A practitioner's guide to building scalable systems

Building Scalable Systems: Distributed Systems basics to System Design Interviews

We want to give you the tools to compare multiple approaches and pick the one that works for your scenario.

Grokking the Distributed and Scalable Systems

# The ideal machine doesn't exist

In an ideal world, you only need a single computer with *just* following two traits:

1. Infinite resources 
2. Guaranteed 100% availability & Uptime

If you have found the magical machine with the above two properties, then don't bother learning about distributed systems. It will save you from a lot of trouble and headaches.

However, in the real world, we are all mere mortals, and computers have limited resources. 

For any web-scale application, a single node would run out of resources. At that point, you need to set up multiple nodes that can work together to serve your users. 

Nonetheless, even before your single node runs out of resources, there's another reason for creating a distributed system, and that is "Fault Tolerance."

# Goals of a Scalable System
The primary three primary goals of a scalable system
1. Scale
2. High availability and uptime
3. Optimal Performance
4. Cost Efficiency


## 1. Scale
*(In coming lessons, we will dedicate a whole chapter on Scaling - discussing both Vertical and Horizontal scaling, etc. but don't worry about it for now)*

Your application needs to scale because of one or more of the following two reasons:

1. **Too many concurrent users**: The number of users accessing your service at a given time exceeds the capacity of a single server.
2. **Too much data:** Your single server cannot handle the amount of data


## 2. Fault Tolerance
Servers crash more often than you think due to a variety of reasons e.g., 

1. Disk crashes, 
2. CPU or RAM overheating, 
3. Software bugs, etc. 

That's why, even before your single server runs out of resources, you need to build redundancy in your system to ensure that your users don't experience any interruption.

## 3. Optimal Use of Resources
With Distributed Systems, you can share resources and redirect resources to tasks that are more urgent when there's high load and then redirect the same resources to low-priority tasks when there's not enough load. e.g. many companies use the same servers to serve users during the day time and run batch processing jobs during the night.

## 4. Cost Efficiency
By building these scalable systems and efficiently using them, companies are able to save cost and do more with less. This could be mean saving money spent on 

- Hardware resources
- Engineering resources
- Maintainance cost

What and Why

Node in a Distributed System

Performance and Numbers

Fault Tolerance and Availability

Understanding TIME in Distributed Systems

Data Centers for software engineers

Scaling Distributed Systems

Load Balancers

Structured Storage at Scale

Sharding in Storage Systems

SQL vs NO-SQL

Blob Storage at Scale

Caching in Distributed Systems

Queues and Asynchronous Processing

Data Analytics and Batch Processing

Logging and Monitoring

Searching at Scale

Conclusion

Why do we need Distributed systems?

The ideal machine doesn’t exist