HOW WE’VE SAVED 98% IN CLOUD COSTS BY WRITING OUR OWN DATABASE

Summary from the TLDR Web Dev newsletter

HiveKit, a company focused on tracking vehicles and people, faced a scaling problem where storing billions of location updates per month became prohibitively expensive with conventional databases. To overcome this, the company built a custom database specialized for its write-heavy load. It also optimized storage by storing delta updates instead of full state for each update. These changes, combined with moving from AWS Aurora to EBS, reduced the company's cloud storage costs by 98%.