Introduction

Samsung account has more than 1.8 billion user account information. It also works with about 60 services and apps operating inside and outside of Samsung, resulting in over 2 million massive traffic per second across the system. Under these circumstances, disruption of service is by no means acceptable. Therefore, we established a column-by-column encryption plan to protect the privacy of the database in the face of these limitations. And using Kafka, we performed the rapid encryption and migration of existing user data, and succeeded in encrypting large amounts of traffic in real time. I would like to introduce in detail the difficulties and solutions we have experienced in this process.

Index

1. What is samsung account?

Samsung account introduction
Samsung account service and data features
Efforts to protect personal information

2. What parts did you think about when designing encryption?

Constraints in Encryption
A Design Method Considering Constraints

3. How did you implement it?

Database Snapshot using Kafka
Changing database schemas and encrypting data
Massive Service Conversion and Rollback

4. What was the difficulty?

Performance tuning
Technical problem
Prepare for complete operational distribution
Actual work process and result