re:Invent 2018 Las Vegas recap - All new AWS services

Slide 1

Slide 1 text

Slide 2

Slide 2 text

5 0 , 0 0 0 + attendees 2 , 1 0 0 + technical sessions 1 0 0 , 0 0 0 + live stream registrations 2018

Slide 3

Slide 3 text

60minutes Over 100 Announcements © 2 0 1 8 , A m a z o n W e b S e r v i c e s , I n c . o r i t s a f f i l i a t e s . A l l r i g h t s r e s e r v e

Slide 4

Slide 4 text

Slide 5

Slide 5 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Agenda • Global Infrastructure • Compute • Storage and Analytics • Databases • AI/ML • Security • IoT • New Business Initiatives (Robots, Space)

Slide 6

Slide 6 text

What do builders want?

Slide 7

Slide 7 text

Slide 8

Slide 8 text

Slide 9

Slide 9 text

Slide 10

Slide 10 text

Slide 11

Slide 11 text

Amazon Global Network • Redundant 100GbE network • Redundant private capacity between all Regions except China Over 150 Global CloudFront PoPs 89 Direct Connect Locations a e o q i h Paris Sweden AWS GovCloud East First 5 years: 4 regions 2016–2020: 13 regions Next 5 years: 7 regions A W S REGIONAL EXPANSION 1 9 R e g i o n s 5 7 A Z s d m c g b n s k v i i i i i i i i Milan i Cape Town

Slide 12

Slide 12 text

N E W !

Slide 13

Slide 13 text

AWS Global Accelerator: Before and After

Slide 14

Slide 14 text

Slide 15

Slide 15 text

N E W !

Slide 16

Slide 16 text

Before Transit Gateway

Slide 17

Slide 17 text

After Transit Gateway

Slide 18

Slide 18 text

Slide 19

Slide 19 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. EC2 Instance Types Burstable T 3 Big Data Optimized H 1 Memory Optimized R 5 High I/O I 3 Compute Intensive C 5 Graphics Intensive G 3 General Purpose GPU P 3 Memory Intensive X 1 X 1 e General Purpose M 5 V i r t u a l P r i v a t e S e r v e r s Bare Metal High I/O I 3 m Dense Storage D 2 F 1 FPGA A m a z o n L i g h t s a i l High-Memory Intensive Z 1 Powered by M 5 a R 5 a • Choose between processors on AWS general purpose and memory optimized instances • 10% lower prices on AMD-based instances • Most applications can run on AMD-based variants with little to no modification M 5 d R 5 d C 5 d Z 1 d • NVMe-based SSD block level instance storage physically connected to the host server • High-speed, low latency local block storage • EBS PIOPS to 1GB/s (64,000)

Slide 20

Slide 20 text

N E W ! Up to 45% lower cost for scale-out workloads

Slide 21

Slide 21 text

Slide 22

Slide 22 text

N E W !

Slide 23

Slide 23 text

Slide 24

Slide 24 text

N E W ! Add-on capability Scale to thousands of cores Native integration on C5n.9xl, C5n.18xl, and P3dn.24xl instance supports MPI and libfabric

Slide 25

Slide 25 text

Management Tools Introducing Predictive Scaling Ramp capacity before you need it with Predictive Scaling Time On-premise capacity provisioning Load/Capacity Time Capacity provisioning with target tracking Load/Capacity Time Capacity provisioning with predictive scaling and target tracking Load/Capacity © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.

Slide 26

Slide 26 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Hibernate EC2 Instances • Signals OS to hibernate (suspend to disk) • RAM is saved to root FS, EBS remains attached • EBS root volume must be encrypted and large enough • Can be used to prewarm instance that take long to boot • Hibernated instances can be stopped https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-instance-lifecycle.html

Slide 27

Slide 27 text

Slide 28

Slide 28 text

Slide 29

Slide 29 text

N E W ! App Mesh with Amazon ECS and Amazon EKS to better run containerized microservices at scale Microservice A Microservice B Microservice C Microservice A Microservice B Microservice C Microservice C Microservice C Before After

Slide 30

Slide 30 text

N E W ! Easily create and maintain custom maps of your applications

Slide 31

Slide 31 text

Introducing AWS Marketplace for Containers AWS Marketplace Software Container products available in AWSMarketplace Choose from more than 160 curated and trusted container products in AWS Marketplace and run them on AWS Container product images are verified and scanned Deploy container products on Amazon ECS, AWS Fargate, or EKS Products are available with free, bring your own license, and usage-based pricing models AWS Marketplace for Containers

Slide 32

Slide 32 text

Slide 33

Slide 33 text

N E W ! GE N E R AL L Y AVAI L AB L E A W S C l o u d 9 AWS Toolkit for PyCharm AWS Toolkit for IntelliJ AWS Toolkit for VS Code GENERALLY AVAILABLE IN DEVELOPER PREVIEW IN DEVELOPER PREVIEW Open source toolkits meeting you where and how you like to work AWS Toolkits for popular IDEs + IDEs

Slide 34

Slide 34 text

N E W ! Languages Lambda support for Ruby + Bring any Linux compatible language runtime; Powered by new Runtime API - Codifies the runtime calling conventions and integration points Same technology powering Ruby support in AWS Lambda Bring any Linux compatible language runtime Custom Runtimes + AWS OPEN SOURCE o f f e r e d b y o f f e r e d b y o f f e r e d b y o f f e r e d b y PARTNER SUPPORTED https://aws.amazon.com/about-aws/whats-new/2018/11/aws-lambda-now-supports-custom-runtimes-and-layers/

Slide 35

Slide 35 text

N E W ! Extend the Lambda execution environment with any binaries, dependencies, or runtimes Lambda Layers BUSINESS LOGIC LIB A LIB B BUSINESS LOGIC LIB A LIB B BUSINESS LOGIC LIB A LIB B BUSINESS LOGIC LIB A LIB B Programming Model Before BUSINESS LOGIC BUSINESS LOGIC BUSINESS LOGIC BUSINESS LOGIC LIB A LIB B After

Slide 36

Slide 36 text

N E W ! Programming Model Store, share, and deploy serverless applications Serverless Application Repository Compose application architectures from reusable building blocks Nested Applications using Serverless Application Repository Deploy new architectures as a set of serverless apps (nesting) Foster best organizational practices and reduce duplication of effort Share components, modules and full applications privately with teams or publicly with others to improve agility + + Integrate Lambda functions into existing web architectures ALB Support for Lambda AWS Lambda Applicatio n Load Balancer AWS Fargate Amazon EC2

Slide 37

Slide 37 text

N E W ! MOBILE APPS CHAT DASHBOARDS IoT DEVICES Amazon API Gateway WebSockets API LAMBDA FUNCTIONS PUBLIC ENDPOINTS ON AMAZON EC2 AMAZON KINESIS ANY OTHER AWS SERVICE Stateful connection A L L P U B L I C L Y A C C E S S I B L E E N D P O I N T S Stateful connection Programming Model This new type of API will enable customers to build real-time two way communication applications backed by Lambda functions or other API Gateway integrations. Web Socket support for API Gateway

Slide 38

Slide 38 text

N E W ! WorkFlow Step Functions Glue AWS services together without writing code Step Functions API Connectors Amazon ECS AWS Fargate Amazon DynamoD B Amazon SNS AWS Batch Amazon SQS AWS Glue Amazon SageMak er +

Slide 39

Slide 39 text

N E W ! Secure

Slide 40

Slide 40 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Firecracker Architecture and Benefits • Firecracker microVMs have the same security as KVM VMs • Designed for low overhead, high density, and fast start times • Built-in fair sharing

Slide 41

Slide 41 text

Slide 42

Slide 42 text

Slide 43

Slide 43 text

Amazon EBS Performance Increases Storage Achieve up to 16,000 IOPS and 256 MB/s of throughput per gp2 volume genaral purpose SSD 2x performance increase for io1 and 60% performance increase for gp2volumes Use fewer volumes and operational staff to achieve maximum performance for your applications Achieve up to 64,000 IOPS and 1,000 MB/s of throughput per io1 volume provisioned IOPS

Slide 44

Slide 44 text

Introducing Amazon S3Intelligent-Tiering Storage Anew storage class that automatically optimizes storage costs for data with unknown or changing access patterns

Slide 45

Slide 45 text

Introducing Amazon EFS Infrequent Access Storage Lower-priced storage for infrequently accesseddata Cost Optimized for Infrequent Access Lower price than standard tier Easy to Manage Users configure lifecycle policies between classes Single File System Store infrequently and frequently accessed files in the same data set

Slide 46

Slide 46 text

Storage Introducing AWS DataSync Online transfer service that simplifies, automates, and accelerates movingdata Transfers up to 10 Gbps per agent Pay as you go Simple data movement to S3 or EFS Secure and reliable transfers Replicate data to AWSfor business continuity Transfer data for timely in-cloud analysis Migrate active application data toAWS Combines the speed and reliability of network acceleration software with the cost-effectiveness of open source tools AWS integrated AWS

Slide 47

Slide 47 text

Introducing AWS Transfer for SFTP Storage Fully managed service makes it easy to integrate SFTP-based file transfers into AWS Move your existing SFTP workflows to AWS in 3steps Seamless migration of existing workflows Data available for archiving and processing in S3 Simple to use Cost effective Fully managed, highly available, and elastically scalable 1 2 3

Slide 48

Slide 48 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Set up Storage 1 Move data 2 Cleanse and prep data 3 Configure and enforce security and compliance policies 4 Make data accessible for analytics 5 Steps for building and managing a data lake

Slide 49

Slide 49 text

N E W ! Enforce security policies across multiple services Gain and manage new insights Move, store, catalog, and clean your data faster with machine learning A service that allows you to build a secure data lake in days Amazon Lake Formation

Slide 50

Slide 50 text

Slide 51

Slide 51 text

N E W ! Amazon FSx for Windows File Server Windows native for fully compatible Windows File System experience No hardware or software to manage Secure and compliant including PCI-DSS, ISO, and HIPAA Up to 10s of GB/s throughput with sub- millisecond latencies (Compatibility with AD, Windows access control, and native Windows Explorer experience) Fully managed Windows file system built on native Windows file servers

Slide 52

Slide 52 text

N E W ! High throughput, low latency – 100s of GBs/s and millions of IOPS Seamless integration with Amazon S3 Secure and compliant including PCI-DSS, ISO and HIPAA Fully managed file system for high compute intensive workloads Amazon FSx for Lustre

Slide 53

Slide 53 text

N E W ! Fully managed, highly- available Kafka clusters Highly available with rolling upgrades Fully compatible – Run your Kafka applications with zero code changes Fully managed and highly available Apache Kafka service Amazon Managed Streaming for Kafka

Slide 54

Slide 54 text

Slide 55

Slide 55 text

Slide 56

Slide 56 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Provisioning capacity for DynamoDB ( S o m e t i m e s i t ’ s h a r d t o k n o w w h a t ’ s b e s t ) HIGH-SCALE APPLICATIONS Estimating how much throughput capacity to provision can be guesswork Not enough experience with app can cause unexpected extreme app usage Spikey traffic can be costly to maintain availability and performance Auto-scaling can cause lag time apps can’t afford

Slide 57

Slide 57 text

N E W ! Amazon DynamoDB Read/Write Capacity On Demand N o m o r e c a p a c i t y p l a n n i n g – p a y o n l y f o r w h a t y o u u s e No capacity planning No need to specify how much read/write throughput you expect to use Pay only for what you use Pay-per-request pricing Ideal for unpredictable workloads Ramp from zero to tens of thousands of requests per second on demand

Slide 58

Slide 58 text

N E W ! Amazon DynamoDB Transactions N a t i v e , s e r v e r - s i d e s u p p o r t f o r t r a n s a c t i o n s

Slide 59

Slide 59 text

Existing time-series databases Relational databases Difficult to scale Manual effort needed for enterprise-grade availability and reliability Limited data lifecycle management capabilities Unnatural for time-series data Rigid schema inflexible for fast-moving time- series data Building with time-series data is challenging Lack time-series analytic functions like smoothing, approximation, and interpolation 1 2 3 Clickstream data IoT Sensor Readings DevOps data

Slide 60

Slide 60 text

N E W ! Timestream is 1,000X faster and 1/10th the cost of relational databases Trillions of daily events Serverless Time-series analytics (interpolation, smoothing, approximation) built in Multiple orders of magnitude improvement in query performance Fast, scalable, fully managed time series database Amazon Timestream

Slide 61

Slide 61 text

Slide 62

Slide 62 text

TRANSACTIONS WITH DECENTRALIZED TRUST 2 \ DMV Track vehicle title history Manufacturers Track distribution of a recalled product HR & Payroll Track changes to an individual’s profile Healthcare Verify and track hospital equipment inventory Common customer use cases on Blockchain Relational databases Blockchain frameworks Using sub-optimal ways to solve the problem LEDGERS WITH CENTRALIZED TRUST 1

Slide 63

Slide 63 text

Mortgage Lenders Processing syndicated loans Small Businesses Transact with suppliers and distributers Retail Streamline customer rewards Financial Institutions Peer-to-peer payments Blockchain Frameworks Hard to scale Set up is hard Complicated to manage Expensive TRANSACTIONS WITH DECENTRALIZED TRUST 2 LEDGERS WITH CENTRALIZED TRUST 1 Common customer use cases on Blockchain

Slide 64

Slide 64 text

Two needs, what to do?

Slide 65

Slide 65 text

N E W ! Cryptographically Verifiable All changes are cryptographically chained and verifiable Transparent Full visibility into entire data lineage Immutable Append-only, immutable journal tracks history of all changes Highly scalable Automatically scale up or down Easy to use Query with familiar SQL operators Fast Execute 2-3X more transactions Fully managed ledger database that provides a transparent, immutable, cryptographically verifiable transaction log owned by a central trusted authority Amazon Quantum Ledger Database

Slide 66

Slide 66 text

N E W ! Create and manage scalable blockchain networks Amazon Managed Blockchain Choose Hyperledger Fabric or Ethereum Create blockchain networks with a few clicks; manage them with simple API calls Scales to support thousands of applications running millions of transactions Easy to move data into QLDB for further analysis

Slide 67

Slide 67 text

Slide 68

Slide 68 text

Slide 69

Slide 69 text

Slide 70

Slide 70 text

Slide 71

Slide 71 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. FRAMEWORKS INFRASTRUCTURE P3 P3dn INTERFACES Infrastructure and frameworks for Machine Learning C O M I N G S O O N Scale model training performance across multiple instances 100 Gbps networking 64 GB GPU memory 768 RAM (System Memory) P3dn.24xl 3X faster network throughput than any other provider 2X as much GPU memory than any other GPU instance from other major providers 100+ GB more system memory than the next largest instance available from other providers “In the cloud, 85% of TensorFlow workloads run on AWS" Nucleus Research, TensorFlow on AWS, 28 November 2018, Analyst: Rebecca Wettemann

Slide 72

Slide 72 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Running inference in production drives the majority of cost for Machine Learning Inference Training Infrastructure costs But what about inference? It’s never been easier, faster, or more cost- effective to train Machine Learning models

Slide 73

Slide 73 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Two main drivers of inference inefficiency: complexity and cost of Machine Learning inference today Elasticity is important One size does not fit all

Slide 74

Slide 74 text

N E W ! EC2 Instance EC2 Instance EC2 Instance GPU Add GPU acceleration to any Amazon EC2 instance for faster inference at much lower cost (up to 75% savings) Amazon Elastic Inference P 3 . 8 X L P 3 P 3 P 3 P 3 Amazon Elastic Inference 36 TOPS GPU M5.large Amazon Elastic Inference

Slide 75

Slide 75 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Starting at 1 TFLOPS Any instance family Simple speech and language models Up to 32 TFLOPS Recommendation engines or fraud detection models Provision Elastic Inference capacity inside VPC 360,000 ResNet-50 Computer vision deep learning model images per hour, inference $0.22 per hour on medium EI accelerator 75% lower cost L O W E S T C O S T A V A I L A B L E Elastic Inference - Scale & Cost

Slide 76

Slide 76 text

Workloads needing entire GPU or that are latency sensitive

Slide 77

Slide 77 text

N E W ! High throughput Low latency Hundreds of TOPS Multiple data types Multiple ML Frameworks INT8, FP16, mixed precision TensorFlow, MXNet, PyTorch, Caffe2, ONNX EC2 instances Amazon SageMaker Amazon Elastic Inference High performance machine learning inference chip, custom designed by AWS AWS Inferentia

Slide 78

Slide 78 text

Slide 79

Slide 79 text

N E W ! Human annotations Data in S3 Automatic annotations Training data Simple, pre-built workflows Mechanical Turk Your own employees Third party vendors Active Learning model >80% confidence <80% confidence Build highly accurate training datasets and reduce data labeling costs by up to 70% using Machine Learning Amazon SageMaker - Ground Truth

Slide 80

Slide 80 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon SageMaker: broad set of built-in algorithms Designed to be 10x faster K-Means Clustering Principal Component Analysis Neural Topic Modelling Factorization Machines Linear Learner (Regression) BlazingText Reinforcement Learning XGBoost Topic Modeling (LDA) Image Classification Seq2Seq Linear Learner (Classification) DeepAR Forecasting A L G O R I T H M S Built into Amazon SageMaker Improving and expanding continually 40% more algorithms added since SageMaker launch

Slide 81

Slide 81 text

N E W ! Natural Language Processing Computer Vision Speech Recognition Text Clustering Text Generation Text Classification Grammar and Parsing Named Entity Recognition Text to Speech Handwriting Recognition Object Detection in Images 3D Images Text OCR Video Classification Speaker Identification Ranking Regression Anomaly Detection Register with AWS Marketplace Automatic validation on SageMaker Package algorithm, models and configuration Self-service listing on AWS Marketplace Browse or search AWS Marketplace Subscribe in a single click Available through Amazon SageMaker Over a hundred algorithms and models that can be deployed directly to Amazon SageMaker AWS Marketplace for Machine Learning S e l l i n g a l g o r i t h m s & m o d e l s o n A W S M a r k e t p l a c e

Slide 82

Slide 82 text

N E W ! Increase or decrease the number of nodes available during training Dynamic Training with MXNet

Slide 83

Slide 83 text

Slide 84

Slide 84 text

Slide 85

Slide 85 text

Slide 86

Slide 86 text

N E W ! Fully managed reinforcement learning algorithms TensorFlow, MXNet, Intel Coach, and Ray RL 2D and 3D simulation environments via OpenGym Simulate environments with Amazon Sumerian and AWS RoboMaker Example notebooks and tutorials New machine learning capabilities in Amazon SageMaker to build, train, and deploy with Reinforcement Learning Amazon SageMaker RL

Slide 87

Slide 87 text

Can we help developers get rolling with Reinforcement Learning? (literally)

Slide 88

Slide 88 text

N E W ! Fully autonomous 1/18th scale race car, driven by Reinforcement Learning AVAILABLE FOR PRE-ORDER ON AMAZON.COM AWS DeepRacer

Slide 89

Slide 89 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. HD Video Camera Gyroscope for direction and orientation Accelerometer for measuring change in speed Two batteries: one to power on-board compute, one to drive motors Dual-core Intel Atom® Processor Introducing AWS DeepRacer All wheel drive, monster truck chassis Suspension mounted high for a view of the road Both accelerometer and gyroscope useful in the future for building more sophisticated models such as finding the perfect racing line or path finding

Slide 90

Slide 90 text

Slide 91

Slide 91 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Build reinforcement learning model DeepRacer League Races at AWS Summits Winners of each DRL Race and top points getters compete in Championship Cup at re:Invent 2019 Virtual tournaments through the year AWS DeepRacer League World’s first global autonomous racing league, open to anyone

Slide 92

Slide 92 text

Slide 93

Slide 93 text

N E W ! Activity stream from app Views, signups, conversion, etc. Inventory Articles, products, videos, etc. Demographics (optional) LOAD DATA (EMR Cluster) INSPECT DATA IDENTIFY FEATURES SELECT ALGORITHMS SELECT HYPERPARAMETERS TRAIN MODELS OPTIMIZE MODELS HOST MODELS BUILD FEATURE STORE CREATE REAL-TIME CACHES Customized personalization & recommendation API F u l l y m a n a g e d b y A m a z o n P e r s o n a l i z e Amazon Personalize Age, location, etc. Real-time personalization and recommendation service, based on the same technology used at Amazon.com Amazon Personalize

Slide 94

Slide 94 text

N E W ! Amazon Forecast Any historical time-series Integrates with SAP and Oracle Supply Chain Custom forecasts with 3 clicks 50% more accurate 1/10th the cost Integrates with Amazon Timestream Retail demand Travel demand AWS usage Revenue forecasts Web traffic Advertising demand Generate forecasts for: Accurate time-series forecasting service, based on the same technology used at Amazon.com

Slide 95

Slide 95 text

N E W ! OCR++ service to easily extract text and data from virtually any document Amazon Textract

Slide 96

Slide 96 text

N E W ! Extract information from unstructured medical text accurately and quickly No machine learning experience required Amazon Comprehend Medical

Slide 97

Slide 97 text

Slide 98

Slide 98 text

Slide 99

Slide 99 text

N E W !

Slide 100

Slide 100 text

Introducing AWS SecurityHub Security Public Beta Centrally view and manage security alerts & automate compliance checks • Enabled in minutes to aggregate security findings from AWS and Partner services across your accounts • Quickly assess security and compliance in one location and take action on findings • Built-in and customizable insights help you track security issues that are unique to your environment • Improve compliance with automated, continuous account-level configuration and compliance checks

Slide 101

Slide 101 text

Introducing AWS Control Tower Management Tools Limited Preview Set-up a multi-account environment in a single location to govern AWS workloads Automate the creation of a landing zone with best practice blueprints AWS Control Tower automates the set-up of a well-architected multi-account environment with best practices, and guides you through a step-by-step process to customize it to your organization Guardrails for policy enforcement Control Tower offers curated guardrails. Guardrails are high-level rules that provide on-going governance for your overall AWS environment Dashboard for continuous visibility The Control Tower dashboard gives you continuous visibility into your AWS environment. You can view the number of organizational units and accounts provisioned, guardrails enabled, and the compliance status of your enabled guardrails

Slide 102

Slide 102 text

N E W ! AWS coming to a data center near you in two ways AWS Outposts VMware Cloud on AWS AWS Native AWS DESIGNED HARDWARE The same that we run in our own data centers OPTION 1 OPTION 2

Slide 103

Slide 103 text

N E W ! AWS Well- Architected Tool Measure and improve your architecture using AWS Well-Architected best practices Implement workplans to improve your architecture Stay up to date as your architecture evolves Review workloads against best practices

Slide 104

Slide 104 text

N E W !

Slide 105

Slide 105 text

N E W !

Slide 106

Slide 106 text

N E W ! Easily develop, test, and deploy intelligent robotics applications AWS RoboMaker Development Environment Simulation Cloud Extensions for ROS Fleet Management

Slide 107

Slide 107 text

© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Space Business? • Build ground stations, sign long-term leases, and purchase excess bandwidth • Expensive to build and difficult to maintain and require high CAPEX investment to scale, and experience network latency and scheduling conflicts • Have opaque investing and pricing for startups

Slide 108

Slide 108 text

N E W ! Easily control satellites and ingest data with fully managed Ground Station as a Service AWS Ground Station Low Earth Orbit (LEO) Medium Earth Orbit (MEO) Simultaneous narrowband S-band, X-band and UHF downlink Receive satellite data into Amazon VPC and Process data in AWS Cloud