System Design Intro

System design refers to the process of creating a system based on the requirements. This involves choosing the technology stack, selecting modules, components, protocols, and more.

You can break system design into two levels.

HLD - HIGH-LEVEL DESIGN

LLD - LOW-LEVEL DESIGN

Let's use a real-life example to explain this concept. Imagine we are planning to construct a city.

1. In a High-Level Design (HLD), you would determine the main components such as Houses, Hospitals, and Schools. For connections, you would plan roads and a metro system.

2. In a Low-Level Design (LLD), you would delve into the details of constructing Houses, Hospitals, Schools, roads, and the metro system.

Requirement gathering:

This is the first step before you go on to design any system. First, we have to understand what the customer wants, and what the actual requirement is, then we go and do our research on it.

Then comes things like Prioritizing, Planning features, Planning resources, and Cost estimations, Components.

Requirement gathering can also be separated into two further categories:

1. Functional

   • Need of customer

   • Features to be used

   • Example - chatting in a chat application

2. Non-Functional

   • Performance

   • Latency

   • Availability

   • Reliability

   • Consistency

   • Ease of use

   • Scalability

   • Security

Some 'back of the Envelop' calculations, handy for calculations:

Units conversions:

• 1 byte = 8 bits

• 1 ASCII character = 1 byte = 8bits

• 1 KB = 1024 bytes ≈10³ bytes

• 1 MB = 10⁶ bytes (million bytes)= 10³ KB

• 1 GB = 10⁹ bytes (billion bytes) = 10⁶ KB = 10³ MB

• 1 TB = 10¹² bytes (trillion bytes) = 10⁹ KB = 10⁶ MB = 10³ GB

• 1 PB = 10¹⁵ bytes (quadrillion bytes) = 10¹² KB = 10⁹ MB = 10⁶ GB = 10³ TB

Units of time:

• Seconds - s

• milliseconds - ms

• microseconds - μs

• nanoseconds - ns

  [ 1 second = 10³ ms = 10⁶ μs = 10⁹ ns ]

Availability:

QPS (Queries per second):

Example: Let's say Twitter has below specifications:

Monthly average users (MAU) = 300 million

Daily average users (DAU) = 150 million

Tweets (a person per day) = 2 tweets/day

Time the data to be stored = 10 years

Query per second (QPS) = (DAU x tweets/day)/ 24 hrs

QPS = ( 150 x 10⁶ x 2 ) / ( 24 x 60 x 60 )

QPS ≈ 8000 QPS

Note: Since peak QPS is 8000, we will build a system to be tolerant for at least 16000 QPS.

That is it for the basic overview, please head to the next article for further details by clicking here.

Thanks for coming this far in the article, I Hope I was able to explain the concept well enough. If in any case, you face any issues please feel free to reach out to me, I'd be happy to help.

You can always connect to me over LinkedIn to share feedback or queries by clicking here -->>Connect with me on LinkedIn.