How does Uber handle billions of payments without missing a single transaction? In this video, we dive into Uber's innovative solution for managing their massive payment infrastructure.

Uber Engineering Blogs
- [ Ссылка ]
- [ Ссылка ]
- [ Ссылка ]

We explore:
- Uber's journey from a monolith to microservices architecture
- The challenges they faced with their initial payment system
- Introduction of the Job/Order based system
- Creation of LedgerStore - Uber's custom ledger database
- The importance of immutability in financial transactions
- Uber's massive data migration process involving 250 billion records
- The Shadow Writer technique for seamless database transition
- Different types of indexes used by Uber (Strongly consistent, Eventual consistent, Time-range)
- How Uber manages over 2 trillion indexes


LINKS
---
MY 12K+ DISCORD 💬
[ Ссылка ]

CONNECT WITH ME ON SOCIAL
📸 Instagram:
[ Ссылка ]

🎚TikTok:
[ Ссылка ]

🐣 Twitter:
[ Ссылка ]

My gear 💻
[ Ссылка ]
-----

TIMESTAMPS
0:00 Intro
0:20 Ubers Problem…
1:16 How the System was Failing
1:47 Rebuilding Their Payment System
2:46 Databases… for Money?
3:26 Why Databases Can’t Handle Finances Correctly
4:48 Rebuilding their Payment Infrastructure (LedgerStore)
5:49 Ubers Big Big Problem…. (Migration)
6:36 How The Migration Worked
7:11 The Big Big Migration
8:31 Leap of Faith
9:01 Implementing Trillions of Indexes
9:23 What is an Index?
10:08 The 3 Types of Indexes
11:57 What we can learn from Uber
12:42 Thanks for Watching

свернуть