1. Microsoft Azure P wer Lunch
6/11/2018
Microsoft Azure P wer Lunch
Today’s Topic:
End to End Data Science using R-Server, RTVS and Nodejs
Date: 8/24/2017
Presented By: Azure Solution Architects from US South Central
© 2014 Microsoft Corporation. All rights reserved. Microsoft, Windows, and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

2. Sree Ram – Solution Architect Data & AI @
Worked in areas of application development, Data warehousing, databases and AI space for over 15 years
Data Science professional with expertise in R programming
Areas of expertise and interest include Spark and big data, NoSQL Databases, Deep learning algorithms, statistical analysis/forecasting, Data warehousing and containerization
My Profile

3. Session Agenda End to End Data Science using R-Server, RTVS and Nodejs
Azure Services Updates
Today’s Topic:
End to End Data Science using R-Server, RTVS and Nodejs
Q & A

4. Azure updates
Azure Cloud Shell: Generate scripts by using mssql-scripter
Public preview: Azure Event Grid
Azure Data Catalog enhanced permissions
Azure Data Catalog support for relationships and related data assets
Azure Monitor: New capabilities for diagnostic settings
Azure Batch updates
Investing deeply in Terraform on Azure
You can now use mssql-scripter, an open-source command-line interface, to generate database scripts in Azure Cloud Shell. 
With mssql-scripter, you can generate data definition language (DDL) and data manipulation language (DML) T-SQL scripts for database objects in SQL Server (running anywhere), Azure SQL Database, and Azure SQL Data Warehouse. You can save the generated T-SQL script to a .sql file or pipe it to standard *nix tools (for example, sed, awk, grep) for further transformations. The generated script can be edited or checked in to source control. Then, you can execute the script in your existing SQL Database or SQL Data Warehouse deployment processes and DevOps pipelines with standard multiplatform SQL command-line tools such as sqlcmd, which is also available in Azure Cloud Shell.
Mssql-scripter is built using Python and incorporates the usability principles of the new Azure CLI 2.0 tools. The source code is on GitHub, and we welcome your contributions and pull requests. An installation and quick start guide is available.
Today we are introducing two new capabilities in Azure Monitor for routing your Azure resource diagnostic logs and metrics to storage accounts, Event Hubs namespaces, or Log Analytics workspaces. You can now create multiple resource diagnostic settings per resource (in public preview), and you can route your metrics and logs to a destination in a different subscription.
Recent updates have added new capabilities to the Standard Edition of Azure Data Catalog to give Data Catalog administrators more control over allowed operations on catalog metadata.
Recent updates have added new capabilities to Azure Data Catalog to deliver support for relationships between registered data assets, and discovering related data assets in the Data Catalog portal.
You can use Azure Event Grid to easily build applications with event-based architectures. You select the Azure resource that you want to subscribe to, and you give the event handler or webhook endpoint to send the event to.
Azure Batch Rendering. At SIGGRAPH 2017, we announced the public preview of Azure Batch Rendering, which updates Batch to support cloud-scale rendering capabilities on a pay-per-use basis. You can easily create Batch pools with one or more rendering applications already installed. You have the ability to pay for these applications on a pay-per-minute basis with your Azure account, instead of using your own licenses and license server. Autodesk Maya, 3ds Max, and Arnold are initially supported, with Chaos Group V-Ray coming soon. With the Batch plug-in for Maya, you can also submit a job to Batch right from Maya. See the blog post and Batch documentation for more details.
Job task counts. A common monitoring operation is to get task counts—the number of tasks per state. There can be many thousands of tasks in jobs, and calculating counts by using the List API can be slower than required. We've added support for quickly obtaining task counts for a job. See the REST API and C# API documentation for more details.
Exposure of ports on pool nodes so they can be addressed externally. We've enhanced network configuration to allow a list of NAT pools to be configured. You can specify protocol, ports, and network security group. See the REST API (will be added shortly) and C# API documentation for more details.
Corey - Today, we’re extending our partnership and will offer an increasing number of services directly supported by Terraform, including Azure Container Instances, Azure Container Service, Managed Disks, Virtual Machine Scale Sets and others. We want to give additional flexibility to express infrastructure-as-code and to enable many more native Microsoft Azure services to be easily deployed directly through Terraform. Learn more about the Azure provider for Terraform.

5. The Family R – Open Source … a.k.a. CRAN-R Microsoft R Client
6/11/2018 4:23 PM
The Family
R – Open Source … a.k.a. CRAN-R
Microsoft R Client
Microsoft R Server
Microsoft SQL Server R Services
Microsoft R Server for Hadoop / Spark
R tools for Visual Studio
Microsoft Machine Learning (MS-ML)
MRAN
© Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

6. 7 “Flavors” of R Explained
6/11/2018 4:23 PM
7 “Flavors” of R Explained
Engine
Licensing
Scalability
Support
Usage
Open Source R
Open Source
Single threaded, In-Memory
Community
It all starts here…
MS R Open
Intel-licensed math improvements
Basis for other MS R products
MS R Client
Open core, (o.s. + commercial) free dnld.
Parallelized but thread-constrained, memory-constrained.
Best desktop platform for developing scalable R
MS R Server Linux
Open core, licensed per core
Parallelized, In-memory and out-of-memory, multi-threaded
Commercial
Server-scale engines for production data science and deployment
SQL Server R Services
Open core, bundled into SQL Server Standard & Enterprise
Parallel scaling, in-memory (Std. edn.) and multi-process parallel (Ent. Edn.)
In-database or standalone Windows engine.
MS R Server for Hadoop
Open core
Parallel scaling across entire MapReduce or Spark cluster
Large-scale, mixed workload big-data analytics & deployment
MS R Server for Teradata
Parallel, out-of-memory scalability inside Teradata engine
In-EDW scalable analytics & deployment
© Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

7. Best-in-class Deployment Experience
Data Scientist
Easy Consumption
Explore and consume services in R directly
Data Scientist
Microsoft R Client
(mrsdeploy package)
Easy Deployment
Turn R into web services in one line of code
Microsoft R Server
configured for
operationalizing R analytics
Services / Sessions
getService
Microsoft R Client
(mrsdeploy package)
publishService
Apps
REST API calls
Easy Setup
In-cloud or on-prem
Adding nodes to scale
High availability & load balancing
Remote execution server
Developer
Easy Integration
Swagger-based APIs: easy to consume with any programming language

8. DEMO

9. Resources Azure Home Page Azure Blog Azure Updates
(Refer blogpost on swagger)