Wireless commonsense fontsfixed

Wireless Common Sense

Shifting the Collection of Prejudices

Robert J. Berger Internet Bandwidth Development Consulting

rberger@ibd.com

Common senseis the collection of

prejudices acquired by age eighteen- Albert Einstein

Tech You Grew Up with Shapes your Prejudices of Wireless Tech Capabilities

Claude Shannon’s Communications

Theory Shows us the Reality beyond our prejudices

C=W log (1+S/N)

Shannon’s Theory

where; C=channel capacity (bits/second)

C=W log (1+S/N)

Shannon’s Theory

where; C=channel capacity (bits/second)W=Bandwidth (How much Spectrum we are using)

C=W log (1+S/N)

Shannon’s Theory

where; C=channel capacity (bits/second)W=Bandwidth (How much Spectrum we are using)

S=signal power, and N=noise power

C=W log (1+S/N)

Propagation Characteristics & Path Loss

• Determines how much of the signal power gets to the receiver

• The higher the frequency- the less easily the signal propagates

• The higher the frequency- the less easily the signal propagates- The more buildings, trees, hills the

more attenuation

• The higher the frequency- the less easily the signal propagates- The more buildings, trees, hills the

more attenuation

• Physical Reality is such a Drag!!

Radio & TV: Broadcast HIGH POWER & low frequency

Radio & TV: Broadcast HIGH POWER & low frequencyMEGA WATTS of Power!!!

Spectrum in the 80 - 800 MHz range

50Khz - 6Mhz Channel Bandwidth

Gets Large Channel Capacity via Power and Significant Spectrum

Tech from the 1930’s

Very simple analog modulationUses spectrum inefficiently

Receiver has difficulty rejecting noise

Cell Phones: Tiny Channel Capacity

Low Signal Power & Advanced Modulation

800 - 1900Mhz a bit more difficult propagation, but still good

But only requires 4kbps Channel Capacity

High speed data phones will have less coverage or need more basestations

Past Experience Does Not Apply to WiFi

Uses Unlicensed Spectrum2.4Ghz & 5 - 6Ghz, Very Low Power

Rapid comidification & evolution -Very low cost & advanced modulation

Developed “Bottoms Up” like the Internet, not from Carriers or

Broadcasters

Unlicensed Wireless Constraints

Original Spectrum considered “Junk Spectrum”2.4Ghz is absorption frequency of water5 - 6 Ghz propagates about 1/2 as well

Can’t interfere, but must accept interference

Very low power, but lots of spectrumVery low power, but lots of spectrum

Very low power, but lots of spectrum

Very short distance with high bandwidth

Average WiFi Link Budget

AP to Client limited by weakest linkAvg clients’ Signal Power: 15dBmAvg AP Sensitivity 1Mbps: -94dBm

AP to Client limited by weakest linkAvg clients’ Signal Power: 15dBmAvg AP Sensitivity 1Mbps: -94dBmLink Budget avg client to AP:

signal power + xmit antenna gain - free space path loss + receiv antenna gain - receive

sensitivity = Margin

AP to Client limited by weakest linkAvg clients’ Signal Power: 15dBmAvg AP Sensitivity 1Mbps: -94dBmLink Budget avg client to AP:

signal power + xmit antenna gain - free space path loss + receiv antenna gain - receive

sensitivity = Margin15dBm + 2dB - (-84) + 7dB - (-94dBM) =

34 dBm Margin at 150 Meters

Link Margin

100 75 50 25

24 2628

35 3739

1 Mbps 11 Mbps

Meters

Link MarginThe Free Space Loss calculations are optimistic

100 75 50 25

24 2628

35 3739

1 Mbps 11 Mbps

Meters

Realistic distance of a laptop to an AP is ~30m in the open

100 75 50 25

24 2628

35 3739

1 Mbps 11 Mbps

Meters

Realistic distance of a laptop to an AP is ~30m in the open

At 150 Meters, One good tree will knock out your link 15

0 75 50 25

24 2628

35 3739

1 Mbps 11 Mbps

Meters

Real World Constraints

Real World ConstraintsWiFi originally designed for indoor LAN

WiFi CSMA does not handle large numbers of clients accessing an AP

Nearby clients and APs not associated to an SSID can cause contention

Asymmetrical Power causes client misbehavior

Obstructions: One Tree: 20dB One Wall: 10dB

WiFi has to Evolve Around Limitations

802.11 Phy has already evolved from 1Mbps to 802.11n with 125Mpbs

Distance not really improved (but more bandwidth for same short distance)

Contention by nearby nodes not addressed yet (802.11s may, Proprietary extensions today)

WiFi Recapitulates Ethernet Evolution

Originally Considered a “Toy”

Started out as LAN evolved to MAN

“Good Enough” and much less expensive

Designed by NetHeads, not BellHeads

Continually pushed beyond original limitations, first by proprietary extensions,

that migrate to standards

Still at the Hub & Bridge

Stage Compared toEthernet Evolution

Mesh: Ideal Architecture for Unlicensed Bands

• Take advantage of

• Take advantage of- Limited range

100 75 50 25

24 26 2831

4035 37 39

• Take advantage of- Limited range- Low Cost

components

components- Good amount of

Spectrum

components- Good amount of

Spectrum- Millions of Nodes

Not Yet Near the Ideal

Not Yet Near the IdealSome use 2.4Ghz for backhaul & access!

Uses standard 802.11 CSMA protocol

Gives great demo: Works great unloaded

As usage grows, contention is multiplied & becomes unusable

Today’s Best Practices for Mesh

Use different bands for Mesh & Access

Coordinate transmissions & directional antennas

Multiple Mesh Radios pluses and minuses

Features

Wireless! High-performance, multi-node and

multi-radio mesh architecture.! Low latency and high throughput

across multiple wireless hops.! Up to six radios per node

(upgradeable).! Full duplex mesh.

System! All nodes auto-discover and self-

configure.! Self-tuning and self-healing mesh

for network optimization.! User definable QoS with voice,

video and data prioritization.! Up to 16 BSSIDs per radio.! Multiple SSIDs (per network and

per node) and VLAN tagging, withconfigurable security parameterson a per-SSID basis.

! Session persistence for roaming,path optimization or failover.

! Manager/One® Web interfaceprovides a full suite of intuitivemanagement tools at the network,node, and radio levels.

! Additional remote managementoptions include SNMP, CLI overTelnet or SSH, HTTP/HTTPS,DHCP, and BOOTP.

! Seamless interoperability with theStrix Access/One® Indoor WirelessSystem (IWS) and Edge WirelessSystem (EWS).

Security! Supports all industry standard

security protocols.! RADIUS, WPA, EAP-MD5, EAP-

TLS, PEAP-TTLS authentication.! 802.11i with AES, WEP and TKIP

encryption.! MAC address Access Control Lists

on a per SSID basis.! Full VPN support.

OWS 3600 SERIES

DELIVERS INDUSTRYÊÂS HIGHEST THROUGHPUT,LOWEST LATENCY ACROSS MULTIPLE HOPS

The award winning Access/One® Outdoor Wireless System (OWS) 3600 is theindustry’s highest highest throughput, lowest latency modular multi-radio meshnetworking system. Utilizing Strix DMATM, the Access/One OWS delivers multi-radio, multi-RF and multi-channel capabilities usingadvanced algorithms to deliver high throughput overmultiple hops from the core to the edge of the network.The OWS intelligently self-tunes, self-configures andself-heals to optimize the overall performance andavailability. The OWS architecture makes 802.11 a fullduplex technology, moving traffic more efficientlythrough the network and utilizing different RFfrequencies and channels for network connectivity andclient access. RF Channels are selected dynamically,making the network more resilient to interference thanstandard mesh networks. Working closely together,these features deliver higher throughput and lowerlatency across multiple hops, supporting real timevoice, video, and data applications. The OWS scales

efficientlyand econ-omicallyminimizing the number of wiredtermination points required in thenetwork, greatly reduces deploymentand operating costs and the Total Costof Ownership (TCO). Extendedoperating temperature ranges andflexible mounting options make theOWS suitable for all types of

deployment scenarios. All OWS nodes can be centrally managed managed usingthe Manager/One® secure Web interface, or the carrier grade SNMP basedmanagement tools. Enhanced features including Virtual/Strix, Priority/One andRogue/One support deployments of mixed use networks where varying securityschemes are implemented based on user type and different levels of priority canbe assigned to the various network traffic. Access/One Network OWS is an idealsolution for carriers, service providers, municipalities, public safety, federalgovernment applications and more.

OWS is the most secure meshnetworking systemavailable, with

the tools to authenticate users,encrypt wireless traffic, andmonitor network activity all

provided as standard features.Secure private networks canoperate in tandem with open

public access networks—with dataintegrity guaranteed.

Strix Systems, Inc.26610 Agoura Road,Calabasas, CA 91302 USA

1-877-STRIXSYS (787-4979) Toll FreeWeb Site: www.strixsystems.com

www.strixsystems.com

Multiple tiers of Wireless backhaul

Expect < 1Mbps delivered19

State of Muni-Wireless(US Centric View)

First phase build-outs

Significant Learning Curve

First Generation Equipment

Be prepared for negative Hype CyclePeople’s expectations too high

Its still very early

WiMax?

WiMax?If WiFi recapitulates

the evolution of Ethernet;WiMax recapitulates the

evolution of all of Ethernet’s competitors

WiMax?

WiMax?No real advantage when comparing unlicensed

WiMax & WiFi

Licensed WiMax can be useful(ATM was useful for some things too)

WiMax & WiFi

Getting a License can be difficult & Expensive

WiMax & WiFi

Getting a License can be difficult & Expensive

Very good for feeding WiFi Mesh

WiMax?

Beware of Hype

Very few deployments of “real” WiMax

No deployments of Mobile WiMax

Wireless Tech of the Future

• Ultrawideband- Ultra Short Range- 2 to 3 x bandwidth

over 802.11n

• 60 - 70 Ghz- Extreme Bandwidth &

Directionality

over 802.11n

• 60 - 70 Ghz- Extreme Bandwidth &

Directionality

• Super Mesh- 802.11s evolution?- Short Range is good- More Bandwidth!

No Silver Bullets on the Horizon

Wireless Common Sense

Shifting the Collection of Prejudices

Robert J. Berger Internet Bandwidth Development Consulting

rberger@ibd.com

Wireless commonsense fontsfixed

Documents

Transcript of Wireless commonsense fontsfixed

Tackling Benchmark Problems of Commonsense Reasoningict.usc.edu/pubs/Tackling Benchmark Problems of Commonsense... · Tackling Benchmark Problems of Commonsense Reasoning ... gordon@ict.usc.edu

Commonsense Reasoning about Chemistry Experiments: Ontology and Representation Ernest Davis Commonsense 2009.

COMET: Commonsense Transformers for Automatic Knowledge ...

Commonsense Cookery Notes

Commonsense Judgments of Infanticide

Pm Commonsense Approach

Moore Commonsense

Commonsense Pain Rich

6 Keys to Commonsense Marketing

ThomasPaine CommonSense

CoSeDat - CommonSense

Commonsense Ai

NLog-like Inference and Commonsense Reasoning

BPN Radical CommonSense Cookbook

The Commonsense Case for Ethical Vegetarianism

Commonsense Reasoning, Commonsense …Commonsense Reasoning, Commonsense Knowledge, and The SP Theory of Intelligence J Gerard Wol August 7, 2018 Abstract This paper describes how

SCIENCE AND COMMONSENSE

Paine CommonSense

Elektronisches Postamt der Commonsense

Heterogeneous Graph Learning for Visual Commonsense Reasoningpapers.neurips.cc/...graph-learning-for-visual-commonsense-reasonin… · Heterogeneous Graph Learning for Visual Commonsense