Interview with Vlad Ilyushchenko, QuestDB
QuestDB is an open source time-series database. We built it to facilitate high ingestion rates and to work in environments where there’s a huge amount of data.
My trading background was the starting point. Everybody uses kdb. There’s no feasible replacement for it, but it’s so painful to use. Back then we were just trying to remove the pain working with data that kdb induces.
Say you’ve got an idea and want to analyze some data. With kdb, it’s tough to overcome the language difficulty. It’s so hard to understand how to interact with the system. You need to hire multiple consultants charging thousands per day. You need to tell them what you want, and there’s always a back and forth telling them what you need; what you get back is not quite right. You interact with this data by proxy, through consultants, and it’s just a pain for people who want to know more about their own data.
I started working on QuestDB about seven years ago as a side-project that I would spend evenings and weekends working on. Regular traditional databases didn’t seem to have time series on their radar at all. Looking back, it was an emergent data type, and people would probably use whatever tool they had at hand to handle it. Typically, users put time series data into Oracle databases or something open source, like PostgreSQL or MySQL.
