Data Community Austria Meeting OKTOBER

Defeat Page Latches in Microsoft SQL Server

Concurrent inserts into a clustered table with an ever-increasing key become a bottleneck since Microsoft SQL Server must protect the corresponding pages with a page latch. There are several techniques like, hash partitioning, reserve bits, in-memory OLTP and the feature introduced in 2019 CTP 3.1 called OPTIMIZE_FOR_SEQUENTIAL_KEY. We will compare these features to understand their performance improvements and the effort required to modify your existing application.

presented by Torsten Strauss (MVP, MCSE, MCT), Senior Platform Engineer @ BP Europe

Lessons learned from optimizing enterprise data models

You know that DirectQuery performance sucks - but, what should you do when DirectQuery is the ONLY available option?

You know that high cardinality increases data model size - but, what should you do when your users want to analyze numbers on a granularity level higher than day? How to reduce the data model size, while at the same time satisfying business requirements?

You know that by leveraging composite models, you can gain significant performance improvements - but why do you still have DirectQuery queries even when you target the data from the aggregated table in Import mode?

In this session, I'll share tips, tricks and workarounds learned and implemented in real-life scenarios on enterprise data models.

As you may assume, this session is not for rookies - I'll assume that you already have at least intermediate Power BI/Tabular model knowledge.

presented by Nikola Ilic, Data Mozart, Microsoft Data Platform MVP | data-mozart.com