Rethinking Relational Database CRUD Design Patterns Since We Live in a Big Data World / Chris Bohn

Опубликовано 16 августа 2018, 11:32

Приглашаем на конференцию HighLoad++ 2024, которая пройдет 2 и 3 декабря в Москве!
Программа, подробности и билеты по ссылке: clck.ru/3DD4yb
--------
HighLoad++ 2017

Тезисы:
highload.ru/2017/abstracts/305...

The past several years have seen the explosive growth of Big Data Analytics to process huge volumes of data. For example, Big Data analytics can be used to process clickstream data to gain insight into user behavior. Clickstream data generally contains only a reference to the user, without any attributes. This is where marrying production data (dimensions) with clickstream data (facts) yields powerful analytics. Unfortunately, there is often an ETL impedance mismatch between production and Big Data data stores. Depending on the Big Data database, this means either very inefficient ETL, or the inability to even load production data into the system. “CRUD” is an acronym that stands for CReate, Update, Delete. CRUD is a fundamental feature of Relational Databases (“RDBs”) that enables records to be created, updated, deleted. But in the internet and Big Data age of today, the entire notion of CRUD may be fundamentally ill suited, and a better design pattern needed.
...

Нашли ошибку в видео? Пишите нам на support@ontico.ru