Today databases of the Massive Parallel Processing class are the industry standard for storing Big Data and solving various analytical problems based on them.
The use cases for mpp databases are varied: they can be both a "workhorse" of corporate BI, and a tool for centralizing data from hundreds of sources in one DWH, and even used for "hot" interaction with ML models working in a productive environment.
This class of technologies is a necessary element in the toolkit of a modern Data Engineer.
In a demo lesson, we will analyze in detail what an mpp-database is, what solutions are on the market today, and even delve into a practical example of using one of the most innovative mpp-systems today: ClickHouse.
We invite everyone to join the demo lesson "Introduction to MPP Databases Using ClickHouse as an Example".
IT- . , , , . , , , , , .
. ( ) , ( ). , , , SQL-, , , — data science , ..
, , , , . , , 20 , . , .
…
, . , !
, - , . Gartner, 2019 66 , 24% — — . Pitchbook, 30 5 8 35 .
, 2015-2020 .
: , , , a16z, , , , . , a16z, : https://a16z.com/investments/.
. , Linkedin 2019 . NewVantage Partners 60% Fortune 1000 , 12% 2012 , McKinsey .
, ( ) - — , .
- , . , , . , .
, : () , () , .
:
Unified Architecture for Data Infrastructure
: (OLTP), SaaS- . , .
:
— , . . - , — .
, .
, AI/ML ?
: - ( ) , ( ).
. (data warehouse). -, SQL ( Python ). (data lake) . , , , . , Java/Scala, Python, R SQL.
, ( ). , , — , , , ACID-, SQL . .
: ? ? , - . , - , .
, , , - , SaaS . , . (, ETL-) .
, . ( ) .
, «» (blueprints) — , , , , .
. -, . , , , . AI ML .
1: -
- - - , , .
. , .
, , SQL ( Python) .
, , . , , data science, / .
2:
, , , Hadoop-
, .
, .
-, , AI/ML, , / , ( , ) (Java/Scala, Python, SQL).
, , , . , - , .
3: .
, , .
, , . , .
, - (. . ), .
, , , / . , , , , - .
. . , -, , , , . , .
, , , , , , , .
- "Data Engineer".
« MPP- ClickHouse».