System Design Questions

As a technologist I love system designs. I like to tell interview candidates you don’t need to practice them, it’s what you do every day. In fact, it’s one of the favorite parts of my technical role. I never liked crossword puzzles, but I could totally do system design problems instead. Here’s a great list ..

Most of these are copied from the system design interview books/tutorials.

• Theory
  • Why do these?
• For All System Design Qs
  • Users and Use Cases
  • High Level Design
  • APIs
  • Data
  • Connections
  • Throughput/Availability
  • Capacity/Constraints
  • Regionalization I18N
  • Non Functional requriemetns
  • Capacity Estimation
  • Money: Monetization + COGS
• Specific Tech
  • Geo Approaches
• Other posts
  • Design and Architecture
  • Cloud first applications
  • Data Systems
  • Security
  • SQL vs No-SQL
• Use cases
  • Proximity Service
  • Google Maps
  • Distributed Message Queue
  • Metrics Monitoring and Alerting System
  • Ad Click Event Aggregation
  • Hotel Reservation System
  • Distributed Email Service
  • S3-like Object Storage
  • Real-time Gaming Leaderboard
  • Payment System
  • Digital Wallet
  • Stock Exchange
  • Rate limiter
  • Consistent hashing
  • Key-value store
  • Unique id generator
  • Url shortener
  • Web crawler
  • Notification system
  • News feed system
  • Chat system
  • Search Auto Complete
  • Youtube
  • Google Drive
  • Nearby Friends
• Resources
  • System design resources

Theory

Why do these?

For All System Design Qs

Users and Use Cases

High Level Design

APIs

Data

Connections

What to call this?

Polling - Wake up every ‘k’ seconds and ask

Push - Usually better (always?), only down side is harder to implement, more server load?

Long Poll - Open a http request, leave it open for a while then answer on the server. - Weird hack - What if client dies

Bi-directional - E.g. WebSocket (ideal) - Ideal - Setup over HTTP - Traverses Fire walls

TCP vs UDP

WebSocket vs HTTP/2 + Server Side Events (SSE)

Throughput/Availability

• Load balancing/bottlenecks/fault points
• Caching, purging, replication, consistency
• Serving data from edge CDN

Capacity/Constraints

• Capacity
• Scale and usage estimates (include memory, data transmission)
• There are 10e5 seconds in a day
• Other

Regionalization I18N

• Per locale laws
• Languages
• Data practices
• Geo-distributed
• Edge/CDN

Non Functional requriemetns

• Latency
• Scalability
• Availability
• Reliability
• Observability
  • Test ability
  • Monitoring
  • Metrics
  • Logging
  • Alarming
• Deployment
• Supportability
• Non Functional Requirements
  • Security, authentication, authorization

Capacity Estimation

Money: Monetization + COGS

Specific Tech

Geo Approaches

• As the crow flies - lat/long search - 2d union is expensive
• Geo Hash - Use each digit to divide space into quadrants
• Quad Tree - Geo Divide but so leafs are ‘equal sized’

Other posts

Design and Architecture

Cloud first applications

Data Systems

Security

SQL vs No-SQL

Todo Link to Data Systems

• Dynamo
• SQL

the table

Type	TPS	Notes
Key Value (Dynamo)	40K	Upto 5 Indexes
SQL (Aurora)	??

Use cases

Proximity Service

Google Maps

Distributed Message Queue

Metrics Monitoring and Alerting System

Ad Click Event Aggregation

Hotel Reservation System

Distributed Email Service

S3-like Object Storage

Real-time Gaming Leaderboard

Payment System

Digital Wallet

Stock Exchange

Rate limiter

Consistent hashing

Key-value store

Unique id generator

Url shortener

Web crawler

Notification system

News feed system

Chat system

• 1:1 vs Group
• Online vs Offlie
• External Notification System
• Presence
  • How to detect user is no longer online
  • How to propagate presence to other users (similar to propagate messages)

Search Auto Complete

Youtube

Google Drive

Nearby Friends

begin

sequenceDiagram
    participant U as Montoring User
    participant U2 as Moving User
    participant W_PS as WebPubSub
    participant FL_PS as FriendLocationPubSub
    participant F_DB as FriendsDB
    par Initialize Connection
        U ->> W_PS: RegisterFriendLocationChange()
        W_PS ->> F_DB: GetFriends()
        W_PS ->> FL_PS: RegisterFriendLocationChange (friends)
    end
    par A friend changes their location, initial sub generates location
        U2 ->> W_PS: NewLocation(location)
        W_PS ->> FL_PS: NewLocation(U2,location)
        FL_PS ->>  W_PS: OnFriendLocationChange
        NOTE over W_PS: Check Close Enough <br> Coalse and Debounce
        W_PS ->>  U: OnFriendLocationChange
    end

end

classDiagram

class FriendsDB {
User:UserId
Friend:UserId
CanSee:bool
}
class UserLocationHistory {
ID: AutoIncrement
User:UserId
Location:LocData
Time: TimeStamp
}

Resources

System design resources

Online: - https://github.com/donnemartin/system-design-primer

Add Links to Books * System Design Interviews 1 and 2