NX JAVA meetup # 14: performance research on prod, pros and cons of Spark in solutions based on Cassandra
Hello!
On August 29, a java-mitap will be held in the St. Petersburg office of Nexign.
Beginning at 19:30.
More details and a link to register - under the cut.
Reports
19: 30-20: 10 - Victor Bodrov “Practice of shooting for food: from idea to implementation”
Victor: “In order to get the most accurate results, we conduct a variety of performance studies on the prod. For our experiments, we created a special infrastructure containing dedicated routes, test data, and entities. This platform allows you to explore the internal problems of the system and integration problems with numerous counterparties.
In the report I’ll tell you how the technical features of this infrastructure are implemented, what unusual organizational difficulties arise in the preparation and conduct of such experiments. How can we quickly eliminate the negative impact of such studies on the stable functioning of sales? ”
20: 30-21: 30 - Anton Blinov “Spark: analytics and operation on the Cassandra cluster”
How to verify the success of migration to a new scheme if you have several data centers with a bunch of nodes, and the amount of data is hundreds of terabytes? And most importantly, how and how to realize this migration?
Together with Anton, we’ll talk about the disadvantages and benefits of using Spark in production solutions based on Cassandra. We will discuss the configuration and the effect of individual parameters on performance in real cases.
We also look at the features of the SCC API and the unobvious nuances hidden behind the dry but promising documentation.
About speakers
Viktor Bodrov, Yandex.Money
He has been researching productivity for many years: it’s like surfing along the waves of growing requirements for services, where you need to balance and be on the crest. He loves cheese, travel and football.
Anton Blinov, Nexign
In IT since 2011. He started his career as a Java developer. He worked in the banking sector for more than 7 years, managed to try everything: from the development of complex SPA-applications of RBS to highly loaded analytical services.
Since 2019, he has been working in the Nexign team on highly loaded BigData services.