PASS Summit 2016 – Blogging again – Keynote 1

.So I’m back at the PASS Summit, and the keynote’s on! We’re all getting ready for a bunch of announcements about what’s coming in the world of the Microsoft Data Platform.

First up – Adam Jorgensen. Some useful stats about PASS, and this year’s PASSion Award winner, Mala Mahadevan (@sqlmal)

There are tweets going on using #sqlpass and #sqlsummit – you can get a lot of information from there.

Joseph Sirosh – Corporate Vice President for the Data Group, Microsoft – is on stage now. He’s talking about the 400M children in India (that’s more than all the people in the United States, Mexico, and Canada combined), and the opportunities because of student drop-out. Andhra Pradesh is predicting student drop-out using new ACID – Algorithms, Cloud, IoT, Data. I say “new” because ACID is an acronym database professionals know well.

He’s moving on to talk about three patterns: Intelligence DB, Intelligent Lake, Deep Intelligence.

Intelligence DB – taking the intelligence out of the application and moving it into the database. Instead of the application controlling the ‘smarts’, putting them into the database provides models, security, and a number of other useful benefits, letting any application on top of it. It can use SQL Server, particularly with SQL Server R Services, and support applications whether in the cloud, on-prem, or hybrid.

Rohan Kumar – General Manager of Database Scripts – is up now. Fully Managed HTAP in Azure SQL DB hits General Availability on Nov 15th. HTAP is Hybrid Transactional / Analytical Processing, which fits really nicely with my session on Friday afternoon. He’s doing a demo showing the predictions per second (using SQL Server R Services), and how it easily reaches 1,000,000 per second. You can see more of this at this post, which is really neat.

Justin Silver, a Data Scientist from PROS comes onto stage to show how a customer of theirs handles 100 million price requests every day, responding to each one in under 200 milliseconds. Again we hear about SQL Server R Services, which pushes home the impact of this feature in SQL 2016. Justin explains that using R inside SQL Server 2016, they can achieve 100x better performance. It’s very cool stuff.

Rohan’s back, showing a Polybase demo against MongoDB. I’m sitting next to Kendra Little (@kendra_little) who is pretty sure it’s the first MongoDB demo at PASS, and moving on to show SQL on Linux. He not only installed SQL on Linux, but then restored a database from a backup that was taken on a Windows box, connected to it from SSMS, and ran queries. Good stuff.

Back to Joseph, who introduces Kalle Hiitola from Next Games – a Finnish gaming company – who created a iOS game that runs on Azure Media Services and DocumentDB, using BizSpark. 15 million installs, with 120GB of new data every day. 11,500 DocumentDB requests per second, and 43 million “Walkers” (zombies in their ‘Walking Dead’ game) eliminated every day. 1.9 million matches (I don’t think it’s about zombie dating though) per day. Nice numbers.

Now onto Intelligent Lake. Larger volumes of data than every before takes a different kind of strategy.

Scott Smith – VP of Product Development from Integral Analytics – comes in to show how Azure SQL Data Warehouse has allowed them to scale like never before in the electric-energy industry. He’s got some great visuals.

Julie Koesmarno on stage now. Can’t help but love Julie – she’s come a long way in the short time since leaving LobsterPot Solutions. She’s done Sentiment Analysis on War & Peace. It’s good stuff, and Julie’s demo is very popular.

Deep Intelligence is using Neural Networks to recognise components in images. eSmart Systems have a drone-based system for looking for faults in power lines. It’s got a familiar feel to it, based on discussions we’ve been having with some customers (but not with power lines).

Using R Services with ML algorithms, there’s some great options available…

Jen Stirrup on now. She’s talking about Pokemon Go and Azure ML. I don’t understand the Pokemon stuff, but the Machine Learning stuff makes a lot of sense. Why not use ML to find out where to find Pokemon?

There’s an amazing video about using Cognitive Services to help a blind man interpret his surroundings. For me, this is the best demo of the morning, because it’s where this stuff can be really useful.

SQL is changing the world.

@rob_farley

Leave a Reply

Your email address will not be published. Required fields are marked *