In the last week, I attended the BigDataConference Vilnius to give a presentation. Actually, this was the second time to give a presentation at the external tech conference since I’ve already given a talk at the HighLoad++ in Moscow too. Though I experienced a little bit difficulty in HighLoad++ because most of the participants spoke Russian that is not understandable to me, both conferences provided me a so nice experience to me. That’s what I’m going to write about today.

BigDataConference is a tech conference discussing various kind of technologies related to high scalability system. We talked about database system, distributed platform, cloud, and AI. To be honest, I’ve not known an event that is covering such broad field. That was a pretty fun event.

I talked about our infrastructure of auto scaling platform in the conference. We are maintaining large-scale distributed systems in our daily work like Hadoop, Presto. Estimating the proper capacity and adjusting the cluster size is generally time-consuming work to be avoided as much as possible. In this talk, I introduced the design and technology we make use of to achieve the automated distributed systems.

Infrastructure for auto scaling distributed system from Kai Sasaki

The talk itself seems to get a popularity because I got around 70% +1s as feedback and many questions. They are very interested in the good practices of the AWS technology for distributed systems. For example, they asked me how to select the best instance type for the specific workload like large-scale data processing. As well as I have no clear criteria for the selection of instance type, they do not have. It would be great if I could get an answer about these question.

But anyway that was a great conference to me because it was valuable chance for me to give a presentation in English. I want to keep working on these activities even in 2019.

Happy New Year!