2025-08-26 –, Potter Auditorium - Kenneth C Rowe Management Building
Getting your app running in one region is hard enough. Getting your app running in multiple regions is at least twice as hard. What about getting your app running in 10 regions with each region potentially being able to be authoritative? Today I'll show you how global replication can help you scale your app to however big you want across as many regions as you want by covering how Tigris works.
Distributed systems thinking is kinda counter-intuitive at first glance. It seems simple enough out of the gate: just copy what you did in Ashburn into another datacenter. The real fun comes in when you deal with state (most applications we'll ever touch do). How do you know which version of the world is the correct one? How do you make sure that changes are copied over? How do you design something that makes data truly global without becoming an unimaginable nightmare to grok?
This talk will cover the common ways that systems go global and how Tigris (globally distributed object storage) makes compromises to go truly global. Along the way you'll learn fun facts about FoundationDB (microservices: the database), how time can help synchronize things, and more!
Xe iaso (lower case i) is the Senior Cloud Whisperer at Tigris Data. Xe has written hundreds of articles on their blog and has worked in site reliability, backend programming, devops, systems administration, and developer relations. Xe wants to help create more people with the senior title so that everyone can benefit from the lessons they've learned.