Feedback on AWS re:invent 2016

46

Transcript of Feedback on AWS re:invent 2016

Page 1: Feedback on AWS re:invent 2016
Page 2: Feedback on AWS re:invent 2016

Reinvent  2016  •  Monday:  gameday  /  hackaton  •  Tuesday:  sessions  /  Keynote  by  Hamilton  •  Wednesday:  Keynote  by  Jassy,  sessions  •  Thurdsay:  Keynote  by  Vogels,  sessions  •  Friday:  sessions  

•  30000  persons  •  400+  sessions  •  4  places  •  5  days  (instead  of  4)  

Page 3: Feedback on AWS re:invent 2016
Page 4: Feedback on AWS re:invent 2016

Amazon  Global  Network  

Page 5: Feedback on AWS re:invent 2016

Inside  an  AZ  

Page 6: Feedback on AWS re:invent 2016

Custom  routers  

Page 7: Feedback on AWS re:invent 2016
Page 8: Feedback on AWS re:invent 2016

Announcements  

Page 9: Feedback on AWS re:invent 2016

COMPUTE  

Page 10: Feedback on AWS re:invent 2016

Lightsail  

us-­‐east  for  the  moment  Shadow  VPC  (peering  possible)  

Page 11: Feedback on AWS re:invent 2016

New  instances  

C5:  Skylake,  up  to  72  vCPU  /  144GB  RAM  

Page 12: Feedback on AWS re:invent 2016

ElasZc  GPU  

Page 13: Feedback on AWS re:invent 2016

IPv6  support  for  ec2  

•  Instances  can  have  publicly  routable  ipv6  address  •  Dual  stack  instances  (ipv4  +  ipv6)  •  To  create  private  subnets  

Ø New  resource:  Egress-­‐only  Internet  Gateway  

Page 14: Feedback on AWS re:invent 2016

MANAGEMENT  TOOLS  

Page 15: Feedback on AWS re:invent 2016

Opsworks  for  chef  automate  

Managed  chef  server  Chef  automate  suite  

Page 16: Feedback on AWS re:invent 2016

Personal  Health  dashboard  

Personalized  view  of  status  Available  via  API  Can  send  noGficaGons  

Page 17: Feedback on AWS re:invent 2016

EC2  system  manager  

Extension  of  Run  Command  and  Simple  System  Management  Integrated  with  AWS  config  

Page 18: Feedback on AWS re:invent 2016

ANALYTICS  

Page 19: Feedback on AWS re:invent 2016

Athena  

Process  S3  data  with  SQL  Uses  Presto  5$  /  TB  processed  

Page 20: Feedback on AWS re:invent 2016

Glue  

Create  data  sources  Generate  transformaGons  (python)  Run:  scheduled  or  event-­‐based    Uses  Spark  under  the  hood  

Page 21: Feedback on AWS re:invent 2016

Batch  

HPC  as  as  service    Compute  env  (“grids”)        On-­‐demand  or  spot  Queues  to  submit  jobs  

Page 22: Feedback on AWS re:invent 2016

DATABASE  

Page 23: Feedback on AWS re:invent 2016

PostgreSQL  compaZble  Aurora  

Page 24: Feedback on AWS re:invent 2016

DEVELOPER  TOOL  

Page 25: Feedback on AWS re:invent 2016

Code  Build  

Fully  managed  CI    Priced  per  build-­‐minutes  Integrated  with  CodePipeline  

Page 26: Feedback on AWS re:invent 2016

ARTIFICIAL  INTELLIGENCE  

Page 27: Feedback on AWS re:invent 2016

RekogniZon  

Page 28: Feedback on AWS re:invent 2016

Polly  

24  languages  47  voices  

Page 29: Feedback on AWS re:invent 2016

Lex  

Natural  language  understanding    Based  on  tech  powering  Alexa  

Page 30: Feedback on AWS re:invent 2016

MONITORING  AND  SECURITY  

Page 31: Feedback on AWS re:invent 2016

Shield  

Page 32: Feedback on AWS re:invent 2016

X-­‐Ray  

APM  soluGon    Analyze  call  graph  View  call  latencies  

Page 33: Feedback on AWS re:invent 2016

MIGRATION  AND  HYBRID  

Page 34: Feedback on AWS re:invent 2016

VMWare  on  AWS  

Pat  Gelsinger  on  stage    with  Jassy  

Page 35: Feedback on AWS re:invent 2016

AWS  Snowmobile  

100  PB    350KW  1Tb/s  (Nx40Gb/s)  

“We’re  gonna  need  a  bigger  box”  

Page 36: Feedback on AWS re:invent 2016

MOBILE  

Page 37: Feedback on AWS re:invent 2016

Pinpoint  

Send  targeted  noGficaGons    Real  Gme  analyGcs:  engagement,  moneGzaGon,  demographics  

Page 38: Feedback on AWS re:invent 2016

CONTAINERS  

Page 39: Feedback on AWS re:invent 2016

Blox  

Opensource    

Page 40: Feedback on AWS re:invent 2016

LAMBDA  

Page 41: Feedback on AWS re:invent 2016

Greengrass  

Run  lambda  code  on  IOT  devices  

Page 42: Feedback on AWS re:invent 2016

Snowball  edge  

Bigger  snowball:  100TB  Horizontal  scaling  via  clustering  In  addiGon  to  s3  import/export,  local  storage  +  compute  

Ø  Process  data  using  lambda  

Page 43: Feedback on AWS re:invent 2016

Lambda  edge  

Run  lambda  code  on  Cloudfront  edges  Limited  capabiliGes  

Ø  50ms  maximum  Ø No  access  to  external  services  

Use  cases:  redirect  /  modify  process  headers  and  cookies  

Page 44: Feedback on AWS re:invent 2016

Lambda  c#  support  

Page 45: Feedback on AWS re:invent 2016

Lambda  Step  funcZons  

Coordinate  lambda  funcGons  Ø  Sequences  Ø  Parallel  processing  Ø  Branch  based  on  data  Ø  Synchronize  Ø Manage  errors  and  retries  

Page 46: Feedback on AWS re:invent 2016

Conclusion  •  More  compeZZon  in  the  cloud  

Ø  AWS  is  developing  some  services  others  had  before  them  •  AWS  now  talks  “Hybrid”  which  is  very  new  •  Not  many  news  in  mobile  /  containers  /  database  •  More  and  more  high-­‐level  services  •  AWS  goes  on  the  market  of  soeware  vendors  

Ø  APM,  Grid,  CI,  ETL  

•  Big  push  in  AI  •  Many  things  around  Lambda  

The  amount  of  new  services  this  year  is  crazy