Title: Networked Systems in Developing Regions Challenges and Opportunities
1Networked Systems in Developing
RegionsChallenges and Opportunities
- Lakshminarayanan Subramanian
- Courant Institute of Mathematical Sciences
- New York University
- Joint work with many fantastic collaborators!
- (Courant, NYU Med, NYU Wagner, UC Berkeley,
Intel, Univ of Washington, Univ of Colorado,
Amrita University, Aravind Eye Hospitals)
2Disclaimers of a realist
- I am not a philanthropist
- NGOs are great, but may not scale or be
sustainable - Information and Communication Technologies (ICT)
can only be an aid but cannot alleviate poverty - I believe in cell-phones
3Causes of poverty
- Lack of awareness/opportunity
- A negative earnings cycle
- Serious healthcare problems
- Prolonged debt
- Inheritance from family/society
4Strategies to eliminate poverty
- Jeffrey Sachs Provide aid to the
underprivileged - William Easterly 2.3 trillion dollars of aid,
nothing much to show. Promote homegrown
solutions - C.K. Prahlad Tap the fortune at the bottom of
the pyramid - Amartya Sen Promote development at the rural
level
5The Bottom A Brief Description
- 3-4 billion people with per-capita equivalent
purchasing power (PPP) less that US2,000 per
year - Could swell to 6-8 billion over the next 25 years
- Most live in rural villages or urban slums and
shanty townsmovement towards urbanization - Education levels are low or no-existent
(especially for women) - Markets are hard to reach, disorganized, and very
local in nature
6The cost of being Poor
Bombay area Dharavi(shantytown) Warden Road Ratio
Credit (APR) 600-1000 12-18 60-75x
Water (100 gal) 0.43 0.011 37x
Phone (cents/min) 4-5 2.5 2x
Diarrhea Meds 20 2 10x
Rice (/kg) 0.28 0.24 1.2x
7Even the Very Poor Spend
- Dharavi, one of the poorest villages in India
- 85 have a TV
- 50 have a pressure cooker
- 21 have a telephone
- but cant afford a house
- Even the poorest of the poor in Bangladesh
- devote 7 percent of income to communications
services (GrameenPhone) - These are valid markets
8ICT A Big Missing Piece
- Technology can impact everyone
- Bottom of the Pyramid
- Not just Internet access
- Health, education, government, commerce
- Enable profitable businesses
- Must be scalable and sustainable
- Poor are a viable market
- Focus on income creation, supply chain efficiency
- Not charity, not financial aid
- Promotes stability, entrepreneurism and social
mobility - First World technology is a bad fit
- New research agenda
9How can ICT help?
- Communications
- Awareness, access to external world, phone calls
- Healthcare
- Where there is No Doctor? Rural healthcare
system - Telemedicine/consultation
- Continuing Medical Education for Health-workers
- Low-cost diagnostic tools
- Finance
- Microfinance audit, insurance schemes
- Education
- Educational modules, distance learning
10Other important areas
- Commerce
- Supply Chain
- Agriculture
- E-Governance
11Where has ICT helped so far?
- Very few ICT based efforts worldwide have led to
a large-scale success - ICT adoption successes
- Cell-phone, Radio, TV
- Why?
- Figure out what they need
- Never thrust a technology
- Economics not in favor
- Make change a gradual phenomenon!
12A five-step approach
- 1. Identify basic real-world problem
- 2. Where can ICT help?
- 3. Research challenges?
- 4. What is the appropriate solution?
- 5. Deployment?
13ICT research challenges
14Network connectivity is key!
- Traditional wire-line connectivity solutions are
not economically viable! - Potential options
- Develop new low-cost connectivity solution!
- Leverage existing low-bandwidth wireless
solutions - Cellular, Satellite, CDMA450, WiMax
- Intermittent links are a fact of life
- Budget constrained links
- SMS
- Power outages
- Physical transportation links
15Research Challenges
- Low-cost high-performance network connectivity
- WiLDNet WiFi-based Long Distance Networks
- Wireless Mesh Networks
- Intermittent distributed systems
- Leveraging the next-generation cell-phone
- Redesigning applications to tolerate
intermittency - Interactive applications in low bandwidth
environments - What can you do using 30 Kbps bandwidth?
- Application specific research challenges
- Security, User-interface, NLP, Image processing
etc.
16Research Challenges
- Low-cost high-performance network connectivity
- WiLDNet WiFi-based Long Distance Networks
- Wireless Mesh Networks
- Intermittent distributed systems
- Leveraging the next-generation cell-phone
- Redesigning applications to tolerate
intermittency - Interactive applications in low bandwidth
environments - What can you do using 30 Kbps bandwidth?
- Application specific research challenges
- Security, User-interface, NLP, Image processing
etc.
17Rural Connectivity in Developing Regions
- Current solutions
- Wired
- Dial-up/DSL
- Optical fiber
- Cellular
- VSAT
- WiMax
18WiFi-based Long Distance Networks
- WiLD links use standard 802.11 radios
- Longer range up to 150km
- Directional antennas (24dBi)
- Line of Sight (LOS)
- Why choose WiFi
- Low cost of 500/node
- Volume manufacturing
- No spectrum costs
- Customizable using open-source drivers
- Good datarates
- 11Mbps (11b), 54Mbps (11g)
19WiLD Deployments
- Bay Area
- 7 links up to 45km
- Experimental testbed
- India at Aravind Hosp
- 12 links up to 15km
- 9 Rural Clinics
- 25,000 patients/yr
- Scaling to 50 clinics
- Ghana
- 5 links up to 100km
- Electronic Library
Aravind Network, Theni, India
Other Projects
- Digital Gangetic Plains
- 30 links
- upto 39 km
- 802.11 APs
- Akshaya Wireless
- 400 nodes
- 2-25 km
- commercial APs
- CRCNet,New Zealand
- 17 links
- upto 13 km
- Soekris SBCs
20Akshaya A Case study
- Joint project between the State of Kerala in
India and Tulip IT - Wireless IP network set up in the district of
Malappuram 630 eCenters in all - Backhaul and last mile links are based on
proprietary technologies - 1 center for every 2000 families
21Total Costs for network deployment
WiFi/WiMax is the most economically viable
solution Fiber/WiMax is the least economically
viable
22Experience with WiLD Networks
- In the field, point-to-point performance is bad
- On a 60km link in Ghana
- We get 0.6 Mbps TCP vs 6 Mbps UDP
- On a relay (single channel)
- We get only 2 Mbps TCP
23WiLDNet Design Overview
- Fix 802.11 protocol problems
- Replace CSMA -gt TDMA
- Enforce synchronization of multiple links
- Variable channel loss
- Adaptive loss recovery
- Combine retransmissions and FEC
24Design Constraints
- No hardware changes
- Modify WiLD routers, not endpoints
- Routers are inexpensive machines
- low processing power
- low energy budget (solar)
- We want to be spectrum efficient
25Problem with 802.11 ACKs
- Low utilization
- Large propagation delays
- Stop wait inefficient
- RTS/CTS makes it worse
- ACK timeouts
- ACK doesnt arrive in time
- Retransmissions until retry limit reached
26Problem Propagation Delay
- Large propagation delay ? high collision
probability
27Design Choices for WiLDNet
- Use Sliding Window flow control
- 802.11 MAC ACKs disabled
- Packet batches sent every slot
- Slot allocation determined by demand
- Replace CSMA with TDMA on every link
- Alternate send and receive slots
28Inter-Link Interference
Simultaneous Receive
Simultaneous Send
Send Receive
29Implicit Synchronization for TDMA
- Every packet is time-stamped in TX slot
- Slots are offset because of propagation delay
- We dont use explicit marker packets to signify
end of TX slot
?
Sender
TX slot
RX slot
Receiver
2P MAC protocol (Raman et al. Mobicom 05)
30Channel Loss From external traffic
- Strong correlation between loss and external
traffic - Source (A) and interferer (I) do not hear each
other
31Loss Recovery Bulk ACKs FEC
- Bulk ACKs
- Aggregate ACKs (bit-vectors) sent with every
packet - Use retransmissions for loss recovery
- Retry limit can be per-packet
- Adaptive FEC
- Sender performs encoding of packets proactively
- Packet level FEC
- Tradeoff of BW and Delay
- Bandwidth efficient use Bulk ACKs
- TCP, bulk traffic
- Delay efficient use Adaptive FEC
- Voice, Video
32WiLDNet Design Recap
- Replace CSMA with TDMA
- Loose time synchronization
- To eliminate inter-link interference
- Overcome variable channel loss
- Adaptive loss recovery layer at link layer
- Using Bulk ACKs BW efficient
- Using FEC Delay efficient
33Evaluation Multiple Hops outdoors
- 2 hop network
- 19km, 1.5km
- WiLDNet Similar throughput
- Same channel OR
- Diff. channels
- More spectrum efficient
Channels Through-put for TCP (Mbps)
802.11 Same 2.11
802.11 Diff. 4.50
WiLDNet Same 4.86
WiLDNet Diff. 4.90
34Deployment
35Wireless Mesh Networks
- Rapidly deployable high performance wireless mesh
networks - Current mesh networks have poor performance in
multi-hop settings - What we are investigating
- Multi-radio, Multi-channel
- Nodes with directional antennas
- Understanding multi-AP interactions
- Visit the 12-node testbed in 715 Broadway!
36Research Challenges
- Low-cost high-performance network connectivity
- WiLDNet WiFi-based Long Distance Networks
- Wireless Mesh Networks
- Intermittent distributed systems
- Leveraging the next-generation cell-phone
- Redesigning applications to tolerate
intermittency - Interactive applications in low bandwidth
environments - What can you do using 30 Kbps bandwidth?
- Application specific research challenges
- Security, User-interface, NLP, Image processing
etc.
37Intermittent Distributed Systems
- How do we build conventional distributed systems
in intermittent environments? - Connectivity is intermittent and unpredictable
- Net bandwidth is limited
- Potentially high delays
- Potentially lossy environments
- Examples
- A distributed system of cell-phones using GPRS
links and SMS messages to communicate - Web search from a rural cafe
38Intermittent Web Search
- A typical search today involves 4-8 queries!
- Can we do web search in one round?
- What we have done
- Change the query interface
- Specify all that you know about what you are
searching for - Intermittent proxy issues multiple queries,
prefetches and bundles response pages - Local proxy enables search within retrieved bundle
39Intermittent ATM
- How do you operate ATMs over intermittent
environments? - Our solution
- Enable offline authentication
- Use approximate consistency results to split a
bank balance into smaller entities - Provide a risk model to enable extra cash
availability at ATMs
40Research Challenges
- Low-cost high-performance network connectivity
- WiLDNet WiFi-based Long Distance Networks
- Wireless Mesh Networks
- Intermittent distributed systems
- Leveraging the next-generation cell-phone
- Redesigning applications to tolerate
intermittency - Interactive applications in low bandwidth
environments - What can you do using 30 Kbps bandwidth?
- Application specific research challenges
- Security, User-interface, NLP, Image processing
etc.
41Low-bandwidth video streaming
- Multi-hop satellite network for distance learning
/telemedicine - Very high delays, low bandwidth
- Questions
- Enhancing QoS on low bandwidth environments
- OverQoS Overlay based QoS
- On the fly transcoding
- Can we develop an appropriate video codec that is
easily adaptable to different rates?
42Intermittent proxy
- Imagine a distributed system over a multi-hop
intermittent/low-bandwidth network - Content distribution, distributed databases,
video streaming, client/server transactions, web
search - Intermittent proxy
- A generic optimization engine that performs
resource allocation across flows on a
low-bandwidth/intermittent link - Maintains soft application state to enhance
performance - Performs local hop-hop recovery
43Research Challenges
- Low-cost high-performance network connectivity
- WiLDNet WiFi-based Long Distance Networks
- Wireless Mesh Networks
- Intermittent distributed systems
- Leveraging the next-generation cell-phone
- Redesigning applications to tolerate
intermittency - Interactive applications in low bandwidth
environments - What can you do using 30 Kbps bandwidth?
- Application specific research challenges
- Security, User-interface, NLP, Image processing
etc.
44Cell-phone based applications
- Cell-phone based Micro-finance
- Use SIM as a cheap identity
- Use programmable smart-phones to provide low-cost
authentication - Benefits Reduce transaction costs and corruption
- Cell-phone based Health record system
- Health-workers in the field use cell-phones to
enter health records - Need a distributed database syncronization/search
mechanism which works over SMS-links - Cell-phone based cheap Inventory management
- Why need RF-ID based systems?
45Low cost paper-watermarking
- Every piece of paper has an inbuilt random
speckle pattern - Can we extract this speckle pattern to watermark
any paper? - Use USB 60x microscope
- Applications
- Verification of paper based records
- Cheap inventory management
46Watermarking initial results
Nearly 90 match
Less than 20 match
47Automated Diabetic Retinopathy
48Medical Education Modules
- Can we create medical education modules to teach
a healthcare worker to become a malaria expert? - WiSE-MD modules for surgical education in NYU Med
- Tailor modules for two focus areas
- Malaria and Infectious diseases
- Obstetrics and Gynecology
49Other interesting problem areas
- Traffic Signaling Networks
- User interfaces
- Text-free interfaces for semi-literate or
illterates - Speech interfaces
- Natural language processing
- People interested in content in local language
- Language translation
50Questions?