Two of our consultants, Michael Betterton and Dawie Kruger, are attending Microsoft Ignite on the Gold Coast this week, and in this blog they share some of the insights from the event.
The opening key note delivered by Scott Guthrie, the VP of cloud and Enterprise for Microsoft had a strong cloud message to the keynote with a key takeaway being Microsoft's focus on a hybrid cloud/on-premise solution. Themes of cloud, analytics, internet of things and new infrastructure approaches prevailed. One of the case studies discussed was Rolls Royce, who have an IoT solution on all of the airplane engines they make that stream data to Azure and uses Azure ML to predict engine failure, displaying results in PowerBI.
The keynote finished with a demo using Azure Cognitive services. This demo had two components, the first was real time facial recognition. The second component was the server-less cognitive services offering which is a lego-brick style drag-and-drop code editor. In the demo the presenter built in sentiment analysis of tweets that fed into a Dynamics instance in Azure, creating cases for negative tweets, all without a single VM provisioning.
Also discussed was the vision for SQL Server & beyond presented by Lindsay Allen – the program manager for SQL. The approach is to modernise business apps by moving them to the cloud either through IaaS (infrastructure as a service) or through PaaS (platform as a service). The PaaS offering was hosting SQL DB’s in the cloud. This is a new offering by MS being released soon that is the “CL Series Databases” PaaS. This lets you easily move on-premise databases to the cloud with little effort into a single elastic cluster. This cluster can scale to 150pb. The smart database features in Azure have also been updated with the ability to suggest index creation/drops, schema issues and parameterisation of queries for PaaS DB’s.
Understanding Windows Containers - using containers you are able to create a base OS or Software layer environment for developers that can be deployed to UAT and Production servers and they assist to eliminate discrepancies that caused application to crash or “work differently” in prod. It takes the concept of virtualisation to the software layer instead of the server layer. Some benchmarks saw new containers being spun up in less than 3 seconds using PowerShell scripts. Containers are not meant to be used for databases as they are blown away and have very small footprints however it does make for some interesting concepts for app development and deployment. More on container can be read here.
There were a few sneak peeks of upcoming features in SQL coming in June. In no particular order they are:
High Availability, Disaster Recovery & Availability Groups will be available across OS (window + Linux), with any combination of OS’s.
Clustered Columnstore indexes available from LOB data types.
The ability to pause and resume online index rebuilds
Adding of machine learning to the query optimizer to attempt to learn from its mistakes. This is a first step only and the key implementation here is Table Valued Functions and assessing the size of the result set.
Python Integration - Apparently R wasn’t enough for MS and soon python will have full integration with the DB engine.
Graph Data Structures – this involves full support for CRUD of graphs & db-engine integration. This had a very short demo only but the query of the graph structure was written into normal T-SQL with its own syntax. It will be interesting to see applications of this in the future.