AWS re:Invent 2016 Fast Forward

82
AWS re:Invent 2016 ▶▶

Transcript of AWS re:Invent 2016 Fast Forward

Page 1: AWS re:Invent 2016 Fast Forward

AWS

re:Invent 2016

▶▶

Page 2: AWS re:Invent 2016 Fast Forward

The CrewAdios Tsai, Angela Tsai, Gaod Chen,

Gdx Wu, Drake Guan, Eric Tsai

KKTV / KKStream / KKBOX

Page 3: AWS re:Invent 2016 Fast Forward

IntroductionTracks

Questions

Page 4: AWS re:Invent 2016 Fast Forward

What isAWS re:Invent?

When

Where

Who

How

Why

Page 6: AWS re:Invent 2016 Fast Forward
Page 7: AWS re:Invent 2016 Fast Forward
Page 8: AWS re:Invent 2016 Fast Forward
Page 9: AWS re:Invent 2016 Fast Forward
Page 10: AWS re:Invent 2016 Fast Forward
Page 11: AWS re:Invent 2016 Fast Forward
Page 12: AWS re:Invent 2016 Fast Forward

Coverage of AWS re:Invent 2016

● Breakout Sessions● Hands-on Labs● Hackathon● Alexa Competition● Exhibition● AWS Certification Exams● Executive Briefing Center (EBC)● Executive-only Sessions● re:Play / Harley Ride / Pub Crawl / 5K running / ...

Page 13: AWS re:Invent 2016 Fast Forward
Page 14: AWS re:Invent 2016 Fast Forward
Page 15: AWS re:Invent 2016 Fast Forward
Page 16: AWS re:Invent 2016 Fast Forward
Page 18: AWS re:Invent 2016 Fast Forward
Page 19: AWS re:Invent 2016 Fast Forward
Page 20: AWS re:Invent 2016 Fast Forward
Page 21: AWS re:Invent 2016 Fast Forward
Page 22: AWS re:Invent 2016 Fast Forward
Page 23: AWS re:Invent 2016 Fast Forward
Page 25: AWS re:Invent 2016 Fast Forward

Breakout Sessions

Page 27: AWS re:Invent 2016 Fast Forward

Tracks of Breakout Sessions● Life Sciences● World Wide Public Sector● Windows● Storage● Serverless Computing Mini Con● Security Services Mini Con● Security & Compliance● Networking● Mobile● Media & Entertainment● Machine Learning Mini Con● Business Apps● IoT Mini Con

● Gaming● Financial Services● Developer Community● Architecture● Alexa● Big Data & Analytics● Healthcare● DevOps & Tools● Databases● Content Delivery● Containers Mini Con● Compute● Big Data Mini Con

Page 28: AWS re:Invent 2016 Fast Forward

New AWS Services

Page 29: AWS re:Invent 2016 Fast Forward

Computing

Page 30: AWS re:Invent 2016 Fast Forward

New EC2 Instance Types

Page 31: AWS re:Invent 2016 Fast Forward

EC2 Elastic GPUs

● Attach GPU to EC2 instances in much the same way as EBS.

● Elastic GPUs plan to support Open Graphics Library (OpenGL).

src: https://www.skyarch.net/blog/?p=9127

Page 32: AWS re:Invent 2016 Fast Forward

Amazon Lightsail (Winter is coming!)● As cheap as DigitalOcean!● Best usage and business for old machines in AWS.

Page 33: AWS re:Invent 2016 Fast Forward

EC2 Systems Manager

Null

Page 34: AWS re:Invent 2016 Fast Forward

● Plans, schedules, and executes your batch computing workloads

● Fully managed● No additional charge

● Dynamic provisioning and scaling

● EC2 and EC2 Spot● Priority-based queues and

scheduling

Page 35: AWS re:Invent 2016 Fast Forward

AWS Step Functions

Page 36: AWS re:Invent 2016 Fast Forward

AWS Step Functions

Page 37: AWS re:Invent 2016 Fast Forward

AWS Step Functions● Amazon States Language (https://states-language.net/spec.html)

"HelloWorld": { "Type": "Task", "Resource": "arn:aws:lambda:us-east-1:123456789012:function:HelloWorld", "Next": "NextState", "Comment": "Executes the HelloWorld Lambda function"}

● Step Functions := microservice + distributed apps + visual workflow.● SWF (Simple Workflow) killer.● Service limits.● Tokyo region!

Page 38: AWS re:Invent 2016 Fast Forward

● Run applications and services without having to manage infrastructure.● No longer have to provision, scale, and maintain servers to run your

applications, databases, and storage systems.

Page 39: AWS re:Invent 2016 Fast Forward
Page 40: AWS re:Invent 2016 Fast Forward

Live Demo

Page 41: AWS re:Invent 2016 Fast Forward

Lambda@Edge● Lambda functions’ invocation at the Edge (CloudFront).● For inspection and alteration of HTTP headers, access control (requiring

certain cookies to be present), device detection, A/B testing, expedited or special handling for crawlers or ‘bots, and URL rewriting.

● Limits:

VS

Page 42: AWS re:Invent 2016 Fast Forward

Storage

Page 43: AWS re:Invent 2016 Fast Forward

AWS Snowball EdgePetabyte-scale data transport with on-board storage and compute

● 100TB per edge.● S3 compatible with NFS.● Snowball cluster.● Lambda functions on-board!● Data encrypted.

Page 44: AWS re:Invent 2016 Fast Forward

AWS Snowmobile

src: https://www.datapipe.com/blog/2016/12/01/aws-reinvent-day-two-recap/

Page 46: AWS re:Invent 2016 Fast Forward

AWS Snowmobile

● Up to 100PB, bigger than Snowball [Edge], and much mobile (because it’s a truck).

● Load data S3 or Glacier.● Thinking about migrating your

data center to AWS? Try this!● Unfortunately, US only.

Page 47: AWS re:Invent 2016 Fast Forward

PostgreSQL “Compatible” Aurora

● 2x the performance of PostgreSQL.● pgSQL and PostGIS.● Aurora: 10GB to 64TB w/o pre-provisioning; 6 copies across 3 AZ; read with

up to 15 replicas; ...

src: http://dev.classmethod.jp/cloud/aws/reinvent2016-postgre-aurora/

Page 48: AWS re:Invent 2016 Fast Forward

DevOps● Building software is a mindset not a skillset.● Development and operations teams are no longer siloed.

Page 49: AWS re:Invent 2016 Fast Forward

AWS Organizations

● Centrally manage groups of accounts.● Control individual account permissions at scale.● Automate AWS account creation.● Simplify billing --- consolidated billing for groups of accounts.● Free!

Page 50: AWS re:Invent 2016 Fast Forward

AWS CodeBuild

● Lightweight Jenkins, focusing on building.● Managed building environment, configurations, and other build specs.● Repo: GitHub, AWS CodeCommit, S3.● Env: Android, Java, Python, Go, Node.js or Docker. (Others are erased)

Page 51: AWS re:Invent 2016 Fast Forward

src: https://www.infoq.com/news/2016/12/AWS-DDoS-AWS-Shield

Page 52: AWS re:Invent 2016 Fast Forward

src: http://bit.ly/2hLQ76w

Page 53: AWS re:Invent 2016 Fast Forward

AWS Shield: Managed DDoS Protection

Standard Protection● Quick detection with always-on network

flow monitoring.● Automated mitigation techniques to protect

against common, frequently occurring Layer 3 and 4 attacks.

● Available to ALL AWS customers at NO additional cost.

● ELB, CF and Route 53.

Advanced Protection● Intelligent DDoS attack detection.● 24x7 DDoS response team.● AWS WAF for proactive rules.● Bill protection against spikes due to DDoS

attack for ELB, CF and R53.● ...

Page 54: AWS re:Invent 2016 Fast Forward

● ●

Page 55: AWS re:Invent 2016 Fast Forward

AWS Personal Health Dashboard

Null

Page 56: AWS re:Invent 2016 Fast Forward

Analytics

EMRElasticsearchAmazon CloudSearchData PipelineKinesisRedshiftAthenaGlueQuickSightPinpoint

Page 57: AWS re:Invent 2016 Fast Forward

● Ad-hoc Queries Against S3● Serverless● $5 per Terabyte

● Built on Presto● Runs Standard SQL● Parallel● Fast

Page 58: AWS re:Invent 2016 Fast Forward

● Fully Managed ETL Service● Build Your Data Catalog● Generate and Edit

Transformations● Schedule and Run Your Jobs

● Integrated with○ S3, RDS, Redshift○ JDBC-compliant Data

Store● Handles Common Errors● Git Integration

Page 59: AWS re:Invent 2016 Fast Forward

● Visualizations● Ad-hoc Analysis● Fast● Easy to Use● Low Cost● Mobile Friendly

● Share and Collaborate● Integrated with

○ S3, RDS, Redshift○ On Premise DB○ SaaS Applications

(e.g., Salesforce)

Page 60: AWS re:Invent 2016 Fast Forward

● Understand User Behavior● Create Targeted Campaigns● Measure Results● Rich, In-app, and Silent

Notifications

● Scheduled and recurring Campaigns

● A/B Testing● Segment and Funnel● Custom Events and Attributes● Export to S3

Page 61: AWS re:Invent 2016 Fast Forward

Artificial Intelligence

Machine LearningRekognitionPollyLex

Page 62: AWS re:Invent 2016 Fast Forward

Amazon EC2 P2 Instances and Deep Learning AMI

P2 Instances

● GPU-based parallel compute capabilities

Deep Learning AMI

● MXNet● Caffe● Tensorflow● Theano● Torch

Page 63: AWS re:Invent 2016 Fast Forward

● Detect Objects, Scenes, and Faces

● Facial Analysis● Search and Compare Faces

● Same Image Analysis Technology of Prime Photo

Page 64: AWS re:Invent 2016 Fast Forward
Page 65: AWS re:Invent 2016 Fast Forward

● Build Speech-enabled Products

● Text to Speech (TTS)● 47 Lifelike Voices● 24 Languages

● Same TTS Technology of Alexa's Voice

Page 66: AWS re:Invent 2016 Fast Forward

● Automatic Speech Recognition (ASR)

● Natural Language Understanding (NLU)

● Same ASR and NLU Technology of Alexa

Page 67: AWS re:Invent 2016 Fast Forward
Page 68: AWS re:Invent 2016 Fast Forward

長持ちの絵に、生麦、生米、生卵。なた豆七つぶ生米七つぶ、七つぶなた豆、七つぶ生米東京特許許可局局長 急遽許可却下ちなみに桃はバラ科モモ属の落葉樹の果実ですけど、スモモはバラ科サクラ属で、まあ同じバラ科ですが厳密にはスモモはモモのうちではないそうですよ♪

Page 69: AWS re:Invent 2016 Fast Forward
Page 70: AWS re:Invent 2016 Fast Forward
Page 71: AWS re:Invent 2016 Fast Forward
Page 72: AWS re:Invent 2016 Fast Forward

Alexa, shutdown VideoPass.Alexa, slack daily.Alexa, build an Alexa.

Page 73: AWS re:Invent 2016 Fast Forward

Others

Page 74: AWS re:Invent 2016 Fast Forward

Amazon AppStream 2.0● Stream your Windows applications to any

devices.● HTML5 browser/player.● AppStream STreaming eXperience (STX)

:= H.264 + Opus + (HLS?).● No DRM but HTTPS.● Competitor to VMFive, Ubitus, … ?!

Page 75: AWS re:Invent 2016 Fast Forward

AWS Greengrass● Running AWS Lambda functions across the cloud and IoT devices.● Partners: Intel, Qualcomm and Canonical.

Page 76: AWS re:Invent 2016 Fast Forward

Blox

Null

Page 77: AWS re:Invent 2016 Fast Forward

Insights● AWS is really smart to pick what to develop. (Nearly) every service has

potential clients and (MVP) spec confirmed before it is launched. That is, you can say AWS is not a creative company, but more like a customer-oriented service provider.

● AWS prefers to provide fundamental services than complete functional services. Ex, multi-CDN proposal; QuickSight costs 1/10th; CloudWatch vs Datadog.

● Microservice/serverless/Lambda is still quite hot. Lambda function is extended to IoT, Snowball, Alexa, CF Edge with integration of Batch and Step Functions. And rumor has it that Lambda would penetrate into more services.

● Amazon Echo-like hand-free IoT is not here yet because of Chinese.

Page 78: AWS re:Invent 2016 Fast Forward

What’s next?Deeper AWS sharing:

● Content Delivery & Video Streaming.● Serverless Computing.● Databases.● Big Data & Machine Learning.● You name it...AWS MiniCon!

Page 79: AWS re:Invent 2016 Fast Forward

AWS topic sharing

Track session report

WE WANT YOU

Page 80: AWS re:Invent 2016 Fast Forward
Page 81: AWS re:Invent 2016 Fast Forward

https://kkvideo.atlassian.net/wiki/display/ConferenceChronicles/AWS+re%3AInvent+2016

Page 82: AWS re:Invent 2016 Fast Forward

See you next year!