Transcript Slide 1
Sensor Network Research Group University of Massachusetts, Amherst Ultra-Low Power Data Storage for Sensors
Gaurav Mathur, Peter Desnoyers, Deepak Ganesan and Prashant Shenoy Motivation:
Number of flash memory storage options exist for sensors
NOR flash (Mica, Mica2, Telos)
Multimedia cards – MMCs (RISE)
NAND flash (UMass Mica2 adapter board)
No comprehensive empirical comparison
of the energy-efficiency of
storage alternatives for sensor devices Many applications require energy-efficient, high-capacity storage
Camera sensing
Acoustic sensing What is the most energy-efficient storage platform How does the energy cost of What are the implications storage compare to that of for sensor devices ?
computation and communication of an ultra-low power storage subsystem on sensor net design ?
?
Comparison Study of Storage Technologies Toshiba Parallel NAND flash adapter Mica2 Atmel serial NOR flash MMC adapter TelosB STM serial NOR flash
Compare MMC, NOR & NAND flash technology Measure active & sleep mode power consumption
Comparison of the per byte cost of operations (in μJ)
Device
Mica2 NOR TelosB NOR
Read
0.26* 0.056*
Write
4.3* 0.127*
Erase
2.36
0.185
Total
6.92
0.368
Hitachi MMC 0.144
0.659
0.0033
0.806
Parallel NAND 0.004* 0.009* 0.004
0.017
* Without ECC.
Cost of performing ECC in software is approx 0.026μJ/byte Parallel NAND flash
is
21 times
Telos STM flash and
407 times
more efficient than Mica2 Atmel flash.
Implications of Low Power Storage
Challenges conventional wisdom of trade-offs
Computation
<
storage
<<
communication
Energy-efficiency of NAND flash enables ultra low power,
almost infinite storage
(~1 GB) High-capacity, Low-power storage impacts the following sensor application domains by supporting:
Higher degree of
local data archival and indexing
, which
supports
in-network query processing
Use of history for efficient
network-level compression Impact on Radio Transmission Costs Comparison of computation, communication and storage costs
Energy Ratio
TI MSP430 instruction
0.0008
1
Toshiba NAND read
0.004
5
Toshiba NAND write
0.009
11
CC2420 Radio Tx
2.6
3250
CC2420 Radio Rx
2.2
2750
Storing data on
NAND flash
is
369 times
more energy efficient than communication
BMAC uses a per-packet preamble that leads to a high per-packet transmission cost Instead, store packets on NAND flash & batch transmit Can lead to a
58x reduction
in communication costs !