!hybrid!cloud,based!sensor!system!for!simplifying! and ...ckrintz/racelab/smartfarm.pdf ·...

5
SmartFarm Hybrid Cloudbased Sensor System for Simplifying and Automa8ng Agriculture Analy8cs Chandra Krintz ([email protected]) Rich Wolski ([email protected]) hDp://www.cs.ucsb.edu/~racelab Computer Science Department UC Santa Barbara Nov 2014

Upload: others

Post on 13-Aug-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: !Hybrid!Cloud,based!Sensor!System!for!Simplifying! and ...ckrintz/racelab/smartfarm.pdf · SmartFarm!!Hybrid!Cloud,based!Sensor!System!for!Simplifying! and!Automang!Agriculture!Analy8cs!

SmartFarm    Hybrid  Cloud-­‐based  Sensor  System  for  Simplifying  

and  Automa8ng  Agriculture  Analy8cs  

Chandra  Krintz  ([email protected])  Rich  Wolski  ([email protected])  

hDp://www.cs.ucsb.edu/~racelab  Computer  Science  Department  

UC  Santa  Barbara  Nov    2014  

 

Page 2: !Hybrid!Cloud,based!Sensor!System!for!Simplifying! and ...ckrintz/racelab/smartfarm.pdf · SmartFarm!!Hybrid!Cloud,based!Sensor!System!for!Simplifying! and!Automang!Agriculture!Analy8cs!

Mo8va8on  

•  Ag  Industry  -­‐-­‐  under  tremendous  pressure  to  increase  food  produc8on  and  yields  to  keep  up  with  demand  –  Popula8on  growth  is  outpacing  produc8on  –  Natural  resources  are  limited  (arable  land,  water,  …)  –  Hard  to  predict  weather  paDerns,  climate  change  

•  Goal  of  our  work:  inves8gate  a  novel,  unifying,  and  open  source  approach  to  agriculture  analy8cs  and  precision  farming  called  SmartFarm    –  Open  source,  secure,  on-­‐premise  (private)  cloud  

•  Easy  to  use,  self  managing,  lights-­‐out  automa0on  (think  Tivo!)  –  Linking  disparate  sensor  technologies  and  analysis  engines  –  To  give  farmers  ac8onable  insights,  alerts,  accurate  predic8on  

Page 3: !Hybrid!Cloud,based!Sensor!System!for!Simplifying! and ...ckrintz/racelab/smartfarm.pdf · SmartFarm!!Hybrid!Cloud,based!Sensor!System!for!Simplifying! and!Automang!Agriculture!Analy8cs!

The  UCSB  SmartFarm  

farm  office  

security  boundary   ac8onable  

analy8cs,  alerts  

D  A  T  A  

External  data    sources  (ingress      only)  

UAVs  

Private  Cloud  

Aerial  imagery  

Internet  data  

climate  data  

servers  

many  sources  of      farm  sensor  data    

APIs  

Public  Clouds  

Page 4: !Hybrid!Cloud,based!Sensor!System!for!Simplifying! and ...ckrintz/racelab/smartfarm.pdf · SmartFarm!!Hybrid!Cloud,based!Sensor!System!for!Simplifying! and!Automang!Agriculture!Analy8cs!

UCSB  SmartFarm  Technologies  

•  Private  cloud  is  API  compa8ble  with  Amazon  and  Google  Clouds  –  Any  applica8on  or  service  that  executes  in  public  cloud  can  also  execute  in  on-­‐farm  without  modifica8on  

–  Facilitated  via  AppScale  and  Eucalyptus  private  cloud  systems  –  Hybrid  configura8ons  

•  Move  computa8on  to  the  data  on-­‐farm;  control  all  data  that  leaves  farm  –  Precision  farming  “app  store”  

•  Or  perform  computa8on  in  the  public  cloud;  shared  datasets  and  services  

•  Leverages  emerging  Internet-­‐of-­‐things  (IOT)  &  big  data  technologies  –  Unified  data  ingress  from  disparate  sensor  systems  

•  Images  and  data  –  Machine  learning  and  data  mining  –  Streaming  (think  Storm)  and  batch  processing  (think  Hadoop/Spark)  

Page 5: !Hybrid!Cloud,based!Sensor!System!for!Simplifying! and ...ckrintz/racelab/smartfarm.pdf · SmartFarm!!Hybrid!Cloud,based!Sensor!System!for!Simplifying! and!Automang!Agriculture!Analy8cs!

Proposed  Research  Goals  

A  novel,  unified  so8ware  system  that  provides  

•  A  cloud-­‐based  soeware  architecture  that  links  data  sources  with  data  aggrega8on,  processing,  and  analy8cs  algorithms,  

•  Data  ingress  for  disparate  sensor  systems  (local  and  remote),  each  poten8ally  with  their  own  data  formats,  8me  scales,  etc.,    

•  A  new  data  model  based  on  document  referen8ality,  geoloca8on,  and  eventual  consistency,    

•  Applica8on  programming  interfaces  (APIs)  for  easy  and  automa8c  integra8on  of  modeling,  analysis,  aler8ng,  audi8ng/repor8ng,  and  visualiza8on  technologies,  and    

•  Privacy  and  security  mechanisms  that  restrict  access  to  data  and  the  system  to  only  authorized  par8es.