Scalability (Vertical and Horizontal)
So first all of what is scaling, Well for example, if we are a new startup and we launch one software on our local system. It was running well until it got hit by thousands of people on the internet and our server goes down because of traffic. So to give service back to our users, we need to upgrade our server. There are two ways of doing that: 1. Vertical Scaling and Horizontal Scaling. Let's talk about each.
Vertical Scaling
Here we upgrade the existing system by adding more resources like RAM, CPU, and Storage.
Here we buy a big machine(machine with big resources). It means that our machine will be bigger and therefore, It can process the request faster.
Key Points:-
- Not required the Load Balancing.
- Single Point of Failure.
- Inter-Process Communication. (This means the communication will be faster).
- Data is Consistent as there is only one system.
- Hardware Limit (We can’t make one system bigger and bigger we will be facing the hardware limit).
Horizontal Scaling
Here instead of upgrading one system, we add another system with resources. It is buying more machines (systems) of a similar type to get the work done. Refer to the below diagram.
In horizontal scaling, if the user increases then, the no. of servers also increases. So we can say that this is linear.
Key Points:-
- Load Balancing Required.
- Resilient (No single point of failure as there are a lot of servers that can attend to the requests).
- May have Network calls (As the communication between the servers are from the network).
- Data Inconsistency.
- Scale well as User increases.
So which scaling do you think is used in the real world?
both, we take some of the good qualities of vertical scaling (fast communication & data consistency) and from horizontal scaling (scales well & no single point of failure). But most probably company uses horizontal scaling with big resources to overcome the issues.
