AWS re:Invent 2017: From Minutes to Milliseconds: How Careem Used Amazon ElastiCache (DAT321)

Published on Nov 29, 2017

AWS architecture for Careem, a fast-growing car-booking service in the broader Middle East, has quickly evolved to support over six million users in eleven countries. Careem also operates in areas with weak GPS signals and unique traffic patterns, resulting in the poor user experience of long driver match times and rider wait times. Careem was storing driver location data in MySQL, but their high volume of concurrent calls and lack of geospatial support in MySQL 5.6 resulted in continuous deadlocks and performance issues. Amazon ElastiCache for Redis helped meet their need for in-memory storage service with advanced data structures. ElastiCache for Redis accelerated their car booking application and reduced ride matching times from several minutes to milliseconds. Learn how their big bottleneck of insert and update operations in MySQL became a quick lookup in ElastiCache for Redis by using Redis Sorted Sets, geohashes, and timestamps.