Build 2017 keynote day 1

Highlights:

My main takeaways:

  • serverless computing and the CosmosDB are going to allow you to quickly create new ultra scalable applications.
  • AI, cognitive services and Cortana skills will allow you to build some real Sci-Fi things.
  • Microsoft graph is improving and adding more features that will allow you to build more personal apps.

Keynote begins
Satya Nadella comes on stage

Satya is reflecting on the past, makes fun of the inverse relationship between the amount of his hair vs the amount of computing power he has.
clip_image001[6]
clip_image002[6]

There are unintended consequences of new technology. We all have a responsibility to ensure that the more dystopian scenarios do not come to fruition
clip_image003[6]

Need to build good principles into technology.
clip_image004[6]
clip_image005[6]

There are now 500 million devices running Windows 10.
140 million Cortana monthly active users http://www.zdnet.com/article/windows-10-installed-base-hits-500-million/
clip_image006[6]

The user experience is going to span multiple devices. Personal digital assistants will be across many platforms. Requires an “Intelligent Cloud & Intelligent Edge” http://www.zdnet.com/article/microsoft-build-2017-buzzword-bingo-on-the-edge/
Lots of devices (like autonomous cars) are generating 100GB/second of data. Need to move processing power to the edge where the data is.
Serverless computing is going to fundamentally change the economics of back end computing, and will be the core of a modern distributed cloud processing.
Will also need to have logic which can span between cloud & edge. Migrate to where the data is.

clip_image007[6]

 

Sam George comes on stage to deliver a demo
Big manufactuing devices, used to be disconnected and would have a regular maintenance scheduled.
Are now connected to the cloud, collecting telemetry and sending to IoT Hub, to do predictive maintenance. Can also predict a failure and shut it down.

clip_image008[6]

Wanted to put the same logic predicting a failure, and run it on the edge devices. IoT Edge allows you to run the same code in the cloud & push onto edge nodes. Can run on devices even smaller than a Raspberry Pi.
https://azure.microsoft.com/en-us/campaigns/iot-edge/

clip_image009[6]

clip_image010[6]

He selects some of the logic that is running in the cloud (Azure functions, etc.) it is then automatically packaged up into Docker containers, and pushed down to the edge nodes to improve response time.
clip_image011[6]
clip_image012[6]
clip_image013[6]
clip_image014[6]

 

Satya back on stage

clip_image015[6]

All accidents that happen in a workplace could avoided, if data was collected and processed.
Can hook on all the edge computing power together with their combined devices (cameras, etc.) to have an AI safety workplace.

https://www.youtube.com/watch?v=pL-c00M2CnI video promo [2mins]

clip_image016[6]
clip_image017[6]

 

Andrea Carl on stage (Director of Commercial Sales)
Wants to show how to digitise a workplace, by using cameras and cognitive services to detect people and objects in real time. And detect safety policy violations. The demo here does 27 million calculations per second.

clip_image018[6]

Workplace 1 has people & object recognition enabled
clip_image019[6]

Workplace 2 only has people recognition enabled
clip_image020[6]

In workplace 2, there is an orange jackhammer resting unsafely against the workbench. Enables object recognition, and a warning is sent
clip_image021[6]

Now there is a new employee on site
clip_image022[6]
clip_image023[6]

A visitor comes and tries to use the jackhammer
clip_image024[6]

 

Satya back on stage
Microsoft graph allows you to see the complexity between people, activities, and devices that they use.

clip_image025[6]
clip_image026[6]

 

Laura Jones Demo (Sr. Product Manager at Microsoft)
Intelligent meetings demo, that uses Microsoft Graph, cognitive services, Cortana, Teams.
They use a Cortana standalone microphone to ask about her day. Tells her about her daily schedule, is told her car fuel is low. Laura requests to send a message, requesting time off work. Sent a reminder off to her car
clip_image027[6]

Is told there is a traffic jam that will delay her. Asks if she’d like to notify the other attendees.
clip_image028[6]

Her reminder pops back up
clip_image029[6]

Sends a message to team channel
clip_image030[6]

They continue the meeting within Teams, with the other team members
clip_image031[6]

She walks into the meeting room, and her phone changes from phone call, into a companion app. Uses the Project Rome SDK. Also the new Microsoft Graph information about devices and activities.
clip_image032[6]

They used voice to request the Meeting bot in the channel to add a task
clip_image033[6]

She gets back to her desk, and Cortana pops up saying that they have auto generated meeting notes. Can also see a transcript of the meeting, with keywords detected and highlighted (more cognitive services coolness)
clip_image034[6]

 

Scott Gu on stage

There are so many new cloud technologies coming out all the time. Is difficult to keep abreast of them all, learning them all, keeping them secure
clip_image035[6]

Focusing Azure innovation on doing the heavy lifting, and making your dev life easier. Around scalability, privacy, etc.

 

Scott Hanselman on stage (Principal Program Manager)
Has customised his Azure dashboard. By pinning useful things.

clip_image036[6]

There is now an Azure shell that you can access directly from portal that has the Azure CLI preconfigured
clip_image037[6]

clip_image038[6]

He ran an Azure CLI command to spin up 1,000 VMs in a scaleset.

The Azure companion app on mobile, also has the CLI
clip_image039[6]

Uses Application Insights to see the application map, and drills into where exceptions are happening
clip_image040[6]

Can grab snapshots of production while it is running, to do a “live debug” without blocking produciton.
clip_image041[6]

clip_image042[6]

Azure security portal helping you keep on top of security issues
clip_image043[6]

 

Scott Gu back on stage

clip_image044[6]

SQL Server runs everywhere. On premises, and in the cloud. In docker containers, etc.
SQL Server has built in machine learning via R & Python
clip_image045[6]
clip_image046[6]

Has a new Azure database migration service. Can move SQL, Oracle, etc. data into SQL Azure
clip_image047[6]

2 new database as a service offerings: Postgres & MySQL
clip_image048[6]

 

Announcing Azure CosmosDB
http://www.zdnet.com/article/microsoft-debuts-azure-cosmos-db-a-superset-of-its-documentdb-service/

Guaranteed SINGLE DIGIT millisecond performance
Jet.com scaled their instance up to support 10 of trilliions of transactions in a single day

clip_image049[6]
clip_image050[6]
clip_image051[6]
clip_image052[6]

 

Rimma Nehme
Imagine you want to build an application that is highly available, accessed by users all over the world.
Took the sample data of the Marvel characters, with all the highly complex relationships between the characters.
Also put in a chatbot to ask questions.

clip_image053[6]

But how do you allow the data to be queuired super fast in near real time?

clip_image054[6]

clip_image055[6]

Just need to select the regions that the data should be replicated to

clip_image056[6]

Can select from 5 different consistency models

clip_image057[6]

Added a single line, to make her existing npm code that interacted with gremlin be able to work with Cosmos DB

clip_image058[6]

 

Scott Gu back on stage

clip_image059[6]

How to modernise applications as they shift to the cloud.
You can take any .net framework app, and containerise it.

clip_image060[6]

 

Maria Naggaga (Program Manager)
Takes an existing .Net application. And adds Docker support. It auto detects it requires IIS.
clip_image061[6]

She adds a line to enable WCF
clip_image062[6]

She right clicks, and configures a CI build. Can deploy to App Services, Service Fabric.
clip_image063[6]

Creates a VSTS build
clip_image064[6]

Deploys out to a 5 node service fabric cluster
clip_image065[6]

App Insights mapping works
clip_image066[6]

 

Scott Gu back on stage
Service Fabric is what most Microsoft services are being built on. It runs in Azure, as well on premises. Gives you maximum flexibility on where you deploy.

clip_image067[6]

clip_image068[6]

Serverless computing allows you to code in an event based way.
Can write code that is triggered off many different events
clip_image069[6]
clip_image070[6]
clip_image071[6]
clip_image072[6]

Azure stack allows you to deploy the same programming model in cloud and in certain places.
Gave example of a carnival cruise ship that can use Azure Stack for when the ship is out in the ocean, disconnected.
clip_image073[6]

 

Julia White (CVP)
Azure & Azure stack.
Same portal, and same deployed web apps & Azure functions on both.
Uses Azure Functions on the cruise ship, to process the ship telemetry and stores it. Ready to push up to the Azure cloud when there is connectivity.

clip_image074[6]

 

Scott GU

clip_image075[6]

App source for finding SaaS apps that are built on top of Azure, Graph, etc. Can put your own SaaS apps up there

clip_image076[6]

 

Harry Shum (ECP, AI & Research)
AI is central to the Intelligent cloud, Intelligent Edge.
Microsoft Cognitive Services.
What is making AI possible now? Big compute power, powerful algorithms, and massive data sets that can be trained on.
Won all awards at a recent competition for image recognition

clip_image077[6]
clip_image078[6]
clip_image079[6]

Can now train up custom datasets for speech, image, search. Useful for story scripts with fantasy names.
https://customvision.ai/

clip_image080[6]

Demoing custom vision service
Takes a photo of a leaf, and has it identified.
clip_image081[6]

Ran the classifier over her Azure Data Lake.
clip_image082[6]
clip_image083[6]

Added bot framework support for Bing Search
clip_image084[6]

Adaptive cards framework. To write your chatbot UI once
clip_image085[6]
clip_image086[6]
clip_image087[6]

 

Yina Arenas
Presentation translator addin for PowerPoint.
She spoke in Spanish, and had it displayed in English & Chinese VERY quickly.
http://translate.it/
https://aka.ms/powerpointtranslator

clip_image088[6]
clip_image089[6]
clip_image090[6]

clip_image091[6]

https://Microsoft.com/ai

Advertisements

2 thoughts on “Build 2017 keynote day 1

  1. Nicely done! The photos I took (of the keynote slides) aren’t as nice as you have. Are you OK if I copy some of the images and use as part of some presentations?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s