Microsoft Azure Cloud Cover Show (HD) - Channel 9

In this episode Chris Risner and Thiago Almeida are joined by Kirill Gavrylyuk, a Principal Engineer working on Cosmos DB. Cosmos DB is Microsoft's globally distributed, multi-model data service, and Kirill takes us through the key new concepts of the service and a very interesting demo. We also have a note at the end, as this is our last Cloud Cover show episode. Links from the show: [00:39] - Introduction to Azure Cosmos DB[00:58] - Scenario - Moving a MEAN app using Mongo DB to Cosmos DB[04:37] - Creating a CosmosDB account in the Azure Portal with different APIs[06:52] - Moving the data from MongoDB to CosmosDB and updating the MEAN app[09:52] - Using Data Explorer to work with the Mongo API of CosmosDB data[11:55] - Replicating data globally[16:00] - Five levels of consistency settings[18:27] - Capacity and throughput configuration[27:00] - Quick note - last Cloud Cover episode!! Welcome to Cosmos DBFollow @chrisrisnerFollow @nzthiagoFollow @kirillg_msft

In this episode Chris Risner and Thiago Almeida are joined by Bahram Banisadr, Program Manager focusing on Event Grid. Event Grid is a new service which enables developers to build event based programming. Events could be anything within Azure such as a blob being created in Blob Storage, a VM starting, and much more. By subscribing to events, you're applications can be notified whenever these events occur. You're also able to create your own custom events which can then be subscribed to. Links from the show: [01:10] - Event Grid Overview[02:22] - Basic Concepts[10:02] - Scenarios[15:45] - Custom Event Demo[26:55] - Event Grid in the Azure Portal[27:55] - Available ServicesIntroducing Azure Event GridEvent Grid landing pageEvent Grid quick startsLike Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @nzthiagoFollow @bahrawesome

In this episode Chris Risner and Noel Bundick are joined by Sean McKenna, Product Manager in the Azure Compute Team focused on Azure Container Instances. Azure Container Instances is a new service from Azure which makes it very easy to spin up containers in the Azure cloud. Primary features of ACI include no VM management, per-second billing, user defined resource (CPU and Memory), and image deployment from DockerHub and Azure Container Registry. Using the ACI Connector for Kubernetes, you're even able to use ACI as a place to deploy containers as part of your Kubernetes orchestration. Links from the show: [00:46] - Introductions to Azure Container Instances[05:14] - ACI Connector for Kubernetes[09:04] - Deploying to ACI via Azure Cloud Shell[17:34] - Using the ACI Connector for KubernetesAzure Container Instances Connector for Kubernetes on GitHubAnnouncing Azure Container InstancesGitHub repo for filing issues with ACIACI TutorialsACI on User VoiceRunning Minecraft on ACILike Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @nzthiagoFollow @seanmckmsftFollow @acanthamoeba

In this episode Thiago Almeida and Bill Barnes are joined by Donna Malayeri, a Program Manager in the Azure Functions team. Donna joins us to show local debugging of Azure Functions written in JavaScript and C# Class Libraries as well as other features and news. Links from the show: [00:42] - What's new in Azure Functions & upcoming work[03:54] - JavaScript Debugging with Visual Studio Code[13:05] - Azure Functions visualizer[14:45] - C# Debugging with Visual Studio[21:30] - Application Insights and Custom Telemetry [24:50] - New Portal UI improvementsUsing .NET class libraries with Azure FunctionsCode and test Azure functions locallyC# Customer Reviews sampleAzure Functions feature request and bugsAzure Functions youtube channelLike Cloud Cover on Facebook! Follow @CloudCoverShowFollow @AzureFunctionsFollow @billbaFollow @nzthiagoFollow @lindydonna

In this episode Thiago Almeida is joined by Mark Scurrell, Principal Program Manager in the Azure Compute group focusing on large scale compute workloads. We cover an update on Azure Batch and it's new Low-Priority VMs feature, including demos and customer scenarios. Links from the show: [01:13] - What is Azure Batch and common workloads[03:00] - Azure Batch workloads characteristics[06:56] - ffmpeg batch rendering demo[20:27] - Elasticity and maximizing utilization[22:30] - ffmpeg batch rendering demo (cont'd)[23:22] - Milliman nested stochastic cores utilization example[25:22] - Low-Priority VMs Overview [30:03] - Low-Priority VMs demo[35:00] - Options and auto-scaling with Low-Priority VMs[40:05] - Azure Batch Rendering Platform & futureBatch documentationAzure CLI – Batch template supportBatch low-priority VMs blogBatch low-priority VMs documentationRendering infoMaya plug-in for renderingLike Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @nzthiago

In this episode Thiago Almeida and Romit Girdhar are joined by OP Ravi, Principal Program Manager working on the Azure Time Series Insights service. It is a fully managed analytics, storage, and visualization service that makes it simple to explore and analyze billions of IoT events simultaneously. OP join us to explain what the Time Series Insights service is, it's core scenarios, and some great demos. Links from the show: [00:38] - Problems covered by Time Series Insights[02:40] - Time Series Insights service intro[06:23] - Core Scenarios covered by Time Series Insights[09:31] - Time Series DemosAzure Time Series Insights page Azure Time Series Insights live demo environment Like Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @nzthiagoFollow @romitgirdhar

In this episode Thiago Almeida and Etienne Margraff are joined by Mark Fussell, Principal Program Manager Lead working on Azure Service Fabric, and Haishi Bai, Senior Software Engineer on Azure. Mark and Haishi join us to go over demos and real world examples and architectures for running new and legacy games using Azure Service Fabric. Links from the show: [00:59] - Service Fabric recap[02:24] - Tic Tac Toe - Reliable actors sample[06:52] - Illyriad Games - Age of Ascent MMO game[09:19] - Fun Rock - Heroes United game[12:34] - GameFabric & Unity Tanks demo - Open Source Game Framework by Fun Rock[20:15] - XBOX Ambassador demo[27:17] - Hosting legacy games and Minecraft demo[38:33] - Tencent implementation for hosting legacy gamesTic Tac Toe sample codeGameFabric FrameworkAzure Service Fabric team blogSetting up a highly-available Minecraft server using Containers on a Linux-based Service Fabric clusterLike Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @nzthiagoFollow @HaishiBai2010Follow @mfussell

In this episode of the ML mini series, we are introducing Microsoft Team Data Science Process, a process to improve the productivity of the data science organizations of enterprises, and demonstrating an open-source data science utility, named Interactive Data Exploration, Analysis, and Reporting, developed by Microsoft. We are trying to answer the following questions: [02:22] What is data science? Why data science is hard?[11:13] What is Team Data Science Process (TDSP), why there is a need for TDSP?[18:47] What does TDSP offer to improve the productivity of data science organizations?[22:05] How does the Interactive Data Exploration, Analysis and Reporting (IDEAR) work? What problems is IDEAR helping address? Links: Recent blog post on TDSP: https://blogs.technet.microsoft.com/machinelearning/2017/04/05/latest-rev-of-utilities-for-microsoft-team-data-science-process-tdsp-now-available/TDSP GitHub repo: https://aka.ms/tdspTDSP Utilities: http://github.com/Azure/Azure-TDSP-UtilitiesData Science Project Template repo: http://github.com/Azure/Azure-TDSP-ProjectTemplateTDSP Learning Map: https://azure.microsoft.com/en-us/documentation/learning-paths/data-science-process

Continuing our mini-series on Machine Learning, in this episode we will be focusing on Microsoft R Server solutions with our guest Bharath Sankaranarayan, Principal Program Manager at Microsoft R product team. In this episode we cover "Campaign Optimization - Predicting How and When to Contact Leads" and then "Predicting Hospital Length of Stay" solutions developed leveraging SQL Server 2016 with R Services that demonstrate the capability of SQL in-db analytics. To follow the solution presented you can try it on your own by deploying the solution. The full solution including the source code is available on GitHub. Content: [00:53] - Bharath's presentation[03:14] - R Server Solutions - Cortana Intelligence Gallery[03:47] - Demo: Campaing Optimization[05:56] - Campaing Optimization - Business Model[11:28] - Campaing Optimization - Walk through source code[23:03] - Demo: Hospital Length of Stay[23:35] - Hospital Length of Stay - Business Model[31:10] - Hospital Length of Stay - Walk through source codeLinks: Microsoft R ServerCortana Intelligence Gallery - SolutionsMicrosoft R Server SolutionsGitHub Source: Campaign Optimization - Predicting How and When to Contact LeadsGitHub Source: Predicting Hospital Length of Stay Like Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @mustafakasapFollow @ispalindromed

This show starts a series of short talks introducing Deep Learning with Microsoft Cognitive Toolkit (CNTK). Topics covered in this series will range from elementary to advanced techniques. The intended audience includes: Data scientists who are new to deep learningMachine learning scientists looking for advanced techniquesEngineers looking for means to scale and leverage cloud computing in AzureContent: [03:40] - Deep Learning Made Easy - Optical Character Recognition Sample[04:12] - Test and Train dataset[06:45] - Logistic Regression[08:00] - Building classifier[09:00] - Weight matrix & Bias[11:05] - Activation function[13:27] - Logistic Regression with Softmax[14:40] - Error or Loss Function[18:05] - Train Workflow[25:10] - Prediction Workflow[27:15] - Where to begin?Like Cloud Cover on Facebook! Follow @CloudCoverShowFollow @chrisrisnerFollow @nzthiagoFollow @mustafakasapFollow @noodlefrenzy