YDB Team

Working with a talented team of engineers, I’ve spent over a decade building and evolving YDB’s low-level components, from the actor system and distributed storage to components that integrate with hardware and operating systems. We design new algorithms and protocols, formally verify them using dedicated tools, analyze complex subsystems with simulation modeling, experiment with cutting-edge technologies and hardware, and continuously focus on optimizing performance and automating system management.

Row-based tables are a core YDB component, handling OLTP workloads at the scale of tens of thousands of nodes. Our tables automatically partition data as volume or workload scales. We address challenges in internal data structure, performance, analytics, and information security, including authentication, event auditing, encryption, and secret management.

Along with the SRE team, ours automates operations for YDB and its related services. Our mission is to empower product developers to tackle their work independently, spinning up a cluster with a single click and running YDB as a service. To make this happen, we’ve launched new cloud environments, we’ve built our own Terraform provider, and we’re actively migrating our workloads to Kubernetes.

I’ve been with the YDB team since day one. My job is to handle both streaming and non-streaming message queues inside the database, along with several infrastructure services that rely on them. In our work, we tackle classic distributed systems challenges, from message reordering to exactly-once processing and fault tolerance. YDB has nearly replaced Apache Kafka at Yandex, which helped us eliminate legacy issues and increase throughput by a factor of 100. We now see our product as ready to compete with message brokers outside the Yandex ecosystem.

I’m responsible for one of the platform’s core areas: analytical (OLAP) tables that rival tools like Azure Data Explorer, BigQuery, ClickHouse, and Greenplum. Our goal is to deliver an analytics tool that lets partners monitor key metrics easily, quickly, and reliably. We continually analyze competitor solutions to develop our own high-quality product for large and medium-sized enterprises.

The main goal of my team is to provide a toolkit for working with YDB. Along the way, we are improving the UX/DevX to make it easy and enjoyable to write code for working with YDB, and to ensure that our SDKs' APIs do not contain any bugs. We are also focused on ensuring the reliability and performance of the drivers, and making YDB similar to what database application programmers are already used to working with.

I have been building DBMSs for over 25 years, and I have participated in the development of SciDB, Huawei GaussDB, and TigerGraph. I am the founder of the Database Internals community, and I have published over 20 scientific papers. Currently, I am leading the query optimizer development team at YDB, and our goal is to build a state-of-the-art optimizer for analytical, transactional, and hybrid workloads.