders, when they are looking to pay to an international card,
your payment gateway is not working.
So let's say they're delivering the t-shirt to a person in your country,
but they're ordering the t-shirt from outside your country.
So they have a different credit card or debit card that they want to use.
Your payment gateway is refusing to accept this currency.
So it would be really nice if the payment gateway could be either changed and
that's stopping, let's say 10 to 20% of all sales. At this point, again,
you use your basic system design principles and say,
let me not design the system.
Is it possible for Shopify or any of its partners who are integrated with it to
solve this problem?
Can you create a support ticket and talk to the payment gateway provider,
have them accept different payments? The business person now says,
I've already done that.
They refuse to accept international cards unless you have your own website.
Any integration with Shopify is not going to work here.
It's the same problem with the delivery provider.
Both of these problems are going to last forever unless you're able to move out
of these providers, the payment provider and the delivery provider.
Since Shopify is not satisfying your requirements and has no clear deadline when
they will, you create a technical wrapper,
a solution to solve these two problems. This is where system design begins.
This is where most businesses start. You have a set of integrations.
You have an existing solution which you're using.
Now you want to extract functionalities away from it,
bring in more control and lesser bugs in your new implementation and meet
necessary business requirements in your new implementation. At a high level,
what do you really need?
You need an integration with Shopify such that when a user wants to make a
payment, they come to your custom page. In this page,
you may have a different payment provider who allows international cards.
For example, you may give the options of Razorpay,
PayPal and Stripe in this payment page.
The other thing you want is a page for delivery tracking.
So once a payment has been completed, once an order has completed,
once an order has been booked,
you want a page where users can enter their user ID.
Maybe they can enter the order ID and they can see the delivery status.
So let's draw this out as a bunch of block diagrams.
Most of these are black boxes. You don't know how Shopify works.
You don't know how PayPal or Razorpay works,
but you do need to know how your system is going to interact with these existing
systems to satisfy customer needs. So when a customer is looking to pay,
you have this entry in Shopify,
which tells you where to redirect the customer.
So that's going to be your payment page.
And when the customer is looking to track the status of an order,
you have the link entered in Shopify.
If they click that and your delivery tracking page opens up, okay,
that's it. That's all you need.
Now if you're going to manage payments and delivery,
then you need to store some data in your system for a particular order.
Did the payment really go through?
This information will need to be stored in your system.
The second thing is for an order, what is the delivery state?
You can only display this on the page if you have the information stored
somewhere in your server.
So the page is going to be asking your server for information and your server
has to give the page that information.
And the way in which this interaction happens between a mobile device,
which has a browser running or a desktop,
which has a browser running and a server is called an API application
programmable interface.
This basically means that if you send the server information in a particular
way, you can expect data to be returned to you in a particular way.
And that's why this is called a contract.
An API contract defines exactly what kind of information needs to be sent to the
server to get a response in a particular format. In your case,
you're looking for delivery information and payment information.
If you give the server the order ID, the right spelling of order underscore ID,
and then you have the order ID given in double quotes.
This is a JSON format.
You're going to get a response telling you what is the payment status for that
ordering. Now, if you display this on the webpage,
the customer is happy because they can see if their order passed or failed.
If they failed, then they can connect with customer support.
If it passed, things are good.
So this is how the interaction between a server and a client happens.
Similarly, delivery status can be managed by a different API.
If you send a JSON request to delivery status for now,
just imagine that this is some sort of a port running on the server,
which is taking electronic messages and designalizing them.
The server will read this message, run a function,
and finally return a response.
So this is wonderful because an API basically for the server is a function.
It takes some parameters as inputs and gives you an output.
Which is what any standard function does.
So you can write the API in any language you like. You can write it in Python.
You can like go in Java.
And as long as the function executes successfully,
the client gets a response.
Now all you need to do as a developer is just write this code,
write these functions, flesh them out. Once you have them written locally,
this can be notepad. This can be PyCharm, any ID that you like.
You take this code, copy paste it in a cloud server, and then you execute it.
You just run that file and your server is running.
If you have incoming requests, your code will execute and give back responses.
That's all you need to do.
And a lot of this is actually managed now by cloud solution providers.
So you don't even have to do the copy pasting so much.
You can write from GitHub and GitHub can have action runners,
which are going to automatically deploy.
We won't get too much into the CICD pipeline.
We are looking at the high level design,
but code execution is a thing that is very,
very important if you want to make your life easy as a software engineer.
Now that we have our server set up,
we want our users to be able to access our website on their browser.
So when they type, let's say fancy t-shirts.io,
they should be able to see the webpage for fancy t-shirts.io and they should be
able to perform certain operations. For now, we are restricting to registrations.
They should be able to log in and they should be able to register,
create a new account.
These are the things which will be affecting the server that will have requests
and responses from the server. In our case, that's a serverless architecture.
But the second part is quite important. We talked about a webpage.
What really is a webpage? It's a static file,
which tells you how it's going to look like.
Maybe there is some HTML and CSS there and what it's going to behave like.
You will have some JavaScript or some way in which the webpage can react to
actions of a user.
This page will need to be stored somewhere and it can be stored in two places.
One would be in AWS servers. If a user is trying to access a page,
they just make an API call and your server says, here's the page.
That's possible.
The other way is to use a content delivery network and let's look at the pros
and cons. In this case, I think by the end of it,
the choice would be quite obvious, but let's look at the pros and cons.
Content delivery network versus actually serving the pages through a server who
a machine or maybe even a serverless function.
If you have a content delivery network,
the benefit of this is that these servers are spread across the globe.
That's already taken care of by companies like Akamai.
AWS also gives its own solution and so on.
So the benefit of having servers distributed across the globe is the pages can be
fetched very quickly.
So if you're staying in Mumbai and you have a server here in India,
you can connect to that.
But your customers in the US may connect to a server in New York and even there
you may have an East and West region.
So people in San Francisco may connect to a West region server while the New
Yorkers connect to an East region server. In India,
you may have a single server in the UK.
You may have a single server for the entire Europe region.
You may have a single server depending on your user locality.
If you see a lot of users in India, you might want to add more servers here.
If you see fewer, then you have fewer servers here.
All of this is largely managed by content delivery networks.
If you try to implement this yourself,
it's going to be quite a bit in terms of operation costs, setup, testing.
That's one drawback.
And it may be a bit expensive because content delivery networks provide
services to many customers and they are able to amortize the costs.
You doing it individually is going to be difficult.
A content delivery network doing it by themselves is usually cheaper.
You get the benefits of scale when you tie up with them. So in this case,
we will go ahead with the content delivery network.
We can use Akamai or Amazon CloudFront.
These have servers distributed across the globe and we are going to be storing
static pages in these servers. That means images, files,
videos are all going to be served from content delivery networks.
The rest of this stuff, which is data, which can be changed,
manipulated will be kept on the server in a database.
We are bifurcating the data.
Why not keep everything in the content delivery network?
Like we said, content delivery networks are distributed across the globe.
That's a benefit when you're looking at low latency.
But if you're looking for updates,
imagine you have a t-shirt which is very popular and you just have 50 of those
pieces remaining.
There are 40 orders placed in the US and 11 placed in India.
The two databases,
the two servers will show data of 50 remaining 50 remaining 40 will get booked
here. 11 will get booked here. By the time they reconcile,
by the time they talk to each other,
it's possible that in total 40 plus 11 becomes 51.
We have overplaced the orders of t-shirts.
So keeping data consistent,
keeping the same view of data in a content delivery network is extremely
difficult.
Any kind of dynamic data is stuff that you want to keep in the server.
So it would be much better if the webpage would go and fetch this information,
how many t-shirts remaining for this product from the server,
get that information and in real time, tell you how many t-shirts are remaining.
If you try to push this to the content delivery networks,
they update slowly and they are challenging to update in such a big
distributed environment.
So it's better to just keep static data like the banner of the page,
the images, the product description, the information,
all that stuff can be kept in the content delivery network.
Everything else you want to keep close to the server,
wherever you need a single source of food.
And so this brings an interesting point.
How do we add files to the content delivery network?
When you are adding files to tens of servers distributed across the globe,
is there an easy way to do it? Yes, there is.
So if you have a file system, which is being read by the network consistently,
then what's going to happen is whenever you add any file to the system,
it's automatically going to be picked up by the content of the network and
distributed across the globe. I'll take an example.
Let's say you have your computer and you have a particular folder called public.
This folder is being checked every 10 seconds by some other computer.
Whenever you add a new file to it, the other computer will see, Oh,
a new file has been added. You know,
the idea of this file is different from all the other files that I have.
I need to add this to my network. Going to pick that file.
It's going to take all the data.
It's going to take the metadata and it's going to distribute it across the
globe. The content delivery network is there. This file system needs to exist.
You can host it in your computer, but like we said,
it's better to have things on the cloud for higher reliability and possibly
lower cost. So an existing solution for this is Amazon S3.
This is extremely popular. It's a distributed file system that you can use.
If you're going to host this locally, you could have used Hadoop file system.
You could have used other file systems also.
You could use Ceph is also an open source file system,
but we are going to use Amazon S3 because it's completely managed,
has very high rates of durability and consistency.
That's a solution that you will propose as a tech expert and also the CDN that
you're going to be using. It's better to have things all under one umbrella.
So I'll be suggesting AWS CloudFront,
but you can use Akamai or CloudFlare.
These are very popular solutions to host your webpages.
Until now we have a server setup and we have our webpages distributed across the
globe. There is one thing though,
your customer has this domain called fancy t-shirts.io,
but how do you make people point to your new webpages?
Shopify somehow magically was able to get people from fancy t-shirts.io into
their server, right into their webpages. But if you have to tell Shopify, Hey,
stop sending your webpages, your webpages are really bad.
I want to show my nice looking webpages. Now, how do you do that? Well,
you need to be able to tell the customers on their browser to not point to
Shopify, but to point to your new webpages. Now, how do you do that?
Let's try to understand how the internet works at a high level.
When a person types an address on the browser, fancy t-shirts.io,
what's going to happen is the browser tries to resolve this address using
something called a domain name server.
The way this works is you pay someone for your internet connection called the
ISP internet service provider. Your request goes to your router.
The router has a hard coded address,
which goes and calls the internet service provider says that this person is
trying to resolve the domain of this address.
That request goes to a domain name server.
They look at fancy t-shirts.io and the domain name server is basically a massive
distributed network.
Similar to the content delivery network that we're talking about.
It's a big distributed network of address maintainers for every website.
It has a corresponding IP address,
fancy t-shirts.io points to 192.1.1.5 facebook.com points to
192.5.6.8.
It's got a lot of domains and the corresponding IP addresses.
So this is a dictionary. It's a map. E-value bet.
The benefit of this is that different domain name servers or different
companies, AWS has its own domain name server called DOW53.
GoDaddy is a massive domain name server company.
They can interact with each other and make sure that everybody can find the
website that they're looking for. So if AWS doesn't know about it,
it can just go and say, maybe GoDaddy knows about it.
Let me go and check with them.
When the domain name is resolved to an IP address,
you get this back on your browser.
The browser usually caches this address because you're likely to use it further.
Now you take this IP address and you go to your ISP again,
saying that I need to route a request to this address.
So if you're trying to fetch the webpage of Facebook,
you hit that IP address,
that will be a CDN for Facebook and it's going to return you a webpage,
an HTML page, which will be rendered on your browser.
Fancy t-shirts.io is going to have a similar webpage on the content delivery
network.
We need the domain name servers to have this mapping so that the internet can
find us. Nobody can remember the IP address.
They have to remember the domain name and the domain name mapping is managed by
these domain name servers.
The easiest way to do this is if you have a GoDaddy website,
then you can point it directly to the webpage.
If you are using AWS technologies anyway,
then you can tell GoDaddy that whenever you get this domain name resolution
request, just forward it to Amazon Route 53.
That is Amazon's domain name server solution.
That's going to resolve the address and give back a response.
Now the user is going to go and fetch that webpage on that IP address,
which is a CloudFront server and that will render on the browser.
So they see a nice beautiful t-shirts website and they can interact with this.
How?
Because whenever you press a button for register or anything in this
website,
the webpage is going to make a request to your server that IP address, right?
That will also be resolved in a similar way.
The domain name server needs to have a mapping for server to IP address that's
being called on the webpage. Once that's done,
you will see responses being rendered on the webpage.
You'll see data being manipulated.
So at this point you have a simple functioning webpage and server.
Customers are able to come to your website. They're able to do actions.
They're able to register, login.
One final thing that we should just touch upon and I don't want to go into too
deep because we are doing a database deep dive later in one chapter is what kind
of database should be used?
This question stumps people and I've seen real startups getting completely
confused by what database should be used. They spend like three days,
four days just thinking about databases, researching.
If you ask my personal opinion,
just go with a database that you are comfortable with.
Any database like a NoSQL database or SQL database is fine for a startup.
You can always change the database also later. If you're not sure,
if you've never worked with databases, go with SQL database.
The reason for this is it's a little more mature as a technology.
And if you're hiring someone for a data analyst position,
or if you're trying to run any analytics,
it's easier to find those queries on the internet.
Just learn SQL is much easier than learning database specific language.
Again, we are thinking like eclipse.
We are thinking like people who are on the business side of things.
We are not just thinking about, Oh, you know, this gives more scalability.
This gives more performance. No three things. Does it meet all the requirements?
SQL does. Is it simpler? Yes, it is. And is it cheaper? That's hard to say.
It depends a lot on the way you're dealing with things,
but our original three parameters are still on top for us.
