Transcript pads_review_arch1 - Power Aware Distributed Systems

Architectural Approaches (Part 1)
Power Aware Distributed Systems Kickoff
August 23, 2000
Sensor Node Baseline
1.
Instrument a state-of-the-art sensor node to understand
power consumption in current systems. Where can we
expect significant latitude in power tradeoff? Which knobs
have the greatest dynamic range? What baseline will we use
for comparison?
2.
Identify hardware knobs that can be provided by modules
(radio and processor systems) that can be altered
dynamically, and externally readable parameters (power,
BER, signal strength, battery, etc.) that can be provided to a
power-aware runtime system.
Rockwell WINS Nodes

Rockwell WINS is a modular
stack consisting of:




Power Board
StrongARM Board
Radio Board
Sensor Board

This architecture is fairly
representative of other sensor
nodes in the community.

We plan to adapt this node to
allow module-level power
instrumentation and logging both
in the lab and in the field.
Module Isolation
Battery Pack
Host
Serial
Port

Insert a power isolation board
between each module in the
WINS stack.

Signals are passed through,
power supplies are isolated.

Microcontroller provides power
monitoring and power control
from a host’s serial port.

Event “snooping” may be
possible to trigger data
acquisition in the field.
StrongARM
PADS Power Isolator
Radio
PADS Power Isolator
Sensor
PADS Power Isolator
Sensor Node Baseline
1.
Instrument a state-of-the-art sensor node to understand
power consumption in current systems. Where can we
expect significant latitude in power tradeoff? Which knobs
have the greatest dynamic range? What baseline will we use
for comparison?
2.
Identify hardware knobs that can be provided by modules
(radio and processor systems) that can be altered
dynamically, and externally readable parameters (power,
BER, signal strength, battery, etc.) that can be provided to a
power-aware runtime system.
Power Aware Research Platform
Testbed

Provide a highly instrumented open
sensor network platform that conforms
to open APIs and COTS hardware
interfaces.
 Must
be inexpensive and easy to replicate.
 Geared for lab and parking lot experiments,
but rugged enough to take to field.
 Leaning towards PC104+ standard.

Too power hungry!
Expensive.
Integrate advanced power aware processing and
communications technology into the PADS research platform
as it becomes available in the PAC/C community.
 Allow
community to test new PAC/C technology in sensor networks
without big leap into proprietary low-power commercial systems.
Node Options

Rockwell WINS – StrongARM Stack
 Advantage
is that Rockwell is a team partner, but some sensitive
components make it not entirely open source and easily replicable.

BWRC Pico Node – StrongARM + Xilinx 4020
 Similar to
WINS. Intended to be open standard, but hardware and
software not as mature.

Sensoria WINS-NG – WinCE PDA + custom box
 With

the exception of PDA, box is closed to developers.
MIT (Anantha Chakandrasan)
 Focus

is on voltage scaling and other in-the-lab power tradeoffs.
Note: All of the above are CPU-centric architectures!
CPU-Centric Architectures

Most nodes are CPU-centric in that the radio, GPS, and
sensors are wired up to serial ports, A/D bus, and/or GPIOs
on an embedded processor such as StrongARM.

A processor must be present in every node.

The CPU must wake up on any event and facilitate
communications between any pair of modules (radio and
sensor modules, for example).

Is this the most power-efficient sensor node system model?

Is there another approach which allows most of the sensor
node to be turned off most of the time?
Distributed Microcontroller Model

Make each module an independent actor on a multi-master
serial bus such as I2C (400Kb, 4Mb*) or CAN (1Mb).

Minimum module processing is a 25mA microcontroller.

Migrate specific processing to the modules (microcontroller
on seismic board, DSP on imager board, etc.).

Create common command set for peer to peer
communication and control of modules.

A StrongARM may be used for application control, but it
could distribute “event handlers” to microcontrollers on the
modules and power down most of the time.
I2C/CAN
Module Emulation
I2C/CAN

Most existing sensor nodes have
an external serial port.

A distributed microcontroller
architecture on a multi-master
serial bus would make it much
easier to integrate modules that
don’t fit the form factor.

Also enables simple module
emulation and module testing
from a workstation or laptop.
Serial
Port
Bridge
Serial
GPS
Unit
Serial
Port
Bridge
Laptop
Distributed Microcontroller Model w/
Local Power Control
Single Chip Camera
Acoustic Sensors
Seismic Sensors
Magnetometers
StrongARM
CPU Module
GPS/Radio
Module
Image System
Module
Sensor Interface
Module(s)
Temperature Sensors
Other Sensors
Battery Pack
DC/DC
Converters
DC/DC
Converters
DC/DC
Converters
DC/DC
Converters
80C554
uController
80C554
uController
80C554
uController
80C554
uController
Power Bus
I2C Interface (400 kb/s)
Distributed Microcontroller Model w/
Central Power
Single Chip Camera
Acoustic Sensors
Seismic Sensors
StrongARM
CPU Module
GPS/Radio
Module
Image System
Module
Sensor Interface
Module(s)
Magnetometers
Temperature Sensors
Other Sensors
PCF8584
I2C Bus
Controller
80C554
uController
80C554
uController
80C554
uController
I2C Interface (400 kb/s)
Module Power Lines
(+5V, +3.3V, +2.5V, +1.5V)
Battery Pack
Microcontroller Power (+5V, +2.5V)
DC/DC Converter
Module
80C554
uController