Grant Fritchey starts out reflecting on expansion of PASS content and offerings, year-round, and explains how they can help you future-proof your career. For this event specifically, he talks about learning pathways, and also about content streams, a new option for purchasing session content after the event. The different streams are architecture, analytics, and data management – now you can just buy the content for a single stream ($99), so you're not paying for content not relevant to you.
He talks about the Community Zone, and recommends everyone network at things like the Birds of a Feather tables. He says it is the biggest exhibit hall ever, and that Microsoft is here in a big way. He says the Data Clinic offers you the ability to connect with product groups and engineers about your challenges.
Ballots open for PASS Board elections at 11:30 PST.
Grant thanks Adam Jorgensen, immediate past president, for all of his work transforming PASS and developing the growth of the community.
Rohan Kumar, Corporate VP of Azure Data
Rohan starts off by saying that "SQL just keeps getting better." And he's right. He says they have invested heavily in three pillars: intelligent performance, mission critical availability, and industry-leading security. He called out Accelerated Database Recovery, designed primarily to handle unplanned failures and the outages those present, but presents all sorts of other benefits. He also explained that Always Encrypted has eliminated several of the limitations inherent in the first version.
Then he talked about performance, explaining that SQL Server 2019 has set a new world-record TPC-H on 30TB, processing over a trillion rows in under two minutes.
Next up is the Bob and Conor show. They talk about how the intention of SQL Server 2019 enhancements is to get the best performance simply by upgrading and without having to change your application. They demo how much faster a database comes back online with Accelerated Database Recovery, and also show off improvements with scalar UDF inlining.
Some other announcements from the keynote:
- Azure Arc is in preview – they call it "any data center, any cloud." Always current, elastic scale, unified management, secure, usage-based billing. Read more about it here and see why Forbes is calling it a game-changer here.
- Azure SQL Database Edge is now in private preview, offering time series, built-in AI, and choice of platform with a 300MB footprint.
- Azure SQL Database Serverless is also in preview – on-demand flexible scale, pay precisely for performance, fully managed. Best for single-database, unpredictable workloads.
- There was a Hyperscale demo – they easily scaled from 4TB to 200TB of sensor data, running on 2,400 cores, and canceled a 20 billion row delete that rolled back immediately thanks to ADR. Hyperscale now generally available for PostgreSQL as well.
- A quick Azure Cosmos DB update – globally distributed, infinitely scalable NoSQL database service (true write scale). 100 trillion tx/day, 1 million containers, 1.5 PB database (horizontally sharded). MongoDB 3.6, autopilot, jupyter notebooks, private IP.
- Azure Synapse Analytics – "a limitless analytics service that brings together enterprise data warehousing and Big Data analytics." VentureBeat talks about it here.
- Azure Data Share offers snapshot or in-place data sharing across services. (Sounds like replication on steroids to me.)
- Azure Data Factory now offers a preview of wrangling data flows.
At the end, they showed a cool demo where they had measured crowd reaction with Raspberry Pi devices installed stealthily around the room.
Finally, there was a cool shout-out to SentryOne Plan Explorer during the warmups: