(4:15) With Microsoft Partner Co-Sell Program, if you build a software for Azure, Dynamics 365 or Power Platform, the Microsoft Enterprise Sales force (tens of thousands of Microsoft sales people) can help you sell your products and grow your business. More than $5B of partner revenues from this in the last 12 months. UiPath is one of such partners (presentation and demo at (6:00))
Visual Studio 2019
(9:40) Demo of Visual Studio 2019’s new features, such as:
(12:30) Editor config to be shared with the team to ensure that styling (eg: spaces vs tabs) is consistent.
(14:10) Intellicode, which is Intellisense showing the most likely options to complete your code according to the context thanks to AI.
(15:30) Live Unit Testing: tests are re-run as change something without even saving it.
(16:20) Live Share debug sessions with remote colleagues: the code is hosted only on the machine of the person that created the shared session and there’s no configuration to do on the remote people’s machines to edit, run and debug it (even if the host and the remote machines run Visual Studio on two different operating systems).
(20:40) Hosted development environment inside of Azure, so that you can use Visual Studio Online to edit and debug code even from an iPad.
GutHub + DevOps
(23:05) Last year Microsoft acquired GitHub, the largest developers community in the world, with 36M+ developers, 100M+ repositories and 1.1B+ contributions. Since then they’d added free private repos, a VS Code Pull Request plug-in, and many more.
(24:23) The integration of GitHub and Azure DevOps allows you to deploy on a number of environments, such as cloud or on-premise Kubernetes. Azure CI/CD services are called Azure Pipelines, and they can be used for any platform and any language, with unlimited free build minutes (parallel jobs have a cost). Demo at (25:42) showing how to use a template to deploy a solution to Kubernetes, and how to visually edit the YAML file that describes the pipeline. (31:40) GitHub and Azure Pipelines are used by projects like Python and companies like Shell.
(35:09) App Service is now available for Linux with a new perpetual free tier and virtual network support.
(35:40) Functions now support Azure API Management, have a premium plan for pre-warmed functions that have a zero cold start, and PowerShell language support (in addition to C# and JavaScript).
(36:05) Azure Kubernetes Service (AKS) is a fully managed orchestration service that has auto-scaling and auto-patching. Virtual Nodes (elastic provision of compute capacity with no infrastructure to manage) is now in General Availability.
(37:15) KEDAis a new Kubernetes-based event-driven container creation with auto-scaling, available for the cloud and on-premises. There are a number of built-in triggers that can drive auto-scaling. Azure Functions can also be deployed on AKS.
(38:09) Presentation of how ASOS.com uses AKS and Azure for its huge scalability requirements (thousands of requests per seconds with an average response time of 48ms). ASOS CTO Bob Strudwick talks about the microservices-based architecture that allows individual scaling of loosely coupled components, how using Azure allows them not to worry about database backups and other infrastructure concerns, and how they used Cloud Services, Azure SQL DB, Azure Cosmos DB, Azure Redis Cache and more. Specifically for their Saved Items service, they used Docker images and AKS for orchestration, GitHub and Azure DevOps, ServiceBus, Azure Functions, Cosmos DB and Table Storage.
Microsoft on the Edge
(43:30) Microsoft now has a range of edge devices, such as Azure Stack, Azure Data Box, Azure Sphere, Azure Kinect and HoloLens. Hardware product demos at (44:55). Besides Microsoft hardware, there are thousands of certified devices from hundreds of partners.
(48:46) BMW used Azure to develop their connected car experience and also to support their manufacturing processes. Presentation from the BMW Group VP of Operations about the Open Mobility Cloud, which is an intelligent platform that constantly learns from contextual information and touchpoints to give personalized help to the consumer. They use AKS, App Service, Azure Functions, Cosmos DB as well as a number of cognitive services to implement conversational experiences. Demo at (52:51).
Artificial Intelligence
(55:45) Azure Artificial Intelligence offers pre-trained AI models with Azure Cognitive Services (Vision, Speech, Search, Language and Decision) that can be deployed and run anywhere (on the cloud or into your own isolated apps) using containers and that can be accessed through a RESTful API. It also allows you to extend this services and to create custom AI models with Azure Machine Learning.
(58:22) Azure Cognitive Search, which allows to do advanced searches on your data, is now in General Availability.
(59:06) Azure Machine Learning can scale training from 1 to hundreds of thousands of servers, works with any Python environment (Visual Studio, Pie Charm, Jupiter Notebooks etc.) and frameworks (TensorFlow, PyTorch, Keras). There is now an Automated ML interface that can create a model from your data with zero code, a new visual drag & drop environment. MLOps is kind of DevOps but for machine learning (create a model and then test, deploy and monitor it).
(1:00:39) Demo of Kroger running AI on edge devices (Azure Intelligent Edge), showing how they are able to use cameras an AI in the stores to identify out of stock items in the shelves, re-stock them and avoid missing sales. Kroger’s hi-res cameras grab terabytes of videos every seconds, and they can’t send all that to the cloud, both for privacy concerns and to avoid lags. So all the image and AI processing happens directly in their stores through Azure DataBox Edge running Docker containers.
(1:06:39) Demo of AI analyzing videos in real-time and on-premises in hospitals, for example to identify when the surgery room is ready for the doctor.
Database services
(1:08:12) Azure offers a number of database services: Azure SQL Database, Azure Cosmos DB, Azure Database for PostgreSQL, MariaDB and MySQL. Azure Hyperscale Databases allows to dynamically scale the resources of a database so that for example it can move from Gigabytes to hundreds of Terabytes of size, and from hundreds to millions of transactions per second. Hyperscale for Azure SQL DB is now available in general availability. The version for PostgreSQL is now available in preview.
(1:10:20) New pricing options for Azure Serverless Databases (eg: the new Azure SQL Database Serverless) allows you to scale DB resources and pay on a per-second basis, which is great for unpredictable workload demands.
(1:11:13) Azure SQL Database Edge: in addition to be able to run on the cloud and on-premises, SQL Server can now also run on edge devices.
(1:12:19) Azure Cosmos DB is a horizontally scalable DB that can put data in different geographical locations, and that has a multi master support that allows to not only read but also write data with a guaranteed SLA of single digit millisecond latency for the 99th percentile (no other DB in the world can promise the same). New capabilities include a built-in Spark Analytics on Operational Data, built-in support for Jupyter notebooks, and the ability to build, train and operationalize ML models directly into the DB store. Demo at (1:13:50) showing how Coca Cola used Cosmos DB.
(1:15:46) A Gigaom benchmark shows how Azure SQL Data Warehouse currently outperforms both Amazon Redshift and Google Big Query, while also being cheaper.
(1:16:48) Power BI supports web, iOS, Android and Windows users. Demo of how to use Power BI, Data Factory and other services to deliver business insights such as the reason for shipping delays: data is imported from SQL Server, Dynamics and SAP (more than 80 connectors for both on-premises and cloud data sources), cleaned and transformed visually with no code, and then analyzed with Power BI. The created dashboard is finally published on the web so that users can interact with it.
(1:24:50) Power BI, PowerApps and Flow (a workflow engine like the one in Azure Logic Apps)allow even non-developers to create rich experiences that work on any device. Demo at (1:28:16) showing how an application is created from the browser-based PowerApps Studio.
