`
`
`
`Eldad Perahia
`Robert Stacey
`
`Next Generation
`Wireless LANs
`
`
`
`
`802.11n and 802.11ac
`
`SECOND EDITION
`
`
`! CAMBRIDGE
`
`
`
`TP Link Ex. 1028
`Page 1
`
`

`

`CAMBRIDGE
`UNIVERSITY PRESS
`
`University Printing House, Cambridge CB2 8BS, United Kingdom
`
`Cambridge University Press is part of the University of Cambridge.
`
`It furthers the University’s mission by disseminating knowledge in the pursuit of
`education, learning and research at the highest international levels of excellence.
`
`www.cambridge.org
`Information on this title: www.cambridge.org/978 1107016767
`
`© Cambridge University Press 2008, 2013
`
`This publication is in copyright. Subject to statutory exception
`and to the provisions of relevant collective licensing agreements,
`no reproduction of any part may take place without the written
`permission of Cambridge University Press.
`
`First published 2008
`Reprinted with corrections 2010
`Second edition 2013
`6th printing 2016
`
`Printed in the United States of America by Sheridan Books, Inc.
`
`A catalog recordforthis publicationis available fromthe British Library
`
`Library of Congress Cataloging-in-Publication Data
`Perahia, Eldad, 1967 — author.
`Next generation wireless LANs : 802.11n, 802.1 lac, and Wi-Fi direct / Eldad Perahia, Intel Corporation,
`Robert Stacey, Apple Inc. — Second edition.
`pages
`cm
`ISBN 978-1~107-01676-7 (hardback)
`1. Wireless LANs.
`I. Stacey, Robert, 1967 — author.
`TK5105.78.P47
`2013
`621.39'8-de23
`
`IL. Title,
`
`2012033809
`
`ISBN 978-1-107-01676-7 Hardback
`
`Cambridge University Press has no responsibility for the persistence or
`accuracy of URLsfor external or third-party internet websites referred to
`in this publication, and does not guarantee that any content on such
`websites is, or will remain, accurate or appropriate.
`
`TP Link Ex. 1028
`Page 2
`
`TP Link Ex. 1028
`Page 2
`
`

`

`Introduction
`
`
`
`Wireless local area networking has experienced tremendous growth in the last ten years
`with the proliferation of IEEE 802.11 devices. Its beginnings date back to Hertz’s
`discovery of radio waves in 1888, followed by Marconi’s initial experimentation with
`transmission and reception of radio waves overlong distances in 1894. In the following
`century, radio communication and radar proved to be invaluable to the military, which
`included the developmentof spread spectrum technology. Thefirst packet-based wireless
`network, ALOHANET, wascreated by researchers at the University of Hawaii in 1971.
`Seven computers were deployed over four islands communicating with a central com-
`puterin a bi-directional star topology.
`A milestone event for commercial wireless local area networks (WLANs) came about
`in 1985 when the United States Federal Communications Commission (FCC) allowed
`the use of the experimental industrial, scientific, and medical (ISM) radio bands for the
`commercial application of spread spectrum technology. Several generations of propri-
`etary WLANdevices were developed to use these bands, inchiding WaveLANby Bell
`Labs. These initial systems were expensive and deployment was only feasible when
`running cable wasdifficult.
`Advancesin semiconductor technology and WLANstandardization with IEEE 802.11
`led to a dramatic reduction in cost and the increased adoption of WLANtechnology. With
`the increasing commercial interest, the Wi-Fi Alliance (WFA) was formed in 1999 to
`certify interoperability between IEEE 802.11 devices from different manufacturers
`through rigorous testing. Shipments of Wi-Fi certified integrated circuits exceeded a
`billion units per year in 2011 (ABIresearch, 2012) and are expected to exceed 2.5 billion
`units per year by 2016 (ABIresearch, 2012), as illustrated in Figure 1.1.
`Such large and sustained growth is due to the benefits WLANs offer over wired
`networking. In existing homesorenterprises, deploying cables for network access may
`involve tearing up walls, floors, or ceilings, which is both inconvenient and costly. In
`contrast, providing wireless network connectivity in these environments is often as
`simple as installing a single wireless access point. Perhaps more importantly though,
`the proliferation of laptops and handheld devices has meant that people desire connec-
`tivity wherever they are located, not just where the network connection is located.
`Network connectivity in a conference room or while seated on the sofa in the living
`roomare just two examples of the flexibility afforded by WLANs.
`There has been a proliferation of small scale deployments providing Internet access in
`coffee shops, airports, hotels, etc., which have come to be knownashotspots. In recent
`
`TP Link Ex. 1028
`Page 3
`
`TP Link Ex. 1028
`Page 3
`
`

`

`2
`
`Introduction
`
`
`
`Wi-FiICShipments(Billions) a
`
`
`
`
`
`
`
`
`
`
`Figure 1.1
`
`Wi-Fi IC shipments.Source: ABlresearch (2007, 2012),
`
`years, carriers with heavily congested cellular networks are deploying hotspots to off-
`load traffic from their cellular networks. Additionally, when these networks are used in
`conjunction with virtual private network (VPN) technology, employees can securely
`access corporate networks from almost anywhere.
`WLANproducts and systems started with 802.11b, 802.11g, and 802.1la standard
`amendments, which provided throughput enhancements overthe original 802.11 PHYs.
`Progress in WLAN technology continued with the development of 802.1 1n. Increased
`data rates were achieved with the multiple-input multiple-output (MIMO) concept, with
`its origins by Foschini (1996)at Bell Labs, In 2004, Atheros demonstrated that 40 MHz
`devices could be produced at almost the same cost as 20 MHz devices. Duringa similar
`time frame, the FCC and ETSI adopted new regulations in the 5 GHz bandthat added an
`additional 400 MHzof unlicensed spectrumfor use by commercial WLANs.
`These events paved the way for the broad acceptance of 40 MHz operating modesin
`802.11n. When spectrum is free, increasing the channel bandwidth is the most cost
`effective way to increase the datarate.
`Typically product developmentlags standardization efforts and products are released
`after the publication of the standard. An interesting event occurred in 2003 when
`Broadcomreleased a chipset based on a draft version of the 802.11g amendment, prior
`to final publication. This set a precedent for the flurry of “pre-n” products in 2005 and
`2006, as industry players rushedto befirst to market. Most of these products were either
`proprietary implementations of MIMO, or based on draft 1.0 of 802.11n, and thus
`unlikely to be compliant with the final standard.
`Throughearly 2007, major improvements andclarifications were madeto the 802.1 1n
`draft resulting in IEEE 802.1 1n draft 2.0. To continue the market momentumandforestall
`
`TP Link Ex. 1028
`Page 4
`
`TP Link Ex. 1028
`Page 4
`
`

`

`<i
`
`¥ = a
`
`1.1. An overview of IEEE 802.11
`— ee — eS
`
`3
`oe
`
`interoperability issues, the IEEE took the unusual step of releasing 802.11n D2.0 to the
`public while work continued toward the final standard, This allowed the Wi-Fi Alliance
`(WEA) to begin interoperability testing and certification of devices based on a subset of
`the 802.11n D2.0 features in May 2007. Wi-Fi certified 802.11n D2.0 products provide
`consumers the assurance ofinteroperability between manufacturers that was not guar-
`anteed by previous “pre-n” products. At the end of 2009, 802.11n wasfinally approved
`and the WFA updated the certification program to reflect support for the approved
`standard. Full
`interoperability was maintained between 802.11n D2.0 and the
`approved standard products. These were majorsteps in speeding up the standardization
`and certification process of new technology.
`As this process was successful for the industry and beneficial for the consumers,
`802.1 lac will follow a similar path. It is expected that 802.1lac products based on an
`early draft will be certified and on the market in early 2013. Completion ofthe 802.1 1ac
`is expected by the end of 2013.' Certification based on the approved standard will take
`place in a similar time frame.
`
`1.1
`
`An overview of IEEE 802.11
`
`layers (PHYs) and a common
`The IEEE 802.11 standard defines multiple physical
`medium access control (MAC) layer for wireless local area networking. As a member
`of the IEEE 802 family of local area networking (LAN) and metropolitan area network-
`ing (MAN) standards, 802.11 inherits the 802 reference model and 48-bit universal
`addressing scheme. The 802 reference model
`is based on the OSI reference model
`described in Table 1.1, In this model, the 802.11 MAC and 802.2 logical link control
`(LLC) sublayers form the data link layer and the 802.11 PHY the physicallayer.
`
`The 802.11 MAC
`
`Theinitial version ofthe 802.11 standard was completed in 1997, Influenced by the huge
`market success of Ethernet (standardized as IEEE 802.3), the 802.11 MAC adopted the
`same simple distributed access protocol, carrier sense multiple access (CSMA), With
`CSMA,a station wishing to transmit first listens to the mediumfora predetermined period,
`If the medium is sensed to be “idle” during this period then the station is permitted to
`transmit, Ifthe medium is sensed to be “busy,” the station has to defer its transmission, The
`original (shared medium) Ethernet used a variation called CSMA/CD or carrier sense
`multiple access with collision detection. After determining that the medium is “idle” and
`transmitting, the station is able to receive its own transmission and detectcollisions. If a
`collision is detected, the two colliding stations backoff for a random period before trans-
`mitting again. The random backoff period reduces the probability of a second collision.
`With wireless it is not possible to detect a collision with one’s own transmission
`directly in this way: thus 802,11 uses a variation called CSMA/CAor carrier sense
`
`'
`
`‘The readeris referred to http://grouper.ieee.org/groups/802/1 1/Reports/802.11_Timelines.htm forthelatest
`update on the timeline of 802.1 Lac
`
`TP Link Ex. 1028
`Page 5
`
`TP Link Ex. 1028
`Page 5
`
`

`

`4
`
`Introduction
`
`Table 1.1 OS! reference model (adapted from ISO/MEC 7498-1, 1994)
`ieS
`
`OSI reference
`
`Description Examplesmodel layers Layer categories
`
`
`
`
`
`Application
`
`Presentation
`
`Session
`
`Transport
`
`Network
`
`Data link
`
`Interacts with the software applications
`that implement a communicating component
`Establishes context between application-layer
`entities
`Establishes, manages, and terminates
`communication sessions
`Provides an end-to-end reliable data transfer
`service, including flow control,
`segmentation/desegmentation and error
`contro]
`Provides the means for transferring variable
`length data sequences from a source device
`to a destination device. Maintains the quality
`of service requested by the transport layer
`Provides the means fortransferring data
`between devices
`
`HTTP, FTP, SMTP
`
`Application
`
`MIME, TLS, SSL
`
`Named pipe, NetBIOS
`
`TCP, UDP
`
`IP (IPv4, IPv6), ICMP,
`[Psec
`
`LLC
`
`Data transport
`
`802.11 MAC
`802.1] PHY
`Provides the electrical and physical
`Physical
`specifications for devices
`
`multiple access with collision avoidance. With CSMA/CA,if the station detects that the
`medium is busy,
`it defers its transmission for a random period following the medium
`going “idle” again. This approach ofalways backingoff for a random period following
`anotherstation’s transmission improves performance since the penalty for a collision is
`much higher on a wireless LAN than on a wired LAN. Ona wired LAN collisions are
`detected electrically and thus almost immediately, while on wireless LAN collisions are
`inferred through the lack of an acknowledgement or other response from the remote
`station once the complete frame has been transmitted.
`There is no doubtthat the simplicity ofthis distributed access protocol, which enables
`consistent implementation acrossall nodes, significantly contributed to Ethernet’s rapid
`adoption as the industry LAN standard. Likewise, the adoptionby the industry of 802.11
`as the wireless LAN standard is in no small part due to the simplicity ofthis access
`protocol, its similarity to Ethernet, and again the consistent implementation acrossall
`nodes that has allowed 802.11 to beat out the more complex, centrally coordinated access
`protocols of competing WLAN technologies such as HyperLAN.
`
`Li2
`
`The 802.11 PHYs
`
`The original (1997) 802.11 standard included three PHYs: infrared (IR), 2.4 GHz
`frequency hopped spread spectrum (FHSS), and 2.4 GHz direct sequence spread
`
`TP Link Ex. 1028
`Page 6
`
`TP Link Ex. 1028
`Page 6
`
`

`

`1.1 An overview of IEEE 802.11
`
`5
`
`spectrum (DSSS). This was followed by two standard amendments in 1999: 802.11b
`built upon DSSSto increase the data rate in 2.4 GHz and 802.11a to create a new
`PHY in 5 GHz. 802.11b enhanced DSSS with complementary code keying (CCK),
`increasing the data rate to 11 Mbps. With higher data rates, IEEE 802.11b devices
`achieved significant market success, and markets for IR and FHSS PHYs did not
`materialize.
`The development of 802.11a introduced orthogonal frequency division multiplexing
`(OFDM)to 802.11. Even though 802.11a introduced data rates of up to 54 Mbps,it
`is confined to the 5 GHz band and, as a result, adoption has been slow. New devices
`wishing to take advantage of the higher rates provided by 802.1la but
`retain
`backward compatibility with the huge installed base of 802.11b devices would need to
`implement two radios, one to operate using 802.11b in the 2.4 GHz band and one to
`operate using 802.11a in the 5 GHz band. Furthermore, international frequency regu-
`lations in the 2.4 GHz band uniformly allowed commercial use, whereas in 1999 and
`2000 the non-military use ofthe 5 GHz band waslimited to select channels in the United
`States.
`In 2001, the FCC permitted the use of OFDM in the 2.4 GHz band. Subsequently, the
`802.11 working group developed the 802,11g amendment, which incorporates the
`802.1la OFDM PHYin the 2.4 GHz band, and adopted it as part of the standard in
`2003. In addition, backward compatibility and interoperability is maintained between
`802.1 1g and the older 802.11b devices, This allows for new 802.11 ¢ client cards to work
`in existing 802.11b hotspots, or older 802.11b embeddedclient devices to connect with a
`new 802.11g access point (AP). Because ofthis and new data rates of up to 54 Mbps,
`802.11g experienced large market success. A summary ofthe high level features ofeach
`PHY is given in Table 1.2.
`With the adoption of each new PHY, 802.11 has experienced a five-fold increase in
`data rate. This rate of increase continues with 802.11n with a data rate of 300 Mbpsin
`20 MHz and 600 Mbps in 40 MHz. Furthermore, in the 5 GHz band, 802.1 1ac provides a
`data rate of 1733 Mbps with 80 MHz and fourspatial streams, and a maximumdatarate
`of 6933 Mbps with 160 MHz and eight spatial streams. The exponential increase in data
`rate is illustrated in Figure 1.2.
`
`Table 1.2 Overview of 802.11 PHYs
`
`
`802.11
`
`B802.11b
`
`802.1la
`
`802.11g
`
`802, 11n
`
`802.1 lac
`
`DSSS
`
`L2
`
`2.4
`
`PHY
`technology
`Data rates
`(Mbps)
`Frequency
`band (GHz)
`20, 40, 80, and 160
`20 and 40
`25 MHz
`20
`25
`25
`Channel spacing
`(MHz)
`
`DSSS/
`CEs
`35,11
`
`OFDM OFDM DSSS/
`CCK
`1-54
`
`6-54
`
`SDM/OFDM SDM/OFDM
`MU-MIMO
`6.5-6933.3
`
`6,.5-600
`
`2.4
`
`5
`
`2.4
`
`2.4 and 5
`
`3
`
`TP Link Ex. 1028
`Page 7
`
`TP Link Ex. 1028
`Page 7
`
`

`

`6
`
`5 eee
`
`Introduction
`= ed a oo Se eea a ee —__
`
`@ 20/25 MHz
`
`» 40 MHz
`B 80 MHz
`
`@ 160 MHz
`
`10000
`
`see
`
`100
`
`10
`
`1
`
`-
`
`
`
`w
`
`a=o-
`
`fc
`
`S
`
`dot11 (2.4 GHz) 11b (2.4GHz)
`
`11a(5 GHz)/ 11n (2.4/5 GHz)
`11g (2.4 GHz)
`
`11ac; 4ss
`(5 GHz)
`
`lilac; 8ss
`(5 GHz)
`
`Figure 1.2
`
`Increase in 802.11 PHY datarate.
`
`1.1.3
`
`The 802.11 network architecture
`
`Thebasic service set (BSS) is the basic building block of an 802.11 LAN. Stations that
`remain within a certain coverage area and form somesort ofassociation form a BSS, The
`most basic form of association is where stations communicate directly with one another
`in an ad-hoc network, referred to as an independent BSS orIBSS. Thisisillustrated as
`BSS | in Figure 1.3,
`Moretypically, however, stations associate with a central station dedicated to manag-
`ing the BSS and referred to as an access point (AP). A BSS built around an APis called an
`infrastructure BSS andis illustrated by BSS 2 and BSS 3 in Figure 1.3. Infrastructure
`BSSs may beinterconnectedvia their APs through a distribution system (DS).
`The BSSsinterconnected by a DS form an extended service set (ESS). A key concept
`of the ESSis that stations within the ESS can address each otherdirectly at the MAC
`layer, The ESS, being an 802.11 concept, encompasses only the 802.11 devices and does
`not dictate the nature of the DS. In practice, however, the DS is typically an Ethernet
`(802.3) LAN and the AP functions as an Ethemet bridge. As such, stations ina BSS can
`also directly address stations on the LAN at theMAC layer.
`
`1.1.4
`
`Wi-Fi Direct
`
`Recognizing the need for improved peer-to-peer operation, the Wi-Fi Alliance has
`developed a specification for direct communication between Wi-Fi devices without
`being associated with an infrastructure BSS. Such communication is possible using an
`
`TP Link Ex. 1028
`Page 8
`
`TP Link Ex. 1028
`Page 8
`
`

`

`a == Se — 1.2 History of high throughput and 802.11n
`
`7
`
`(ad-hoc)
`
`BSS 1
`
`Figure 1.3
`
`BSS, DS, and ESS concepts.
`
`independent BSS, as defined in the 802.11 specification; however, it was preferable to
`create a mode of operation closer to that of the infrastructure BSS.
`In a Wi-Fi Direct network, one device, called the group owner (GO), assumesa role
`similar to that of an AP while the other devices associate with that device as they would
`an AP. The Wi-Fi Direct network is thus similar to an infrastructure BSS except that (1)
`the GO does not provide access to a distribution system, and (2) like its peers, the GO
`could be a mobile, battery powered device, and thus also need to enter a low powersleep
`state whenidle.
`The Wi-Fi Direct standard builds on the 802.11 specification, specifying protocols by
`which devices can discovereach other, how a device assumestherole ofgroup owner and
`the protocol for absence from the session channel (for power managementorto visit an
`infrastructure BSS channel).
`
`1.2
`
`1.2.1
`
`History of high throughput and 802.11n
`
`The High Throughput Study Group
`
`Interest in a high data rate extension to 802.11a began with a presentationto the Wireless
`Next Generation Standing Committee (WNG SC) of IEEE 802.11 in January 2002.
`Marketdrivers were outlined, such as increasing data rates of wired Ethernet, more data
`rate intensive applications, non-standard 100+ Mbps products entering the market, and
`the needfor higher capacity WLAN networks (Jones, 2002). The presentation mentioned
`techniques such as spatial multiplexing and doubling the bandwidth as potential
`approachesto study for increasing data rate.
`
`TP Link Ex. 1028
`Page 9
`
`TP Link Ex. 1028
`Page 9
`
`

`

`8
`
`Introduction
`
`After many additional presentations, the High Throughput Study Group (HTSG) was
`formed withits first meeting in September 2002. The primary objective of HTSG wasto
`complete two documents necessary for the creation of the High Throughput Task Group
`(TGn). These are the project authorization request (PAR) form andfivecriteria form. The
`PAR defined the scope and purpose ofthe task group as follows:
`
`The scopeofthis project is io define an amendmentthat shall define standardized modificationsto
`both the 802.11 physical layers (PHY) and the 802.11 medium access control layer (MAC) so that
`modes ofoperation can be enabled that are capable afmuch higherthroughputs, with a maximum
`throughput ofat least 100 Mbps, as measured at the MAC data service access point (SAP), IEEE
`(2006)
`
`Bythis statement, the standard amendment developed by TGn must contain modes of
`operation that are capable of achieving at least 100 Mbps throughput. Throughputis the
`measure of “useful” information delivered by the system and by using throughputas the
`metric, both MAC and PHY overhead must be considered. 802.1 1a/g systems typically
`achieve a maximum throughput ofaround 25 Mbps;thus this statementrequiredatleast a
`four-fold increase in throughput. Meeting this requirement would in essence mandate
`PHY data rates well in excess of 100 Mbpsas well as significant enhancements to MAC
`efficiency.
`Additional explanatory notes were included with the PAR outlining many evaluation
`metrics. These include throughput at the MAC SAP,range, aggregate network capacity,
`powerconsumption, spectral flexibility, cost complexity flexibility, backward compati-
`bility, and coexistence (IEEE, 2006),
`The five criteria form requires that the study group demonstrate the necessity of
`creating an amendmentto the standard. Thecriteria include (1) broad market potential,
`(2) compatibility with existing IEEE 802.1 architecture, (3) distinct identity from other
`IEEE 802 standards, (4) technical feasibility, and (5) economic feasibility (Rosdahl,
`2003). The goalis to create a standard amendmentwhichresults in marketable products,
`but that will also be differentiated from other potentially similar products.
`In addition to completing the PAR andfive criteria forms, HTSG also began develop-
`ment of new multipath fading MIMO channel models (Erceg et al., 2004) and usage
`models (Stephens et a/., 2004), The channel models and usage models were used to
`create a common framework for simulations by different participants in the standard
`development process.
`
`122
`
`Formation of the High Throughput Task Group (TGn)
`
`The PAR was accepted and approved by the 802 working group, creating Task Group
`n (TGn) with the first meeting of the task group held in September 2003. The standard
`amendment developed by the task group would be proposal driven, meaning that
`members of the task group would make partial or complete technical proposals, with
`the complete proposals proceeding through a down-selection process culminating in a
`single proposal upon whichthe standard amendment would be based. Partial proposals
`would be informative and could be incorporated in a complete proposal along the way, To
`
`TP Link Ex. 1028
`Page 10
`
`TP Link Ex. 1028
`Page 10
`
`

`

`1.2. History of high throughput and 802.11n
`
`i]
`
`that end, the task group began development ofthe functional requirements (Stephens,
`2005) and comparisoncriteria (Stephens, 2004) documents. These two documents would
`provide, respectively,
`the technical requirements complete proposals must meet and
`criteria by which complete proposals would be compared.
`The task group began with nine functional requirements. One of the functional
`requirements was a catch-all, requiring that proposals meet the PAR and five criteria. A
`second requirement wasareiteration of the PAR requirement to achieve 100 Mbps
`throughput at the top of the MAC.
`Furthermore, since it was expected that notall regulatory domains would allow a single
`device to use multiple 20 MHz channels (an easy wayto achieve the throughputobjective),
`the second requirement addedarestriction that 100 Mbps throughput be achieved in a
`single 20 MHz channel, To enforce efficient use of spectrum, another requirement was
`added for a mode of operation with a spectral efficiency of at least 3 bps/Hz.
`Four functional requirements addressed operational bands and backward compatibil-
`ity. One of these requirements was that the protocol should support operation in the
`5 GHz band dueto the large availability of spectrumthere. Another requirement was that
`at
`least some modes of operation be backward compatible with 802.1la systems.
`Noteworthy was the fact that there was no requirement to support operation in the
`2.4 GHz band. However, if a proposal did support 2.4 GHz band operation,
`it was
`required that there be modes ofoperation that were backward compatible with 802.11¢
`systems. In this context, some flexibility was given, allowing an 802.11n AP to be
`configured to accept orreject associations from legacy stations.
`The 802.1le amendment to the standard, nearing completion at the time, added
`manyfeatures for improving the quality of service (QoS) in 802.11 systems. Many of
`the perceived applications for 802.11n involved real time voice and video which
`necessitate QoS. Therefore a functional requirement was included which mandated
`that a proposal allow for the implementation of 802.1le features within an 802.11n
`station,
`The comparisoncriteria in Stephens (2004) outlined metrics and required disclosure
`of results which would allow for comparison between proposals under the same
`simulation setup and assumptions. The comparisoncriteria incorporated the simulation
`scenarios and usage models defined in Stephens etal. (2004). During the development
`of the comparison criteria, the task group realized that members of the task group did
`not always share the same definitions for common terms. Therefore definitions for
`goodput, backward compatibility, and signal-to-noise ratio (SNR) were provided. The
`comparison criteria covered four main categories: marketability, backward compati-
`bility and coexistence with legacy devices, MAC related criteria, and PHY related
`criteria.
`Under marketability, the proposal must provide goodputresults for residential, enter-
`prise, and hotspot simulation scenarios. Goodputis defined by totaling the numberofbits
`in the MACservice data units (MSDU)indicated at the MACservice access point (SAP),
`and dividing by the simulation duration (Stephens, 2004). Two optional criteria included
`describing the PHY and MAC complexity. The PHY complexity was to be given relative
`to 802.1 1a.
`
`TP Link Ex. 1028
`Page 11
`
`TP Link Ex. 1028
`Page 11
`
`

`

`10
`
`Introduction
`
`To ensure backward compatibility and coexistence with legacy devices, a proposal
`was required to provide a summary of the means used to achieve backward compatibility
`with 802.11a and, if operating in 2.4 GHz, 802.11g. Simulation results demonstrating
`interoperability were also required. The goodput of a legacy device in an 802.11n
`network and the impact of a legacy device on the goodput of 802.11n devices were
`also to be reported.
`The MAC related criteria included performance measurements and changes that
`were made to the MAC.In the residential, enterprise, and hotspot simulation scenarios a
`numberofdifferent metrics were to be captured and reported. These included the ability to
`support the service requirements of various applications,
`including QoS requirements.
`Measurements of aggregate goodput of the entire simulation scenario were required to
`indicate network capacity. MAC efficiency was to be provided, whichis defined as the
`aggregate goodputdivided by the average PHY data rate. To ensure reasonable range for
`the new modesofoperation, throughput versus range curves were also to be provided.
`The PHY related criteria included PHY rates and preambles, channelization, spectral
`efficiency, PHY performance, and PHY changes. In addition, the comparison criteria also
`defined PHY impairments to be used in combination with channel models for PHY
`simulations. Each proposal was required to generate simulation results for both additive
`white Gaussian noise (AWGN) and non-AWGN channels, Furthermore, simulation
`conditions to analyze the impact on packeterror rate (PER)of carrier frequency offset
`and symbol! clock offset were also defined.
`
`1.2.3
`
`Call for proposals
`
`The TGn call for proposals was issued on May 17, 2004, with the first proposals
`presented in September 2004. Over the course of the process two main proposal teams
`emerged, TGn Syne and WWiSE (world wide spectral efficiency). The TGn Sync
`proposal team was founded by Intel, Cisco, Agere, and Sony with the objective of
`covering the broad range of markets these companies were involvedin, including the
`personal computer (PC), enterprise, and consumerelectronics markets. The WWiSE
`proposal
`team was formed by Airgo Networks, Broadcom, Conexant, and Texas
`Instruments. These semiconductor companies were interested in a simple upgrade to
`802.1la for fast time to market. Many other companies were involved in the proposal
`process and most ended up joining one of these two proposal teams.
`The key features of all the proposals were similar, including spatial division multi-
`plexing and 40 MHz channels for increased data rate, and frame aggregation for
`improved MACefficiency. The proposals differed in scope (TGn Sync proposed numer-
`ous minor improvements to the MAC while WWiSE proposed limiting changes) and
`support for advanced features such as transmit beamforming(initially absent from the
`WwWIiSE proposal).
`A series of proposal down-selection and confirmation votes took place between
`September 2004 and May 2005. During that time, mergers between proposals and
`enhancements to proposals took place. The TGn Syne proposal won the final down-
`selection vote between it and WWiSE, butfailed the confirmation vote in May 2005.
`
`TP Link Ex. 1028
`Page 12
`
`TP Link Ex. 1028
`Page 12
`
`

`

`=
`
`1.2 History of high throughput and 802.11n
`= i —
`=
`— ee
`
`11
`
`1.2.4
`
`Handheld devices
`
`During this period interest arose in a new emerging market of converged Wi-Fi and
`mobile handsets, The shipment of dual mode Wi-Fi/cellular handsets had grownsignifi-
`cantly from 2005 to 2006. Someparticipants in the proposal process believed that
`handsets would be the dominant Wi-Fi platform within a few years (de Courville et al.,
`2005), At the time, converged mobile devices were projected to grow worldwide at a
`compound annual growth rate of 30% (IDC, 2007).
`A contentious issue for handheld proponents was the high throughput requirement
`for 100 Mbps throughput. This,
`in essence, would force all 802.11n devices to
`have multiple antennas. This is a difficult requirement for converged mobile devices,
`since they already contain radios and antennas for cellular 2G, 3G, Bluetooth, and
`in some cases GPS. Concern was raised that mandating 802.11n devices to have
`multiple antennas would force handset manufacturers to continue to incorporate
`single antenna 802,11la/g into handsets and not upgrade to 802.11n. Not only does
`this diminish the capabilities of the handset device, it burdensall future 802.11n deploy-
`ments with continued coexistence with 802.1la/g embedded in these new handset
`devices.
`For this reason an ad-hoc group was formed to create functional requirements
`supporting single antenna devices. Two new requirements were added to the functional
`requirements document in July 2005. The first requirement mandated that a proposal
`define single antenna modes of operation supporting at least 50 Mbps throughputin a
`20 MHz channel. The second requirement dictated that an 802.11n AP orstation
`interoperate with client devices that comply with 802.11n requirements but incorporate
`only a single antenna. This requirement resulted in 802.11n making mandatory atleast
`two antennas in an AP, but only one antenna in a non-AP device.
`
`1.2.5
`
`Merging of proposals
`
`After the failed confirmation vote, a joint proposal effort was started within the task group
`to merge the two competing proposals. Due to entrenched positions and the large
`membership of the group, the joint proposal effort proceeded very slowly, As a result,
`Intel and Broadcom formed the Enhanced Wireless Consortium (EWC) in October 2005
`to produce a specification outside the IEEE that would bring products to market faster.
`With muchofthe task group membership ultimately joining the EWC,this effort had the
`effect of breaking the deadlock within the IEEE, and the EWCspecification, which was
`essentially a merger of the TGn Syne and WWiSE proposals, was adopted as the joint
`proposal and submitted for confirmation to TGn where it was unanimously passed in
`January 2006.
`
`1.2.6
`
`802.11n amendmentdrafts
`
`The joint proposal was converted to a draft 802.11 standard amendment for higher
`throughput (TGn Draft 1.0), and entered letter ballot.
`In letter ballot, IEEE 802.11
`
`TP Link Ex. 1028
`Page 13
`
`TP Link Ex. 1028
`Page 13
`
`

`

`12
`
`Introduction
`
`working group members(not just task group members) vote to either adopt the draft asis
`or reject it with comments detailing changes needed. The draft requires at least a 75%
`affirmative vote within the IEEE 802.11 working group in order to proceed to sponsor
`ballot whereit is considered for adoption by the broader IEEE standardsassociation. TGn
`Draft 1.0 entered letter ballot in March 2006 and,not unusually, failed to achieve the 75%
`threshold for adoption, Comment resolution began May 2006 on the roughly 6000
`unique technical and editorial comments submitted along with the votes.
`With resolution of the TGn Draft 1.0 comments, TGn Draft 2.0 went out forletter
`ballot vote in February 2007and this time passed with 83% ofthe votes. However, there
`were still 3000 unique technical and editorial comments accompanyingtheletter ballot
`votes. It is typical for the task group to continue comment resolution until a minimum
`number of negative votes are received; thus comment resolution for TGn Draft 2.0
`continued between March 2007 and September 2007, resul