This talk describes in detail the architecture of distributed newSQL Yandex Database: what it's built on (data storage layer and "tablets" infrastructure), what components it consists of (actors, "tablets", the engine of distributed ACID transactions), what exactly allows it be distributed and horizontally scalable. This architecture motivates the creation on its basis of a product called Yandex Database Serverless: who might need it, how it became possible, what does it have in common with the usual Yandex Database, and what technical solutions distinguish them.
This talk might be interesting to database developers, to those who use databases, and who are close to the topic of serverless computing.
Download presentation