Edge computing in SaaS: how we reduced latency and what went wrong [ukr]

We wanted to make our service lightning-fast for users anywhere in the world. Edge computing looked like the perfect solution. In practice, we achieved lower latency — but also ran into a whole bunch of unexpected problems.

In this talk, Igor will cover:
• how they designed edge architecture for global users;
• edge providers and infrastructure: what we chose and why;
• which optimizations actually made a difference;
• architectural trade-offs that shaped our system design;
• when edge turned into an “edge-case” and forced us to find unconventional workarounds;
• our failures — and the best practices we derived from them;