Analytics Architectures on Microsoft Fabric

on YouTube ↗48m

" summary

Compares concrete architectures on Microsoft Fabric and shows measurable wins: a 10TB TPC‑H run matches Synapse DW3000 on an f64 at roughly 1/8 the cost, ingestion clocks ~6 TB/hour on f64 and up to ~34 TB/hour on larger skews, query bursting and resource isolation provision nodes per query, Spark offers 128 cores per f64 with 3x burst concurrency, one Lake plus shortcuts and mirroring avoid copying, V‑order parquet gives up to 50% read improvement, and fixing bad schema types cuts a query from 8:10 to 5:15.

AI-written summary · about

" chapters (8)

1. 0:00Introduction
2. 2:56Analytical workloads challenges
3. 8:14Data warehouse limits
4. 15:38Data Lake drawbacks
5. 23:50Cloud migration benefits
6. 31:51Lakehouse promise
7. 40:45Data mesh implementation
8. 47:47Q&A

" related talks

• 46mMicrosoft Fabric Synapse data warehouse Q&AOlivia Shea, Joanna Podgoetsky, Priya Sathy, Shawndell Elfring, Bogdan Crivat, Steve Howard, Faisal Mohamood · Microsoft Build 2023
• 30mDataEngBytes 2023 - BRI-T1-04 - Mimoune DjouallahMimoune Djouallah · DataEngBytes Brisbane 2023
• 23mArchitecting for Performance, Scalability and Real-time Data with AWSDaniel Smith, Mitch Kemp · Aerospike Summit 2022: Mission Now
• 06mAchieve predictable performanceAlexander Grbic · Velocity 2017
• 40mBest Practices and Optimization with the Spark EngineJohan Hille Bringsdal · Øredev 2024
• 48mAccelerate insights with real-time Azure SQL data and Microsoft FabricBob Ward, Anna Hoffman · Microsoft Build 2024
• 26mIntegrating Microsoft Fabric and Databricks for Modern Data AnalyticsShaurya Agrawal · Commit Your Code 2025
• 50mMicrosoft Fabric Data Factory Q&AWee Hyong Tok, Olivia Shea, Nellie Gustafsson, Justyna Lucznik, Shawndell Elfring, Miquella de Boer, Chris Finlan, Cristin Ford, Mark Kromer, Martin Lee, Piero Morano · Microsoft Build 2023