There was so much data and analytics news out of yesterday's opening keynote at Microsoft's Ignite conference in Orlando that I had three separate posts to cover it. Those posts covered Azure Synapse Analytics, SQL Server 2019's general availability and new data protection features in Power BI. In fact, though, there were numerous other data and analytics announcements yesterday. My goal with this post is to cover them too.
Azure data services anywhere: cloud databases down to earth
Among other announcements yesterday, Microsoft launched Azure Arc -- its technology that "enables deployment of Azure services anywhere and extends Azure management to any infrastructure." In the data realm within that, Microsoft also announced Azure data services anywhere (ADSA), which allows both Azure SQL Database and Azure Database for PostgreSQL Hyperscale to run on-premises or, indeed, on any infrastructure. And since ADSA is enabled by Arc, customers can manage all their Azure SQL resources through a unified management experience on the Azure Portal.
In addition to Arc, though, it seems Kubernetes (K8s) is what makes ADSA possible. K8s is emerging as the multi- and hybrid cloud enabler industry-wide, and that phenomenon is bidirectional: just as K8s allows cloud-native Azure SQL technologies to run on-premises and in other public clouds, it is also what enables the on-premises SQL Server 2019 product's Big Data Clusters to be deployed to Azure Kubernetes Services (AKS) and other K8s services in the cloud. With K8s, hybrid doesn't just mean on-premises software running on Infrastructure as a Service (IaaS) virtual machines in the cloud. Now, it can also mean cloud-managed services running on-premises and across public cloud platforms.
Microsoft says more services, beyond the two Azure SQLs, will come to ADSA. Perhaps that will one day include HDInsight (HDI), Microsoft's managed service for Apache Hadoop, Spark and other open source analytics engines and frameworks. That would make sense since, ultimately, Arc sounds a lot like Google Anthos, which Google Cloud's James Malone told me may allow Cloud Dataproc (Google Cloud's HDI equivalent) to run in cloud K8s services beyond Google Kubernetes Engine (GKE).
Core database, edge and machine learning
Beyond ADSA's implementation of Azure Database for PostgreSQL Hyperscale, Microsoft announced yesterday that the core PostgreSQL Hyperscale service is now generally available. Made possible by Microsoft's acquisition of Citus Data in early 2019, the service allows open source PostgreSQL to scale to huge data volumes for operational workloads. The public preview of Azure Database for Postgres was announced in May at Microsoft's other big technology confab, Build.
Microsoft also used Build as the vehicle for announcing the private preview of Azure SQL Database Edge; and yesterday at Ignite, the company announced that product has now moved to public preview. In May, Microsoft explained that the product integrates SQL DB and Azure Stream Analytics. Yesterday, Microsoft announced that the non-edge version of SQL DB now also integrates with Stream Analytics, as well as with Power Apps. On the hardware side, SQL DB now features new infrastructure options with greater memory and compute, allowing its maximum memory to be increased by over 400%.
Also read: Microsoft 'Builds' its data story, in the cloud and at the edge
Beyond the world of relational databases, the Azure Machine Learning (Azure ML) service now incorporates Automated ML enhancements, the ML Designer (an adaptation of the classic Azure Machine Learning Studio service) and a notebook facility that can accommodate both the Python and R programming languages. And according to Microsoft's press materials, Azure ML also gains security, governance, and responsible AI capabilities such as role based access control, quota, cost management, fairness and interpretability. Along with this, Microsoft announced the availability of ONNX Runtime 1.0, a scoring/inferencing engine for trained ML models that works across platforms and hardware.
Non-stop data news cycle
This year's Ignite conference featured perhaps the biggest list of data and analytics announcements of any Microsoft event to date. While Microsoft has been a data company for decades, its position as a leading cloud company now means data is even more important. Data powers the cloud. Analytics and machine learning add value to that power and make the cloud relevant to the enterprise. With that in mind, you can bet this volume of data and analytics news out of a tier 1 Microsoft event is no aberration.