Challenged Gateway for WSN

1
Challenged Gateway for WSN Focus on limited
infrastructure deployments
http//www.wsnblog.com
C. Noda 1, C. Pérez 1, M. Zennaro 2 and B.
Pehrson 3 "Henri Poincaré Group of Complex
Systems (1) Physics Faculty, University of
Havana http//www.complexperiments.net Aeronomy
and RadioPropagation Laboratory (2) ICTP, Trieste
http//wireless.ictp.trieste.it Royal Institute
of Technology KTH, Sweden (3)
Workshop on Wireless Computing and Sensor
Networks Havana. May, 2007
2
Challenged Gateway for WSN Focus on limited
infrastructure deployments

• WSN meets ICT4D main technological requirements
• Autonomous connectivity, low-cost equipment,
  appropriate user
• interfaces and power resilience
• We believe WSN technologies has much to DC
• A barrier many researchers face when attempting
  to deploy
• a WSN
• lack of easy-to-use, afordable, flexible tools
  for the study of
• a scientific phenomena in real, non-laboratory
  settings.
• Our goal is to build an experimental platform to
  help non-technical researchers deploy and use a
  wireless sensor network in Developing Countries

http//www.wsn4d.org
3
Challenged Gateway for WSN Focus on limited
infrastructure deployments
WSN Applications CitySense ZebraNet Duck
Island Military Survileance Shooter
Localization Volcano Monitoring
http//www.wsnblog.com
4
Challenged Gateway for WSN Focus on limited
infrastructure deployments
Two main Assumptions in WSN research
today There are POWER and CONNECTIVITY
Gateways Xbow Stargate - XScale Network
Interface and Single Board Computer 400MHz,
Linux Single Board Computer Archrock But
commonly the gateway is A Notebook 800.00
USD
http//www.wsn4d.org
5
Challenged Gateway for WSN Focus on limited
infrastructure deployments
Current Research Work
Adapted from J. Beutel, Swiss Federal Institute
of Technology Zurich
6
Challenged Gateway for WSN Focus on limited
infrastructure deployments

• In order to advance in rural WSN applications
• A challenged Gateway is in need
• Very low power and limited connectivity both in
  time and bandwith
• Thus our Gateway Desing should focus on
• Price
• I/O
• Speed
• Memory
• Java performance
• Power consumption

Shall we use SNAP? Or maybe we should consider
other platforms? Maybe an embedded linux?
http//www.wsn4d.org
7
Challenged Gateway for WSN Focus on limited
infrastructure deployments
The Linksys NSLU Pros 133 MHz,
ARM-processor 32MB RAM 8MB flash 10/100 Ethernet
port two USB 2.0 ports lt 1 Watt power circa 150
USD and there is a hack http//www.nslu2-linux.o
rg/ Cons proprietary design scarce
documentation
Tmote Connect is based on it http//www.moteiv.co
m
http//www.wsn4d.org
8
Challenged Gateway for WSN Focus on limited
infrastructure deployments
http//www.wsn4d.org
Tmote Sky Moteiv.com 60 Eu
JSimm
SNAP Imsys.se 122 Eu
TILT Systronix
9
Challenged Gateway for WSN Focus on limited
infrastructure deployments
We need something robust enough to be
deployed in the Jungle In a River in
Desert Low Cost so we can do DC apps Yet
flexible and indeed simple to use
The KISS principle
(Keep It Simple, Stupid)
If S stands for simplicity S x y z x is
good research and prototyping y is play z is
reduction of unnecessary elements
10
Challenged Gateway for WSN Focus on limited
infrastructure deployments
http//www.wsn4d.org
11
Challenged Gateway for WSN Focus on limited
infrastructure deployments
12
Challenged Gateway for WSN Focus on limited
infrastructure deployments

• power aware
• sleeping extends
• battery life
• store and forward

13
Challenged Gateway for WSN Focus on limited
infrastructure deployments
Goals Come up with a sound and robust hardware
design based on an open developement platform
backed by a supporting community And we
contribute with open source software... Soo
cool! Right?
To be considered Delay Tolerant Netwoking (DTN)
14
Challenged Gateway for WSN Focus on limited
infrastructure deployments
Final Design SNAP 512 MB (Flash
Storage) WRT54GL IP55 Enclosure 12v 7Ah
Battery
15
Challenged Gateway for WSN Focus on limited
infrastructure deployments

• Maybe we are WRONG -/
• Sofware Development nightmare!!
• porting TinyOS java tools to J2ME
• code nesC and Java for each app
• shall we be able to program nodes in the field
  remotely?

Are we missing something ?? middleware for
sensor networks... what are other people doing?
Do we want an ad hoc or a general solution? Well
we have company.. SeaMoster The South East
Alaska MOnitoring Network for Science,
Telecommunications, Education, and Research
Involves NOAA, NASA and Microsoft
Lets bactrack a little bit...
16
Challenged Gateway for WSN Focus on limited
infrastructure deployments
Tenet http//tenet.usc.edu/ O. Gnawali,
B. Greenstei, K. Jang, A. Joki, et al, SENSYS,
2006. Boulder, CO. The Tenet architecture is
motivated by the observation that future
largescale sensor network deployments will be
tiered, consisting of motes in the lower tier
and masters, relatively unconstrained 32-bit
platform nodes, in the upper tier.

• Masters provide increased network capacity
• constrains multinode fusion to the master tier
• motes process locally-generated sensor data
• simplies application development
• allows mote-tier software to be reused
• currently implemented for ARM and Fedora core

17
Challenged Gateway for WSN Focus on limited
infrastructure deployments
And then we discovered the Gumstix!
Design Flexibility, Low Production Costs, Rapid
Time to Market 128 MB RAM, 32MB Flash 400MHz CPU
with 100MHz bus 100 CPU usage requires
0.22A_at_3V And it runs linux.... And there is
super duper wiki with docs
...now we can explore DTN !!
18
Challenged Gateway for WSN Focus on limited
infrastructure deployments

• And now comes EasyWay...
• Aimed at decomplicating fast deployments for
  Scientific Applications
• Based on Django a high level Python Web
  framework
• Been coded in 3 blocks
• Motes reprogramming and network configuration
• Tasklets settings, data collection and storage
• Data visualization
• MySQL
• Matplotlib
• GraphViz

Submitted to NSDR 2007 with ACM SIGCOMM
2007 Kyoto Japan, 27th - 31st August, 2007
19
Challenged Gateway for WSN Focus on limited
infrastructure deployments
Supporters (been currently lobbied) Regione
Friuli Venezia Giulia http//www.regione.fvg.it/
(Future apps in Malawi) IRDC, Canada Adapted
Consulting, Inc http//adaptedconsulting.com
SIDA, Sweden (through www.sics.se)
20
Challenged Gateway for WSN Focus on limited
infrastructure deployments
THANK YOU
claro.noda (at) gmail.com Feedback is Welcome!