System Design for Flash Sales: Sharding vs. Zookeeper vs. Redis vs. Kafka

In this video, we dive into the exciting world of flash sales during festival seasons. Discover how limited stocks of popular items like iPhones and MacBook Pros are booked in the blink of an eye. We explore four high-level system design approaches to ensure fair and efficient booking processes. From using Kafka and Redis to sharding and Apache Zookeeper, we compare the advantages and limitations of each approach. Join us to uncover the secrets behind successful flash sale booking systems! Don’t forget to like and comment if you find this information helpful! In this comprehensive breakdown, we also address key challenges like preventing overselling and ensuring response times of under 100 milliseconds. Whether you’re a tech enthusiast or preparing for interviews, this video provides valuable insights into building robust systems for flash sales. Don’t miss out on mastering the art of booking your favourite gadgets!

credits
1. Some references take from here: System Design Fight club: https://youtu.be/462JSRNPDNs
2. A good article on zookeeper https://github.com/saptarshibasu/zookeeper-study
3. How redis counters are not consistent: https://github.com/nateware/redis-objects/issues/72
4. Apache zookeeper multi instance setup: https://youtu.be/9DXb_nIk99U
5. Apache zookeeper data models: https://youtu.be/dPKJvAHCAH8
6. Apache zookeeper playlist: https://www.youtube.com/playlist?list=PL1F54YqZCTdx4JmaKY99bb9hoD5gHS9bm
7. Redis increment counter: https://redis.io/commands/incr/

Chapters
00:00 Intro
01:49 Requirement
05:08 Kafka
09:29 Redis counter
15:50 Zookeeper
23:06 Sharded counter
26:26 Comparison