cs 525m – mobile and ubiquitous computing...
TRANSCRIPT
CS 525M – Mobile and Ubiquitous Computing Seminar
Mark FiguraMike Scaviola
Introduction
• Real-world measurement of 802.11b wireless performance– Signal/noise ratio– Average bandwidth
• Divided into 4 phases– Phase 0 – distance– Phase 1 – building materials– Phase 2 – environmental effects– Phase 3 – interference
How we tested (equipment)
• Proxim ORiNOCO 802.11b pcmcia card• Linksys “Instant Wireless” Wireless Access
Point Router• Two laptops from the ATC
And the power?
• A car starter battery along with a power inverter does wonders!
Software
• 2 measurements taken…– Signal/noise with NetStumbler– Average bandwidth with QCheck
Why?
• NY Stock Exchange• Short of “ray-tracing” how can you tell what
factors are involved in 802.11b performance?– Are there any simple rules for planning
the design of your home wireless network?
Phase 0 - distance
•First test on the football field•As expected, SNR drops as distance increases•At about 15 yards, throughput started to become erratic. By 30 yards, we could not get an accurate measurement.
SNR
0
10
20
30
40
50
60
Distan
ce (ya
rds)
2 3 4 6 7 9 11 13 15 18 20 21 23 24 25 28 30
Throughput
00.5
11.5
22.5
33.5
44.5
5
Dis
tanc
e (y
ards
) 2 3 4 6 7 9 11 13 15 18 20 21 23 24 25 28 30
Phase 0 – distance
Our setup on the football field
Phase 0 – distance
• Second test in a long hallway in the Campus Center
• Throughput is erratic, but confined to two distinct plateaus
throughput
00.5
11.5
22.5
33.5
4
dist
ance 3 6 9 12 15 18 21 24 27 30 33 36 39
throughput
Phase 0 – distance
• Just like in the football-field experiment, the SNR decreases as one would expect with distance…
• At the maximum distance tested which was longer than the distance on the football field, we had 2x the SNR
SNR
010203040506070
Tim
e (G
MT
)
02:3
2:01
(G
MT)
02:3
5:04
(G
MT)
02:3
8:06
(G
MT)
02:4
1:07
(G
MT)
02:4
4:10
(G
MT)
02:4
7:12
(G
MT)
02:5
0:13
(G
MT)
02:5
3:13
(G
MT)
02:5
6:15
(G
MT)
02:5
9:17
(G
MT)
03:0
2:29
(G
MT)
03:0
5:31
(G
MT)
03:0
8:40
(G
MT)
03:1
1:43
(G
MT)
03:1
4:49
(G
MT)
SNR
Phase 0 – distance
• Third test in a large open room (Harrington Auditorium)
• At about 15 yards, bandwidth dropped significantly just like in the previous experiments.
• At 53 yards, it returned to full bandwidth!– Shape of the room
Bandwidth
0
1
2
3
4
5
6
1 5 9 13 17 21 25 29 33 37 41 45 49 53
Bandwidth
Phase 0 – distance
Phase 1 – building materials
• Unfortunately, by the time that we were about to conduct the building materials test, we realized it was ill conceived
• Learned from the distance tests that radio waves are much less line-of-sight than we thought
• We didn’t have the resources to do a better test, so we decided to continue with the test as-is.
Phase 1 – building materials
4
4.1
4.2
4.3
4.4
4.5
4.6
4.7
Material
Mbp
s
No Obstruction
3/4" Particle Board
Cement Board
Sheetrock
Polyisocyanurate Insulation
Ceiling tile (hard)
Ceiling tile (foam)
1/2" 3ply birch plywood
3/8" 3ply Pine
3/4" 4ply Plywood
3/8" 4ply Plywood
Insulation (pink panther)
Phase 1 – building materials
Phase 2 – environmental
• What are the effects of humidity on wireless, if any?
• Very slight downward trend
SNR
010203040506070
Tim
e (G
MT
)
23:5
9:27
(G
MT)
00:0
5:24
(G
MT)
00:1
1:18
(G
MT)
00:1
7:11
(G
MT)
00:2
3:03
(G
MT)
00:2
8:54
(G
MT)
00:3
4:46
(G
MT)
00:4
0:41
(G
MT)
00:4
6:34
(G
MT)
00:5
2:29
(G
MT)
00:5
8:23
(G
MT)
01:0
4:16
(G
MT)
01:1
0:08
(G
MT)
01:1
6:00
(G
MT)
01:2
1:55
(G
MT)
SNR
Phase 3 – interference
• Tested with two sources of noise in the 2.4ghz band– Microwaves– Roommate’s phone that ruins my
wireless headphones
Phase 3 – interference
Effect of Microwave Oven on Throughput
0
1
2
3
4
5
6
1 2 3 4 5 6 7 8 9 10 11
Channel
Mb
ps Microwave Off
Microwave On
Phase 3 – interference
Effect of 2.4Ghz Phone on Throughput
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
1 2 3 4 5 6 7 8 9 10 11
Channel
Mbp
s No Phone
Idle Phone
In Use
Conclusions
• Wireless signals may be too complex to conduct research with based solely on empirical evidence
• Although we did not reach many of out goals, we learned a lot about wireless– Line of sight is less important than
structures that tend to “focus” radio waves such as the ends of Harrington
– Without surfaces to bounce off of, 802.11b behaves very poorly as shown by our outdoor test
Conclusions
– Interference with other devices using the same band is a big problem
– If you get poor performance, try moving just a little bit – small changes in position or orientation can have huge effects on performance
– Always keep a close watch on your laptop
Questions?