ClickHouse in Production: Yandex.Metrika Site visitor Embedded JS HTTP Server Message Broker Daemons Pipeline mt-nano mt-log mt-giga Key-Value Store Dictionary Data DB Java API BI Tool Site Owner Data ClickHouse in Production: Yandex.Metrika Site visitor Embedded JS HTTP Server Message Broker Daemons Pipeline mt-nano mt-log mt-giga Key-Value Store Dictionary Data DB Java API BI Tool Site Owner Data ClickHouse in Production: Yandex.Metrika Site visitor Embedded JS HTTP Server Message Broker Daemons Pipeline mt-nano mt-log mt-giga Key-Value Store Dictionary Data DB Java API BI Tool Site Owner Data

Module Input Data Module Module Module Module Input Tables Computing Task Result Table Pipeline = Directed Acyclic Graph (DAG) of modules Module = Input + Task + Output Task = Query or external external program Query = “CREATE TABLE ... AS SELECT ...” A Database System and A ML Pipeline Begin Content Area = 16,30 10 Why ClickHouse Limited hardware resources & time → efficiency matters Performance Minimalism Fundamental Algorithm Components Carefully chosen algorithm targeted at supporting the ML pipeline Featured: Data structures, graph algorithms, statistical operators, ... Write Once, Available

ds/2018/03/shutterstock_66670864.jpg 69 / 77 Тестирование ClickHouse, которого мы заслуживаем Pipeline PVS build GitHub cloud integration perf stress fuzzy functional unit style 70 / 77 Тестирование тест производительности двух версий 74 / 77 Тестирование ClickHouse, которого мы заслуживаем Full Pipeline PVS build clang-tidy GitHub cloud integration perf stress fuzzy functional unit style

特性 采集 大数据基础 PaaS平台 游戏 营销活动 Dbbinlog 数据库采集 Game DB 数 据 管 理 + 元 数 据 TDBANK 准实时传输管道 Kafka-Pipeline 实时管道 TDW 数据仓库 采 集 存 储 大数据应用 PaaS平台 数据挖掘与内容推荐 PaaS 精准 推荐 知识 图谱 精细化 运营 … iData 大数据分析PaaS