Planning around Technology Prof. Eric A. Brewer UC Berkeley ICT for Developing Regions September 17, 2003

Download Report

Transcript Planning around Technology Prof. Eric A. Brewer UC Berkeley ICT for Developing Regions September 17, 2003

Planning around Technology
Prof. Eric A. Brewer
UC Berkeley
ICT for Developing Regions
September 17, 2003
Today’s Focus
 How to think about technology…
 Not
very technical
 Goal: map applications into technology options
 Architecture
 Networking
 Devices and Sensors
General Architecture
Data
Center
Internet
Data
Center Data
Center
Proxies,
Basestations
cell
Devices or sensors
“disconnected”
Data Centers

Best place to store persistent data



(device is second best)
Can justify backup power, networking, physical security
Cheapest source of storage/computer per user


100-1000x less than a personal device (!)
Factors: shared resources, admin cost, raw costs (power,
disks, CPUs)
Need at least two for disaster tolerance
 Plan about 3 in urban centers (for one region)
 Berkeley will be the data center for our early work…

Networking
The key enabler… distribution for information.
 Focus on wireless




Packets allow efficient use of media


Vastly cheaper to deploy
Wide range of options
“multiplexing”
IP = “Internet Protocol”


Global names for every device
Way to route packets to names
Basics

Bandwidth = data/sec



Latency = transmission delay in seconds



Modem = 56 kilobits/sec
DSL = 384 kilobits/sec
Ethernet = milliseconds
Satellite = ¼ second
Power should be mostly tied to transmission


Double distance => quadruple power
Sequence of short hops usually lower power
(but higher latency)
Communication Attributes
 Directional or Omnidirectional
 Broadcast or Unicast
 Wireless
is always broadcast
 Shared or Single User
 Wireless
is always shared
 Synchronous or Asynchronous
 Cellular
is synchronous: full-time two-way connection
 E-mail is asynchronous: send now, receive later
Understanding Range
Range costs power (squared)
 Long distances best covered with directional antennas





User density matters:




10x difference in range for low cost
“Point-to-point” links
Can go 50km at reasonable cost
Range also limited by total users
Urban areas thus use short-range wireless
Rural areas need long-range
Claim: need islands of coverage (with point-to-point)

Islands are the dense areas (e.g. villages)
Example: India
Mumbai (Bombay)
Chennai (Madras)
Mumbai
Broadcast vs. Unicast

Broadcast: same data to everyone


Unicast: data to one person


Web page, e-mail
Wireless broadcast is *cheap*



TV, radio, newspaper
Easy to reach many users with one transmission
Example: satellite TV
Claim: we will need to exploit broadcast



Broadcast common data even for unicast applications
Example: common web content
Goal: minimize unicast data
Synchronous vs. Asynchronous

Synchronous: two-way communications




Low latency, e.g. Phone = ¼ sec or less
Expensive: dedicated resources along whole path
Examples: phones, games, web browsing
Async One-Way: e.g. TV, Internet movie

Delay allowed: e.g. “buffering” for Real Player




Hides losses: resend data before it is missed
Typical delay is 10 seconds
Good fit for broadcast
Examples: Streaming, Market Prices, Weather alerts
Sync vs. Async (2)

Async Two-Way: e.g. e-mail request

Delay in getting response (= two async one-way)


Savings:





Semi-interactive, but potentially much cheaper…
No need for dedicated resources
Can “store-and-forward” data (like real mail)
Can hide problems (e.g. power out) by waiting
Examples: e-mail, correspondence classes, medical
diagnosis (non-emergency), coordinating money transfers, ecommerce (e.g. catalogs)
Open Questions:


How much cheaper? How much more reliable?
When is async good enough?
Intermittent Networking

Physical:


Low-earth orbit satellites: connect only while they are
overhead
“Mules” – moving basestation collects data




Basestation could be on a bus
Weather, e.g. some places only get radio on clear nights
Overloaded network may delay transmission
Extended coverage:


User may periodically enter the coverage area
E.g. coverage only near market or school
The Case for Intermittent

Pros:





Cost: better use of resources, more tolerant of problems
Reliability: delay hides transient problems
Ease of deployment: can be more ad hoc, less coordination
than a synchronous system
Coverage: Intermittent coverage >> full time coverage
Cons:



Not really interactive, or only interactive in some areas
Need to design apps around this (new) model
Don’t know what delay is OK (depends on the app)
Cellular Phones

Advantages:




User interface works well (no need for literacy)
High demand for voice, but some data provided
High-volume phones and basestations
Disadvantages:




Hard and expensive technically
Voice may be inefficient use of bandwidth
Poor fit for rural users (low density of users)
Typical data rate < 9600 baud
Example: Cellular Standards
Cell Range
Cost
Bandwidth
bits/sec
GSM
10 km
$100K
9600
~800
PHS
100-300 m
$3K
9600
~200
35 km
?
9600
~800
System
MiniGSM
Simult.
Users
Satellites
Pros: great for broadcast, coverage
 Cons: limited shared bandwidth, mostly one-way, high
latency
 GEO (geostationary – stays in one place)




35000 km up, very high latency
Mostly one-way due to power
LEO (low-earth orbit), <1000 km up



Intermittent coverage (only when overhead)
Better latency and power, cheaper
Need lots of satellites (800?)
Two-way using Satellites

Option 1: transmit up to the satellite



Very high power required, and large antenna
Very low bandwidth (9600 or less)
Option 2: use a different network


Telephone is most common, could use wireless
Tricky but possible to do TCP
Variation: upload data later when convenient
(intermittent)
 Good for mostly broadcast applications, coverage

WiFi: 802.11
Made for wireless local-area network
 Three variations:


b: 1 Mb/s, oldest and cheapest, $5/chipset



a: 11 Mb/s, shorter range
g: 54 Mb/s, shorter range, more channels


Best for long-distance links
Best for omni coverage of local area (if cheap)
Decent for long distance: (802.11b)



Special directional antennas & alignment, ($20-$200)
30km seems easy, 88km is current record
Longer distances need towers, 40’ typical
WiMax (802.16)

New standard for metro areas

Target is 30km radius, $20K for basestation


High bandwidth, low density




Cost should come down (as with 802.11)
Receiver per village, not per person
100 villages per basestation plausible
Probably requires licensing (not free for all)
Complements 802.11 for local coverage

Expect this combo to be very popular…
Proxies and Basestations
Cellular => one basestation per cell
Proxies are similar: locally shared computing/storage
 10x less than devices (per user!)



Target cost: $200 or about 20¢ per user
Good way to share computing, storage
Relay point for networking


Typically stationary, may be solar powered
“Mules” are mobile proxies
Easy to upgrade in place (like Tivo)
 Not good for long-term storage (“persistent data”)


No backup plan, unreliable power and security…
Example Proxy Tasks

Caching of frequently used data


Temporary storage



Packets waiting for LEO
Transactions in progress
Computation-intensive tasks:




Such data can be broadcast to proxies
Speech recognition
Sensor data analysis
Process data for “dumb” device
Can double as “big” device, e.g. Internet kiosk
Devices
 Wide range of devices are possible
 Two broad classes:
 General-purpose
computer (Simputer, PDA, laptop)
 Task specific: water testing, telephone, camera
 Claim: task specific is a better choice
 Much
simpler to use (and train for)
 Can be more cost effective
 Volume should be high enough to justify
Shared Devices

Sharing reduces the cost per user



Village phones have 4-8x better cost/performance
May only need one per village (water testing)
Some design issues:

Does behavior depend on which user?



E-mail does, telephone does not
Need for authentication, accounting?
Claim: Most projects should start with shared devices

Can move to personal devices if demand warrants…
Device Costs

Big costs:





General driver: too many components



Packaging
Screen: color is >3x monochrome
Battery
UI: keyboard, buttons, touch screen, …
Solution: better integration into silicon
Silicon is ~free… (marginal cost)
We are working on real details of costs…
How to think about devices…
Sharing is good (but design for it)
 Task-specific is good (simpler and cheaper)
 UI and size have a lot to do with cost…
 Dependence on infrastructure is good





Reduces cost (and theft)
Can push work into infrastructure as needed
Increases extensibility, decreases obsolescence
… but requires thought about disconnected operation
Sensors
 Sensors on silicon => very cheap
 Basic:
temp, humidity, pressure, moisture, light…
 Moderate: acceleration, magnetic, position
 Simple chemical: gases, smoke
 Complex biological: toxic agents
 Actuators: control environment as well
 Much
harder: usually want human in the loop…
Sensor Applications

Environmental monitoring



Aid to Infrastructure



Electrical grid loss/theft
Water distribution loss/theft/quality
Commerce


Water testing
Weather warnings?
Butterfat testing for milk pricing
Health


Home environment
Human tests? (or animal)
Example: Great Duck Island
Monitoring petrels on an island in Maine
 About 50 sensors



Some in burrows, detect occupancy
Some to track environmental conditions
Fits architecture:





1.25 inch
Data center in Berkeley
Network: local wireless in patches, satellite to UCB
Proxies: process/manage data collection
Mix of sensors
Intermittent connectivity (to save power)
Summary

Decide on the task





Intermittent networking may be sufficient



Not about internet access (per se), about applications
Drives UI, leads to simplicity, lower cost
Long-term: single chip per app is ideal…
Wide range of devices can share infrastructure
Greater coverage, much lower cost, more reliable
Not always interactive
Share, share, share



Data centers are big shared resources, >100x savings
Proxies are locally shared resources, >10x savings
Share devices as well