Tarantool is an OpenSource in-memory database with features you would expect from traditinoal DBMS: primary and secondary indexes, synchronous replication, transactions and sharding. It has high programmability end extensibility thanks to embedded LuaJIT.

Often in-memory databases and data grids are hard to use because they force developers to write code in compiled languages and go through complex delivery cycle.

We believe that there is a reason why Lambda architecture and FaaS in general became so successful. So we would like to talk about our product (Tarantool Data Grid) that takes ideas from in-memory data grids and FaaS and combines them together to reach even higher speed of innovation in places where it matters most: between modern internet applications and core business services.

Topics I would focus on:

  • History and business value of Tarantool Data Grid
  • Writing code close to data. How we embedded an IDE to a data grid. LSP and dynamic code reloads.
  • Role-based clustering and horizontal scaling with SWIM protocol
  • Scaling from one Docker container on developer's machine to a cluster with 100-s of nodes
  • How to support multiple teams working with the same codebase and data inside the cluster
  • Use cases from a few of our customers in retail and investment banking
Head of Tarantool Enterprise
Leading Enterprise division of Tarantool in-memory database.



(Pacific Time Zone)