实施湖仓一体从 BI 到 AI.pdf

1、Stefan Groot&Joey Radstaat2024 Databricks Inc.All rights reservedIMPLEMENTING LAKEHOUSE:FROM BI TO AI2024 Databricks Inc.All rights reservedW E ENABLE DATA&ANALYTICS TEAMS IN A HIGHLYREGULATED INDUSTRYTO CREATE VALUE OUT OF DATA2024 Databricks Inc.All rights reservedwe enable20+domains globally700+d

2、ata developers80+data applicationswith1architecture2 distinctive platforms a central platform team2024 Databricks Inc.All rights reserved“We have tools but no data”We need it tomorrow!”“Let us do it ourselves”You have to solve this security issue”“Its a data mess”“Why is it so expensive?”2024 Databr

3、icks Inc.All rights reservedTHE SCALABILITY CHALLENGEAND THE CURSE OF A CENTRAL PLATFORM TEAMH O W WE ORGANIZE FOR SCALETOPICS OF TODAYWhy to create a holistic architecture(with 2 platforms)How to structure a platform engineering tribeHow we deploy secure environments at scaleHow we centrally govern

4、 data access2024 Databricks Inc.All rights reservedFROM DATA MESS2024 Databricks Inc.All rights reservedTO DATA MESH!and s t i l l having a central data lake:)IT CAN BECOME MESSYVERY QUICKLYDomain data applicationData distribution platformSource applicationSource applicationSource applicationDomain

5、data application2024 Databricks Inc.All rights reservedDomain data applicationBACK TO THE BASICSWHAT DO OUR USERS NEED TO CREATE VALUE?2024 Databricks Inc.All rights reservedtoolsaccess to access todata2024 Databricks Inc.All rights reservedWE COMBINE 2 TRENDS IN ITDATA LAKEHOUSE+PLATFORM ENGINEERIN

6、GEnterprise Data LakeAnalytics PlatformKnowledge&best practicesFrameworksReusable componentsPlatform ServicesData developer portalX-as-a-ServicesData value creationML/AI appsData explorationData appsUnity Catalog Control PlaneData MarketplaceONE ARCHITECTURETO PROVIDE TOOLS+DATA2024 Databricks Inc.A