hands-on deep dive into roaming and mesh...
TRANSCRIPT
Hands-on deep dive into roaming and mesh performance
This seminar took place on 18-Jul-2018
at octoScope’s San Jose office.
Videos of the seminar on YouTubePart 1 octoScope - introduction to the octoBox personal testbed and its new mesh testing featuresPart 2 dB Performance - demo of 802.11r fast transitionPart 3 SmallNetBuilder.com - demo of roaming in a mesh and observations on 11kvr capabilities, or lack thereofPart 4 dot11 Labs - overview of a mesh roaming test in a test house and in the octoBox
www.octoscope.com
Hands-on deep dive into roaming and mesh performance
What’s the secret of good wireless coverage? Overprovisioning of APs and extenders?
Proper operation of a wireless mesh? Band steering and roaming of the stations?
In this seminar we show what the latest testing reveals about common AP/extender
systems, such as Netgear Orbi and Google Wifi. We look in-depth at the roaming, mesh,
band steering and load balancing test methods.
Our presenters are acclaimed wireless test experts from SmallNetBuilder.com, dot11 Labs,
dB Performance and octoScope. The tests they demonstrate were performed in the
octoBox personal testbed and in a test house.
2
www.octoscope.com
Tim Higgins – SmallNetBuilder.com
Tim Higgins spent a few decades as an engineer and manager in the
automatic test equipment industry. This taught him how to make products
tell their secrets, which he now shares via SmallNetBuilder.com.
For almost 20 years, Tim has been setting the standard for test, analysis
and review of consumer Wi-Fi and other networking products. His work has
been cited by Bloomberg Business, USA Today, Wall Street Journal, New
York Times, Computerworld, Forbes and other leading publications.
3
www.octoscope.com
Dennis Bland – dB Performance
Dennis Bland, President of dB Performance, has been involved in WLAN
technology since 2001, with ten years of WLAN fast-roaming and security
experience. Dennis participated in the IEEE 802.11e and Wi-Fi Alliance
WPS working groups, collaborated with Cisco in implementing CCX test
plans, deployed an IEEE 802.11k/r/v software suite and performance test,
and recently released an IEEE 1609 security stack for automotive
applications. Prior to dB Performance, Dennis worked in both hardware
and software development roles at Devicescape, Wind River, and Nortel.
Dennis is the sole inventor of three patents and has designed resonant-
cavity antennas and wide-bandwidth HDTV antennas as a side hobby.
4
www.octoscope.com
Craig Smith – dot11 Labs
Craig Smith, the Principal of dot11 Labs, is one of the Wi-Fi pioneers who
has been working with Wi-Fi back when 802.11b was cutting edge
technology. Craig’s career includes wireless companies such as Intersil,
2Wire, Airgain and Xirrus. In 2016, Craig founded dot11 Labs, which
specializes in Wi-Fi performance testing and offers a unique set of open-air
and octoBox test services.
5
www.octoscope.com
Leigh Chinitz
Leigh Chinitz, Ph.D. has worked in the wireless and data communications
field since 1993 with companies including Casa Systems, Cisco and
Motorola. Prior to joining octoScope, he was a wireless strategist at the
Casa CTO office. Leigh joined Casa from Cisco where he was a Wi-Fi
Advanced Services Solutions Architect. His background includes Wi-Fi,
UMTS, CDMA, femtocell, small cells and regulatory work with the FCC. He
is a frequent speaker at industry conferences, and a contributor to the Wi-Fi
Alliance, IEEE, TIA, and ITU standards development. Leigh has a BS in
Physics from Yale, a Ph.D. and an MBA. He holds 20 US patents.
6
www.octoscope.com
Large, shielded anechoic
chambers
• Pros:No interference from outside sources
Repeatable measurements possible
Measures antenna patterns (in 4π steradians)
• Cons:Large and expensive
Complicated to use
Shared among all engineers at one, or even multiple,
locations
8
www.octoscope.com
• ProsCheaper and easier than using a chamber
Gives a measurement of performance vs. range
Gives a sense of performance in a real environment
• ConsExtremely susceptible to outside interference
Increasingly difficult to find locations without significant wireless
Very difficult to get repeatable resultsSmall changes in the environment will change the results
Walk-testing in a real environment
9
www.octoscope.com
Shielded from outside interference
Performance in a real-world environment
Repeatable results
Inexpensive
Easy to useP
hysi
cal D
ista
nce
Measure performance vs. range
11
www.octoscope.com
Shielded environments
Programmable Attenuation
Multi-path emulation
Repeatable results
Inexpensive
Easy to useR
F A
ttenu
atio
n
Interference Generation
12
www.octoscope.com
octoBox Personal Testbed
• Reduce wireless test time from weeks to hours
Complete isolation and repeatable RF environment minimizes
time-consuming open-air testing
Automation accelerates data collection, improves test coverage
and product quality
• Demonstrate highest achievable performance
Ideal MIMO environment for highest possible throughput
Supports latest technologies, such as 160 MHz 802.11ac,
802.11ax, MU-MIMO and Beamforming
• Qualify User Experience
Emulate real-world challenges
Programmable range of condition from best MIMO environment
to challenging real-life impairmentsBOX-18
BOX-38
Multipath emulator
14
www.octoscope.com
The Pal
• The Pal is a versatile Wi-Fi wave 2 test instrument that functions as an expert
analyzer, traffic partner, virtual station emulator and a load generator for testing
throughput, capacity and device functionality.
15
www.octoscope.com
The superPal™
Qualcomm Hawkeye QCN5054/QCN5024; Silicon Labs EFR32
10 GbE10 GbE
Interference
Trigger
5 GHz
5 GHz
2.4 GHz
Bluetooth/Thread
Bluetooth/Thread
Synthesizer
11a/n/ac/ax
4x4 MIMO 160 MHz
8x8 MIMO 80 MHz
11b/g/n/ac/ax
4x4 MIMO
Processing
subsystem
quad-core
ARM Cortex
64-bit
2 GHz
Hexagon
DSP engine
Hexagon
DSP engine
Hexagon
DSP engine5 synchroSniffing probes – 3 Wi-Fi, 2 Bluetooth
5-probe tri-band expert analysis and monitoring
3 AP traffic partners
3 STA traffic partners
96 vSTA (virtual station) traffic partners
16
www.octoscope.com
octoScope CustomersOperators Labs Chipset vendors Equipment vendors
17
www.octoscope.com
Tests Supported by the octoBox Testbed
Adjacent Channel Interference
Co-Channel Interference
Bluetooth
Baby monitor
ZigBee
Radar
…
MIMO = multiple input multiple output
MU-MIMO = multi-user MIMO
Throughput
Range
Orientation
Channel width (20/40/80/160 MHz)
Multipath
Channel frequency
Interference
Partner device (802.11a/b/g/n/ac)
Number of partner devices
Network load
Motion
Data rate / Modulation Coding Scheme (MCS)
Forwarding rate
Packet Error Rate
Data rate adaptation
Association capacity
Receiver performance
Roaming
Auto Channel Selection
MIMO-OTA
MU-MIMO
Exponential number of tests vs. variables
18
www.octoscope.com
Wireless Test Applications
• Performance
MIMO OTA throughput
MU-MIMO gains
Load testing
Mesh, Roaming
RX sensitivity
• User Experience
Adaptation to impairments, such as
path loss, interference, multipath, load
Roaming behavior – find sticky clients
DFS (dynamic frequency selection)
19
www.octoscope.com
camera
Plastic rails and hardware
Plastic turntable
Sub 6 GHz test
antennas
Filtered data
ports (Ethernet,
USB, etc.
octoScope’s BOX-38 sub 6 GHz chamber
20
www.octoscope.com
Sure, sure. Sounds great. But what about
for more complicated wireless topologies?
21
www.octoscope.com
Interest in mesh system testing using
the octoBox has never been higher
26
www.octoscope.com
Why mesh (and other multi-AP solutions)?
The old days of “pretty good”, “good enough” coverage are over.
Almost (?) everything is
wireless now, and people care
Coverage has to be everywhere, and to make
that happen, companies are employing mesh
and extender systems, implementing roaming,
band steering, and load balancing.
28
www.octoscope.com
Mesh / extender product examples
eero
Netgear
AmpliFi
Linksys
Luma
TP-Link
Nest Cam
AirTies
29
www.octoscope.com
Ethernet vs. mesh backhaul
Traditional
WLAN
Mesh
Mesh links
Client links
Wired links
Mesh portal (Root)
Wired connection to each AP
30
www.octoscope.com
Wireless backhaul topologies
Linear
Star
Full or partial mesh
31
www.octoscope.com
4-node full mesh
All nodes are connected
to all other nodes
Client can attach to any one of these nodes.
32
www.octoscope.com
4-node full mesh
Pal
1:4
1:2
1:2
1:2
1:2
Pal
Pal root
1
2
Pal
Pal
3
33
www.octoscope.com
Linear topology
(aka “string of pearls”)
root
1
2
3
root1 2 3
34
www.octoscope.com
Mesh tests
• Self-forming, self-healing
• Throughput vs. hops
• Load balancing
• Band steering
37
www.octoscope.com
Self-forming, self-healing
root
1
2
3
root
1
2
3
root
1
2
3
Configure path losses for
several possible paths
through the mesh (e.g.
between 1 and 3) and
observe traffic flow adapt.
38
www.octoscope.com
Band steering
Pal
1:4
1:2
1:2
1:2
1:2
Pal
Pal 4
1
2
Pal
Pal
3• Associate the local Pal on 5 GHz to 4.
• Move the local Pal away from 4 by varying
the quadAtten.
• Watch to see whether 4 transitions the Pal
to 2.4 GHz.
• Use the common Pal as the sniffer.
sniffer
42
www.octoscope.com
Load balancing
Pal
1:4
1:2
1:2
1:2
1:2
Pal
Pal 4
1
Pal
Pal
2
3
• Initially point all the STAs in 2 and 3 to 4, and
isolate them from everything else.
• Then provide an equal or better path for all the
STAs to 1. And crank up the load on 4 from its
local Pal.
• See whether 4 will start kicking out STAs to send
them to 1.
• Use the common Pal and the Pal local to 1 as
sniffers.
sniffer
sniffer
43
www.octoscope.com
Roaming
Controls motion,
velocity, pinging,
graphical reporting
45
www.octoscope.com
Pal
1:4
1:2
1:2
1:2
1:2
Pal
Pal
169.254.240.15
169.254.240.34
169.254.240.32
169.254.240.26
169.254.240.25
169.254.240.12
192.168.16.6
169.254.246.226
2C:27:9E:90:00:C0
169.254.240.13
169.254.240.21
169.254.240.20
169.254.240.23
169.254.240.28
192.168.8.1
5GHz - 08:02:8e:9b:fc:d9
2.4GHz -0e:02:8e:9b:fc:d7
192.168.8.2
08:02:8e:9c:36:d6
0e:02:8e:9c:36:d4
192.168.8.3
5GHz-08:02:8e:9b:23:5f
2.4GHz-0e:02:8e:9b:23:5d
169.254.240.24=20dB
20dB fixed192.168.16.8
04:f0:21:24:02:97
169.254.246.8
192.168.16.74
169.254.246.74
04:f0:21:23:ff:c0
root
1
2
4
192.168.8.4
5GHz – b0:b9:8a:4d:53:e2
2.4GHz – b6:b9:8a:4d:53:e0
0 Hop
Measure 1-hop throughput
by setting attenuators 12,
15, 20, 23, 25, 26, 28, and
34 to 63dB
47
www.octoscope.com
Pal
1:4
1:2
1:2
1:2
1:2
Pal
Pal
169.254.240.15
169.254.240.34
169.254.240.32
169.254.240.26
169.254.240.25
169.254.240.12
192.168.16.6
169.254.246.226
2C:27:9E:90:00:C0
169.254.240.13
169.254.240.21
169.254.240.20
169.254.240.23
169.254.240.28
192.168.8.1
5GHz - 08:02:8e:9b:fc:d9
2.4GHz -0e:02:8e:9b:fc:d7
192.168.8.2
08:02:8e:9c:36:d6
0e:02:8e:9c:36:d4
192.168.8.3
5GHz-08:02:8e:9b:23:5f
2.4GHz-0e:02:8e:9b:23:5d
169.254.240.24=20dB
20dB fixed192.168.16.8
04:f0:21:24:02:97
169.254.246.8
192.168.16.74
169.254.246.74
04:f0:21:23:ff:c0
root
1
2
4
192.168.8.4
5GHz – b0:b9:8a:4d:53:e2
2.4GHz – b6:b9:8a:4d:53:e0
1 Hop
Measure 1-hop throughput
by setting attenuators 12,
15, 20, 23, 25, 26, 28, and
34 to 63dB
48
www.octoscope.com
Pal
1:4
1:2
1:2
1:2
1:2
Pal
Pal
169.254.240.15
169.254.240.34
169.254.240.32
169.254.240.26
169.254.240.25
169.254.240.12
192.168.16.6
169.254.246.226
2C:27:9E:90:00:C0
169.254.240.13
169.254.240.21
169.254.240.20
169.254.240.23
169.254.240.28
192.168.8.1
5GHz - 08:02:8e:9b:fc:d9
2.4GHz -0e:02:8e:9b:fc:d7
192.168.8.2
08:02:8e:9c:36:d6
0e:02:8e:9c:36:d4
192.168.8.3
5GHz-08:02:8e:9b:23:5f
2.4GHz-0e:02:8e:9b:23:5d
169.254.240.24=20dB
20dB fixed192.168.16.8
04:f0:21:24:02:97
169.254.246.8
192.168.16.74
169.254.246.74
04:f0:21:23:ff:c0
root
1
2
4
192.168.8.4
5GHz – b0:b9:8a:4d:53:e2
2.4GHz – b6:b9:8a:4d:53:e0
2 Hop
Measure 2-hop throughput
by setting attenuators 12,
15, 21, 23, 25, 26 and 28 to
63dB
49
www.octoscope.com
smartBox
Ports used for attaching
test devices
Ports available
for connections
to other
environments
51
www.octoscope.com
smartBox™
Instruments built into an octoBox:
- synchroSniff on 3 channels
- Expert monitoring
- Interference
- vSTA (96)
Pal-24
Pal-5
Pal-5
iGen
GigE switch
1:4
52
www.octoscope.com
synchroSniffer probe
synchroSniffer probe
synchroSniffer probe
synchroSniffer
57
www.octoscope.com
Two synchroSniffer probe view at client as it roams
Mesh node 1
Mesh node 2
Roaming packets
59
www.octoscope.com
Insight from the experts!
Craig Smith
Dennis Bland
Tim Higgins
60
www.octoscope.com
Agenda
• Introduction from octoScope
• Wi-Fi Roaming Secrets Revealed!
Tim Higgins, Small Net Builder
Looking at how roaming in Wi-Fi works. STAs are in control, but APs have a big influence.
• MESH Test in Open-Air and octoBox
Craig Smith, dot11Labs
Reviewing an extended open-air roaming test, where a bug was observed and then recreated in the octoBox. A
prototype smartBox and synchoSniffer packet capture were used to isolate the issue. The bug fix was tested and
confirmed in the octoBox and open-air.
• 802.11r Fast-Transition Demo
Dennis Bland, dB Performance
802.11r BSS Fast Transition, with Apple iPAD roaming between two Cisco 1602 APs.
Determine the precise roaming time between AP1 and AP2, using three octoBoxes, two quadAttens, and two Pals to
capture traffic on different channels. octoScope's synchroSniffer will merge the packet captures into a single file to
easily analyze roam time.
61
802.11r Fast-Transition Demo July 18, 2018
Contact: Dennis Bland
Email:
www.octoscope.com
About dB Performance
• IEEE 802.11k/r/v and Cisco CCX software provider for embedded OEMs
• CCX pre-certification testing
• 802.11k/r/v protocol and performance testing
• IEEE 1609 security stack implementation
• 802.11p/DSRC
• C-V2X R14+ (peer-to-peer 5G)
• Wi-Fi Alliance Member and Cisco Solution Partner
63
www.octoscope.com
802.11r Today
Becoming more important…
• Multiple APs in the home – consumers want seamless media connectivity
• New commercial markets (DSRC, trackside AP, IoT)
But difficult to implement…
• Network compatibility issues with legacy clients (see Apple Adaptive Roaming)
• Vendor-specific implementations of STA-side FT (Fast Transition) and 802.11k
• Possible conflicts with AP-side TM (Transition Management)
• No industry-standard performance test plan available
Qualitative and quantitative testing is essential
• octoScope testbed for consistent, repeatable results and stress testing
• RF Air interface and MIMO capability allows for “black box” testing
64
www.octoscope.com
Testbed Setup to Measure 802.11r Performance
Pal-5
Pal-5
Legend:
Cisco 2504 WLAN Controller- PoE switch
- DHCP server
- RADIUS server
AP Console- AP configuration
- iperf endpoint
Cisco 1602 AP1
Apple iPAD Mini 2
octoBox server- Pal Configuration
- synchroSniffer
- Wireshark
Cisco 1602 AP2
Channel 36 capture
Channel 48 capture
4 SMA connectors
1 Ethernet cable
4x4 high gain antenna module
4 RF cables
3 dipole antennas
quadAtten
Data filter
Box-26
Box-26
Box-26
DUT
4x4 MIMO 1:2 splitter1:2
1:2
65
www.octoscope.com
Test Procedure
1. Configure AP1 quadAtten with no attenuation and AP2 quadAtten with 40 dB attenuation, so that
the DUT will initially favor AP1.
2. Associate DUT to AP1 using FT-802.1X with EAP-PEAP/MSCHAPv2 authentication.
3. Configure DUT to send continuous multicast data packets at least every 10 msec.
4. Start packet capture on two Pals to monitor the respective AP1 and AP2 channels.
5. To simulate the roam event, configure the AP1 quadAtten to increase attenuation by 2 dB/sec, and
configure the AP2 quadAtten to decrease attenuation by 2 dB/sec.
6. The test completes after 30 seconds, after which the capture files can be analyzed.
66
www.octoscope.com
Test Results
1. Using synchroSniffer™, merge the packet capture files from both Pals into a single, high-accuracy
time-synchronized .pcap file.
2. Using Wireshark, search for the 802.11 reassociation request sent from the DUT to AP2.
3. Calculate the time between the 802.11 reassociation request and the first multicast data packet
sent from the DUT to AP2. This is the 802.11r roam time.
67
Wi-Fi roaming
secrets revealed!
18 July 2018
Tim Higgins
Smallnetbuilder.com
www.octoscope.com
Agenda
•Who Am I?
•What Will You See?
•Roaming Testbed Overview
•Live Demo
•Review of Other Test Captures
•Q & A
69
www.octoscope.com
Who Is Tim Higgins?
•Aka The SmallNetBuilder Guy
•20+ years building big-iron mixed signal
test systems
•20+ years testing and reviewing
consumer networking gear
70
www.octoscope.com
What Are We Looking At Today?
•STAs rule in Roaming, but APs matter
too!
•How do APs influence the roam?
• 11k, v
•Deauths, disassociations
•Witholding, delaying probe
responses
71
www.octoscope.com
The Testbed
Ethernet (LAN)
Test Traffic
Test Console &
Traffic
Generator
Pal-24
Pal-5
Pal-245
Windows
STA
AndroidR
F S
plitt
er m
atrix
Android STA
Atten 5
Atten 4
octoBox anechoic
RF chambers
Atten 2
Box 3 - Root
Box 2 - Hop 1
Box 1 - Hop 2
Gigabit Switch
Atten 1
synchroSniffer
Atten 3
• AP / Mesh nodes go in
Root & Hop 1 boxes
• Attens 4 & 5 are cross-
ramped to roam STA
• Pal-24 & Pal-5 configured
as synchroSniffer to
capture all roaming mgmt
frames
• synchroSniffer moves with
STA; if the STA doesn’t
see it, it doesn’t matter.
STAs
Ethernet
(optional backhaul)
72
www.octoscope.com
The Testbed
Samsung Galaxy
Tab A 8" (2017
version
SM-T380
Windows STA:
Lenovo M600
Win 10 Pro
Intel Wireless-AC
8260
octoScope Pals
AP1AP2
73
www.octoscope.com
Why Pal For Roaming Test?
• Configurable STA characteristics
✓ # MIMO streams
✓ 802.11 type (a/b/g/n/ac)
✓ Channel Width
✓ Channel
✓ MCS rate
✓ And more!
• Predictable & Controllable roaming
✓ Roam Threshold (when)
✓ Roam Target Threshold (where)
✓ Band preference (where)
✓ 11v BSS Transition Mgmt request
support
74
www.octoscope.com
Why Pal For Roaming Test? - Log comparison
Windows & Android• Timestamp
• STA status (“Connected” only)
• IP address
• Associated BSSID
• Channel or frequency
• RSSI (dBm) for Android / Signal % for
Windows
• Link Rate
• Atten 1 value
• Atten 2 value
octoScope Pal• Timestamp
• IP address
• STA status (associated, scanning,
etc.)
• Associated BSSID
• Channel or frequency
• RSSI
• Tx Link Rate
• Rx Link Rate
• Roam status
• Roam scan channel
• Channel congestion
• Throughput
• Atten 1 value
• Atten 2 value
77
www.octoscope.com
RoamingTest Sequence
1. Start synchroSniffer packet capture
2. Set Attenuator 5 to 0 dB, Attenuator 4 to a larger value. This ensures STA will
connect to AP1.
3. Associate STA
4. Start traffic between AP and STA (optional)
5. Roam device to AP2 by increasing Atten5 and decreasing Atten 4 until set limits are
reached. Log STA info (RSSI, channel) after each attenuator change.
6. Continue to monitor connection after roam to check for post roam band-steering or
other changes
7. Reverse roam from AP2 to AP1 by decreasing Atten 5 and increasing Atten 4
8. Monitor post roam as above
9. Stop traffic
10.Stop synchroSniffer capture and save
11.Close STA log
78
www.octoscope.com
Demo: Google WiFi w/ Pal-245 STA
Ethernet (LAN)
Test Traffic
Test Console &
Traffic
Generator
Pal-24
Pal-5
Pal-245
RF
Spl
itter
mat
rix
Atten 5
Atten 4
octoBox anechoic
RF chambers
Atten 2 = 3 dB
Box 3 - Root
Box 2 - Hop 1
Gigabit Switch
Atten 1
synchroSniffer •Wi-Fi backhaul
•Unused attenuators are
set to 63 dBSTAs
79
www.octoscope.com
Windows STA
Ethernet (LAN)
Test Traffic
Test Console &
Traffic
Generator
Pal-24
Pal-5
Windows
STA
RF
Spl
itter
mat
rix
Atten 5
Atten 4
octoBox anechoic
RF chambers
Box 3 - Root
Box 2 - Hop 1
Gigabit Switch
Atten 1
synchroSniffer • APs in Root & Hop 1
boxes
•Windows STA: Lenovo
M600, Win 10 Pro, Intel
Wireless-AC 8260
• STA data capture via SSH
& netsh commands
STAs
Atten 2 = 3 dB
80
www.octoscope.com
Android STA
Ethernet (LAN)
Test Traffic
Test Console &
Traffic
Generator
Pal-24
Pal-5
AndroidR
F S
plitt
er m
atrix
Android STA
Atten 5
Atten 4
octoBox anechoic
RF chambers
Box 3 - Root
Box 2 - Hop 1
Box 1 - Hop 2
Gigabit Switch
synchroSniffer
• AP / Mesh nodes go in Root
& Hop 1 boxes
• Android STA: Samsung
Galaxy Tab A 8" (2017
version - SM-T380 – 1x1
11n)
• Custom Android app used
for STA telemetry
• Ethernet backhaul needed
due to lower signal levels
STAs
Ethernet
backhaul
Atten 2 = 3 dB
Atten 1 = 63 dB
Atten 3 = 0 dB
81
www.octoscope.com
Thank You
Thanks!For more roaming fun, see:
•Wi-Fi Roaming Secrets Revealed
•Wi-Fi Roaming Secrets Revealed - Part 2
•Wi-Fi Roaming Secrets Revealed - Part 3
83
Mesh test in open-air and octoBox July 18, 2018
Craig Smith – dot11 Labs
www.octoscope.com
Who is Craig Smith?
• 17 years of Wi-Fi Experience
• Employed by Intersil, 2Wire, Airgain & Xirrus
• Worked as Applications Engineer, FAE, Systems &
QA Engineer
• Founded dot11Labs in 2016
• dot11 Labs is an octoScope test lab
85
www.octoscope.com
Testing & Debugging a 2 Node MESH
• Define a 5 minute multi-step open-air roaming test
• Observe sub-optimal performance with the Orbi MESH System
• Reproduce performance issue in octoBox
• Collect packet capture using synchroSniffer & smartBox
• Upgrade Orbi Firmware and confirm fix
86
www.octoscope.com
Devices Under Test
# Vendor Model Hardware Software
1 Netgear Orbi RBK40 RBR40, RBS40 (2 nodes) V2.0.0.56 & V2.1.10.4
2 Apple MacBook Pro 13.3” 802.11n/ac 3x3 10.12.6
87
www.octoscope.com
Test House (Open-Air)
4,667 sqft. home
• Two levels
• Construction: Drywall on stud
• Built 1994
• Lot size: 3.7 acres
90
www.octoscope.com
Floor Plan
2nd level1st level
Open to 1st levelOpen to 2nd level
95’
35’
60’
91
www.octoscope.com
Open-Air Test Configuration
Root Client
Client assoc to Root Node Client assoc to MESH Node
Root ClientMESH
Node
92
www.octoscope.com
Moving the Client in Open-Air
1st Turntable
Root
2nd Turntable
MESH
30 second walk to next TT
93
www.octoscope.com
Open-Air Floor Plan
1st level
2nd level
Open to 1st level
95’
MESH Node
Orbi
router
1 2Root Node
1st, 3rd, 5th
2nd, 4thOrbi
satellite
94
www.octoscope.com
Realtime Throughput
IxChariot Downlink TCP traffic
Root:
CH 48
MESH Node:
CH 48
MESH Node:
CH 4
Root:
CH 48
Root:
CH 4
TT 1 TT 2 TT 1 TT 2 TT 1 Walk Walk Walk Walk
97
www.octoscope.com
Realtime Throughput
TT 1 TT 2 TT 1 TT 2 TT 1 Walk Walk Walk Walk
1st Test Run
2nd Test Run
98
www.octoscope.com
Non-Roaming Throughput
Channel
BSSID
TT 1 TT 2 TT 1 TT 2 TT 1
Time
99
www.octoscope.com
dot11 Labs octoBox Photos
• 3 Enclosures
• Dual-Purpose
BOX-38
smartBox
MESH
Node
Client
Root
BOX-38
BOX-18
101
www.octoscope.com
BOX-38smartBox
MESH
Node
Client
Root
Backhaul
(static)
4x4 high gain
antenna module
quadAtten
4 RF cables
octoBox Set-Up
Fronthaul
(dynamic)
Fronthaul
(dynamic)
BOX-38
BOX-18
1:2
4x4 MIMO 1:2
SPLITTER
1:2
102
www.octoscope.com
Calibrating Open-Air
1st level
95’
MESH Node
Orbi
router
1Root Node
Orbi
satellite
Location Node Band
(GHz)
BSSID Open-air
RSSI (dBm)
1 Root 2.4 A0:40:A0:5E:A8:F5 -27
Root 5 A0:40:A0:5E:A8:F8 -36
MESH 2.4 A0:40:A0:6A:7A:4D -62
MESH 5 A0:40:A0:6A:7A:50 -76
40dB
difference
between TTs
at 5GHz
40dB
103
www.octoscope.com
Calibrating Open-Air
1st level
95’
MESH Node
Root Node
Location Node Band
(GHz)
BSSID Open-air
RSSI (dBm)
2 Root 2.4 A0:40:A0:5E:A8:F5 -70
Root 5 A0:40:A0:5E:A8:F8 -86
MESH 2.4 A0:40:A0:6A:7A:4D -34
MESH 5 A0:40:A0:6A:7A:50 -47
40dB
difference
between TTs
at 5GHz
40dB
2
Orbi
router
Orbi
satellite
104
www.octoscope.com
Realtime Throughput
IxChariot Downlink TCP traffic
Root:
CH 48
MESH Node:
CH 48
Root:
CH 4Root:
CH 4
Static
Throughput
To Root
Ramp
Attenuators
Root:
CH 4
Static
Throughput
To MESH Node
Ramp
AttenuatorsStatic
Throughput
To Root
Ramp
AttenuatorsStatic
Throughput
To MESH Node
Ramp
Attenuators
Static
Throughput
To Root
106
www.octoscope.com
octoBox Results – Non-Roaming Throughput
Time
Channel
BSSID
107
www.octoscope.com
PCAP Analysis
Root Node:
CH 48MESH Node:
CH 48
Root Node:
CH 4
108
www.octoscope.com
Why Does the Client Deauth the 5G Root?
• Reason Code: Class 3 frame received from nonassociated STA
110
www.octoscope.com
Static
Throughput
To Root
Ramp
Attenuators
Static
Throughput
To MESH Node
Ramp
AttenuatorsStatic
Throughput
To Root
Ramp
AttenuatorsStatic
Throughput
To MESH Node
Ramp
Attenuators
Static
Throughput
To Root
octoBox - Realtime Throughput
FW 2.0.0.56
FW 2.1.4.10
IxChariot Downlink TCP traffic
113
www.octoscope.com
Open-Air - Realtime Throughput
TT 1 TT 2 TT 1 TT 2 TT 1 Walk Walk Walk Walk
FW 2.0.0.56
FW 2.1.4.10
115
www.octoscope.com
Open-Air – Non-Roaming Throughput
Time
Channel
BSSID
116
www.octoscope.com
Boston area headquarters305 Foster Street
Littleton, MA 01460 USA
Tel: +1.978.222.3114
California office780 Montague Expressway, Building 1
San José, CA 95131 USA
Tel: +1.978.339.9431
118