Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Innovating on AWS – Making it real

Innovating on AWS – Making it real

The goal of the session is to demonstrate some of the newer and exciting capabilities including as part of the AWS platform. In the first part of this session, we'll provide the audience with an overview of how CloudFormation is used to build out environments quickly and easily using the “infrastructure as code” model. We'll then dive into computer vision using Amazon Rekognition, to demonstrate how this can be used to accelerate the development of AI applications, maximizing time to market and minimizing the effort required.

Nicolas DAVID

May 01, 2018
Tweet

More Decks by Nicolas DAVID

Other Decks in Technology

Transcript

  1. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Nicholas ANSELL Professional Services Senior Consultant, WWPS Nicolas DAVID Professional Services Consultant, WWPS Innovating on AWS – Making it real
  2. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS Services accelerate customers in their quest to provide powerful, feature-rich, scalable solutions that are secure, repeatable and cost-effective.
  3. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Part One: Infrastructure as Code • Quickly and easily build out and maintain AWS environments and Services Part Two: Computer Vision • Automated CCTV Analysis • Facial Recognition: Celebrity Imposter Challenge! Agenda
  4. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Part One Infrastructure as Code
  5. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Infrastructure as Code ? Techniques, practices, and tools from software development applied to creating reusable, maintainable, extensible, and testable infrastructure.
  6. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Infrastructure as Code ? Goals • Immutable infrastructure • Fast Recovery • Atomic Deployments & Validation
  7. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Infrastructure as Code ? Structure Speed Health Security
  8. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS CloudFormation Gives developers and systems administrators an easy way to create and manage a collection of related AWS resources Understands dependencies and supports rollbacks and versioning Allows for reusable component design strategies Supports JSON and YAML formats
  9. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS Quick Start catalog https://aws.amazon.com/quickstart
  10. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS Quick Starts • Automated reference deployments on AWS • Help streamline installation and configuration • AWS Partner Network team • Primarily technical program • Focus on enabling customers
  11. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Broad spectrum of AWS Quick Starts
  12. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Broad spectrum of AWS Quick Starts
  13. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Broad spectrum of AWS Quick Starts
  14. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Broad spectrum of AWS Quick Starts
  15. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Starting simple Amazon VPC Architecture Modular and scalable virtual networking foundation on the AWS Cloud
  16. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Amazon VPC Architecture Quick Start • Up to four Availability Zones for high availability and disaster recovery • Separate subnets for unique routing requirements. For each Availability Zone, this Quick Start provisions one public subnet and one private subnet by default. • Additional layer of security. This Quick Start provides an option to create a network ACL protected subnet in each Availability Zone. These network ACLs provide individual controls that you can customize as a second layer of defense. • Independent routing tables configured for every private subnet to control the flow of traffic within and outside the Amazon VPC. • Highly available NAT gateways, where supported, instead of NAT instances. • Spare capacity for additional subnets, to support your environment as it grows or changes over time.
  17. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Amazon VPC Architecture Quick Start
  18. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. • 1 step Sign up, sign in • 2 steps Choose region and key pair • 4 steps Create VPC • 4 steps Create Internet gateway • 24 steps Create 12 subnets • 24 steps Create 4 NACLs • 16 steps Create 4 NAT gateways • 27 steps Create 9 route tables • 2 steps Create routes • Many more steps Add more stacks Sign up, sign in Choose region and key pair Launch Quick Start Reducing complexity Manual deployment VPC Quick Start
  19. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Adding your requirements SharePoint Server 2016 on AWS Automated gold-standard deployment for trial or production use
  20. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. SharePoint Server 2016 on AWS • Microsoft SharePoint Server 2016 farm on AWS, based on your choice of traditional or streamlined topology • Redundant servers for Active Directory Domain Services (AD DS), SQL Server 2014, and SharePoint Server 2016 in two Availability Zones for high availability • Amazon Virtual Private Cloud in each Availability Zone to provide a private, isolated section of the AWS cloud for your SharePoint Server farm and services • Microsoft Windows Server 2012 R2 with Active Directory Domain Services (AD DS) for user authentication and authorization, and Remote Desktop Gateway for remote Windows administration • Microsoft SharePoint Server 2016 instances for web front-end and application services • SQL Server 2014 instances configured in a Windows Server Failover Cluster (WSFC) for database services
  21. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. SharePoint Server 2016 on AWS Quick Start
  22. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Infrastructure as code & Immutable Infrastructure - Customer Use case The Information and eGovernment Authority (iGA) in Bahrain is responsible for the AWS environment for ~70 different government departments and ministries.
  23. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS CloudFormation Tips Or using AWS CLI: https://docs.aws.amazon.com/cli/latest/reference/cloudformation/estimate-template-cost.html
  24. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS CloudFormation Tools AWS CloudFormation Validation Pipeline Available as QuickStart ! https://docs.aws.amazon.com/solutions/latest/aws-cloudformation-validation-pipeline/
  25. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS CloudFormation Tools Troposphere Python library to create AWS CloudFormation descriptions https://github.com/cloudtools/troposphere CloudMapper CloudMapper creates network diagrams of AWS environments https://github.com/duo-labs/cloudmapper CloudFoundation The CLI tool for creating, managing, and deploying large AWS CloudFormation templates and projects https://github.com/jcolemorrison/cloudfoundation TaskCatBETA Test all the CloudFormation things ! https://github.com/aws-quickstart/taskcat
  26. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Part Two Computer Vision
  27. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. A brief history of Time Machine Learning 1980’s R&D slowed. Hardware too expensive and not powerful enough … now AWS Machine Learning Services Cloud Computing introduced pay- per-use, but ML development still expensive 2000’s GPUs first used for Machine Learning 1950’s The birth of Machine Learning
  28. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Machine Learning at Amazon Optimize Fulfilment Centres Suggest Products Amazon Echo Devices Pilot Autonomous Drones
  29. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Our mission is to “share our learnings and ML capabilities as fully managed services and put them into the hands of every developer and data scientist”
  30. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS Machine Learning Stack Frameworks & Infrastructure AWS Deep Learning AMI GPU (P3 Instances) Mobile CPU IoT (Greengrass) Vision: Rekognition Image Rekognition Video Speech: Polly, Transcribe Language: Lex, Comprehend, Translate Apache MXNet PyTorch Cognitive Toolkit Keras Caffe2 & Caffe TensorFlow Gluon API Driven ML Services Managed Platform Services Amazon Machine Learning Mechanical Turk Spark & EMR Amazon SageMaker AWS DeepLens
  31. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Amazon Machine Learning customers
  32. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Zoolz are an Advanced AWS partner. This Intelligent Cloud backup solution allows customers to backup and analyze photo images in the cloud. The service currently processes 3-4 million photo images per month using Amazon Rekognition and growing fast.
  33. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Amazon Rekognition Features Object and Scene Detection Facial Analysis Face Comparison Facial Recognition
  34. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. There are ~6570 hours of footage produced by each CCTV Camera in one year** As an example, London is estimated to have ~420,000 cameras Therefore, the CCTV cameras in London alone produce around 2.7 billion hours of CCTV footage in one year! CCTV analysis must be automated, it is not a nice-to-have! ** if 6hrs of non-activity is removed each day Scenario: Automated CCTV Analysis
  35. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Rekognition Video Demonstration
  36. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. aws s3 cp $FILE_PATH s3://$BUCKET/$FILENAME Amazon Rekognition Video - The Code aws rekognition get-label-detection --job-id "$JOBID" \ --endpoint-url "$ENDPOINT" --region "$REGION" aws rekognition start-label-detection --video "S3Object={Bucket="$BUCKET",Name="$FILENAME"}" \ --endpoint-url "$ENDPOINT" --region "$REGION” \ --notification-channel "SNSTopicArn=$SNSARN:AmazonRekognition-demo,RoleArn=$ROLEARN" \ { "Labels": [ { "Timestamp": 0, "Label": { "Confidence": 88.43360137939453, "Name": "Automobile" } …
  37. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Analyze pre-recorded or real-time video to identify which CCTV cameras captured something: • A person of interest • A disaster such as a fire or explosion Save time analyzing CCTV footage, jump to the point in the timeline that rekognition identified the information needed. Use CCTV or Drones to identify trends, or make real-time decisions based on volumes of people or cars in particular areas of a city Combine video analysis with GEO data to create searchable object databases Automated Video Analysis Use-cases
  38. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Amazon Rekognition Celebrity Imposter Challenge!
  39. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Real celebrity or imposter? Rekognition Response….. Imposter! Image Source: http://www.scarlettentertainment.com
  40. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Real celebrity or imposter? Rekognition Response….. Celebrity! "MatchConfidence": 100.0, … "Name": "Vin Diesel" Image Source: http://www.youtube.com
  41. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Real celebrity or imposter? Rekognition Response….. Celebrity! "MatchConfidence": 100.0, … "Name": "Katrina Kaif" Image Source: https://en.wikipedia.org
  42. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Real celebrity or imposter? Rekognition Response….. Celebrity! "MatchConfidence": 100.0, … "Name": "George Wassouf" Image Source: https://www.forbesmiddleeast.com
  43. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. aws s3 cp $FILE_PATH s3://$BUCKET/$FILENAME aws rekognition recognize-celebrities --image \ "S3Object={Bucket="$BUCKET",Name="$FILENAME"}” { "CelebrityFaces": [ { "MatchConfidence": 100.0, "Face": { … } "Confidence": 99.99882507324219, "Quality": { … } "Name": "Lewis Hamilton", "Urls": [ "www.imdb.com/name/nm2619894" ], … } Amazon Rekognition Celebrity Challenge - The Code
  44. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. AWS Services accelerate customers in their quest to provide powerful, feature-rich, scalable solutions that are secure, repeatable and cost-effective.
  45. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Want to get started? “aws cloudformation getting started” “aws rekognition getting started”
  46. © 2018, Amazon Web Services, Inc. or its affiliates. All

    rights reserved. Thank you! Please complete the session survey in the summit mobile app.