ok.tech: Data Explained

UPD Presentations of event participants are available here

June 13, we invite everyone who works with data to the Moscow office of Odnoklassniki, at ok.tech: Data Explain. Together with colleagues from OK.ru, Mail.ru Group, ivi.ru, Yandex.Taxi and other technology companies, we will discuss the evolution of storages and databases, talk about the advantages and disadvantages of different approaches to data storage, as well as how these approaches affect convenience of different data interaction teams.

The event will be held in the format of an open discussion between the speakers and the audience, so prepare your questions and do not hesitate to ask them. The event will be moderated by Alexey Chernobrovov.

Register for the event.

Under a cat theses of performances and the schedule.

The border between Hadoop and MPP databases.

Fedor Lavrentiev, Head of Analytics Tool Development, Yandex.Taxi

Hadoop burst into the world of storage as a cheap and easily scalable solution for processing loosely structured data. Over time, the data in Hadoop became increasingly structural, and on the other hand, there was a trend towards parallelization of classical relational databases. And now they are facing. In the report, I would like to talk about where the borders are now in the MPP world.

The evolution of DWH in Odnoklassniki.

Denis Gabaidulin, Developer at OK.ru

DWH at Odnoklassniki helps businesses make data-driven decisions. This is a fascinating, difficult and responsible task. We process tens of billions of events daily and run 1000+ regular jobs. Where did we start and where did we come from, and where do we move on? Why did you write your analogues luigi, spark job server and create your own monitoring?

Switching to ClickHouse: what awaited us

Olga Makarova, Product Manager at ivi.ru

ClickHouse does not slow down if used correctly.
I’ll tell you a few situations that cost us a lot of time and nerves, and how not to get into them.
The antipatterns of using CH are the rake we stepped on.

Where to use elasticsearch today.

Ilya Sereda, Lead Data Engineer tutu.ru

For many years of using elasticsearch in our company, we have solved a bunch of different problems with it. I’ll tell you which solutions proved to be good, and which added a headache to us. I will also tell you what to expect during operation, and in solving which problems we abandoned elasticsearch in favor of other solutions.

timetable

18:30 - 19:00 - Registration of participants
19:00 - 19:05 - Intro from Alexei Chernobrov
19:05 - 19:20 - The border between Hadoop and MPP databases, Fedor Lavrentiev, Head of Analytics Tool Development Yandex.Taxi
19 : 20 - 19:35 - The evolution of DWH in Odnoklassniki, Denis Gabaidulin, OK.ru
19:35 - 19:50 - Switching to ClickHouse: what was waiting for us, Olga Makarova, Product Manager at ivi.ru
19:50 - 20:05 - Where to use elasticsearch today, Ilya Sereda, Lead Data Engineer tutu.ru
20:05 - 20:30 - Questions to speakers
20:30 - 20:45 - Break
20:45 - 22:00 - Discussion

Register for the event.