IEEE 802.11 System Design

  • Upload
    abin444

  • View
    228

  • Download
    0

Embed Size (px)

Citation preview

  • 8/7/2019 IEEE 802.11 System Design

    1/5

    490

    IEEE 802.11 SystemDesignNeeli R. Prasad

    Lucent TechnologiesZadelstede 1-10,P. 0.Box 755

    3430 AT Nieuwegein, The NetherlandsEmail: nprasad8lucent.com

    ABSTRACT: The present generation of the Wireless LAN(WLAN) products are implemented on Personal ComputerMemory Card International Association (PCMCIA) cards(also called PC card) that are used in laptops, computers andportable devices. The technical issues for WLA N systems aresize, power consumption, bit rate, aggregate throughput,coverage range and interference robustness. Scarcity ofspectrum is the biggest issue in wireless communication. Thechallenge is to serve the largest number of users with aspecified system quality. For this purpose n etwork architectureand study thereofQlay a very impo rtant role.This paper first discusses the considered WLAN, after whichexplanations of WLAN system design taking intoconsideration the critical issues such as Coverage, Cellplanning, Interference, Power management and Securityespecially for IEEE 802.1 1 WL AN based on Direct S equenceSpread Spectrum (DSSS) in 2.4 GH z ISM band.

    1. INTRODUCTIONThe tremendous growth of wireless communications andpenetration of Internet has brought about major changes in thefield of Local Area Networks. Benefits of wirelesscommunications and Internet offer gains in efficiency,accuracy and lower business costs. These benefits broughtforward growth in the market of Wireless Local AreaNetworks (WLANs) which in turn brought proprietarystandards in the market, this chaos was resolved byharmonizing effort of IEEE with an international standard onWLAN s: IEEE 802.11 [l], ,The standard specifies Medium Access Control (MAC) [2]and Physical Layer (PHY) for wireless connectivity for fixed,portable and moving stations within a local area. IEEE 802.1 1operates using either radio technology or infrared techniques.Each approach has it own attribute, which satisfies differentconnectivity requirements. Majority of these devices arecapable of transmitting information up to several 100 metersin an open environment. In Figure 1, a concept of WLANinterfacing with a wired network is given. The components ofWLANs consist of a wireless network interface card, oftenknown as station, STA, and a wireless bridge referred to asaccess point, AP. The AP interface the wireless network withthe wired network (e.g. Ethernet LAN) [ l , 3,4] .WLANs using radio waves work at the ISM (industrial,scientific and medical) frequency band of 2.4 GHz. Th e releaseof the ISM band meant the availability of unlicensed spectrumand prompted significant interest in the design of WLANs. Anadvantage of radio waves is that they can provide connectivityfor non line-of-sight situations also. A disadvantage of radiowaves is the electromagnetic propagation, which might causeinterference with equipment working at the same frequency.

    Security might also be a problem, because radio wavespropagate through the w alls.

    STA: StationAP: Access PointFigure 1 A wireless local area networ k.,

    WLANs based on radio waves usually use spread spectrumtechnology [3 , 5, 61. Spread spectrum spreads the signal powerover a wide band of frequencies, which makes the data muchless susceptible to electrical noise than conventional radiomodulation techniques. Spread spectrum modulators use one ofthe two methods to spread the signal over a wider area:frequency hopping spread spectrum,FHSS, r direct sequencespread spectrum, DSSS.Infrared LANs working at 820 nm wavelength provide analternative to radio wave based WLANs. Although infrared hasits benefits it is not suitable for mobile applications due to itsline-of-sight requirement. There are two kinds of infraredLANs: diffused and point to point.Th e basic IEEE 802.11 and the differences between DSSS an dFHS S are discussed in Section 2. Th e WLAN considered in thispaper is given in Section 3. While its system design is given inSection 4. caleable radio system aspects is described in Sec tion5. Finally, conclusions are given in Section 6.

    2. IEEE 802.11IEEE 802.1 I is a standard for wireless system s that operates inthe 2.4 - 2.5 GHz ISM (industrial, scientific and medical)band. This ISM band is available worldwide and allowsunlicensed operation for spread spectrum systems. Table 1lists the available frequency bands and the restrictions todevices, which use this band for com munications [4, 7, 81. The802.11 standard focuses on the M AC (m edium access control)and PHY (physical layer) protocols for access point basednetworks and ad-hoc networks.Th e 802.1 1 standard supports DSSS with differential encodedBPSK and QPSK, FHSS with GFSK (Gaussian FSK), andinfrared with Pulse Position Modulation (PPM).

    0-7803-5893-7/00/$10.002000EEE

    Authorized licensed use limited to: Amal Jyothi College of Engineering. Downloaded on July 07,2010 at 13:37:42 UTC from IEEE Xplore. Restrictions apply.

    http://nprasad8lucent.com/http://nprasad8lucent.com/
  • 8/7/2019 IEEE 802.11 System Design

    2/5

    491

    EuropeJapanFrance and Spain

    Table 1 Frequency BandsI Regulatory Range I Maximum Outwt ILocation

    2.400-2.4835 100mW (EIRP*)2.400-2.4835 10mWA4Hz2.400-2.83S2.475 100mW (EIRP)

    -I (GHz) I PowerNorth America I 2.400-2.4835 I 1000mw

    Busy Medium

    CantentionWindow-/ pircJcqff-Window / y "I rame

    Figure 2 Basic CSMNCA behavior.IEEE 802.11 DSSS: DSSS systems spread the signal energyacross a relatively wide band by increasing the occupiedbandwidth. A DSSS transmitter converts a bit stream intosymbol stream where each symbol represents a number of bitsdepending on the (PSK) modulation technique. The symbolinformation is converted into a compiex-valued signal, which isfed to the spreader. This spreader multiplies its input signal witha PN (pseudo noise) sequence, which is called a chip sequence.The result of this multiplication is a signal with a widerbandwidth. The in-phase and quadrature components of thespreader output signal are fed to a quadrature modulator. The

    transmitter front-end provides filtering, upmixing and poweramplification.The 802.11 DSSS is based on 1 1-chips Barker sequence. The 1 1-chip spreading makes the occupied bandwidth larger andincreases the effective bandwidth fiom 1 MHz to 11 MHz. The802.11 standard specify bit rates: 1 Mbit/s with BPSK, 2 M bit/swith QPSK, 5.5 Mbit/s and 11 Mbit/s with Complementary CodeKey (CCK).ZEEE 802.11 FHSS: FHSS systems hop from narrow band tonarrow band within a wide band. FHSS wireless LAN stationssend one or m ore data packets at one carrier frequency, hop toanother carrier frequency and send one or more packets and.continue this hop-transmit sequence (slow frequency hopping).The time these FHSS radios dwell on each frequency is fixed.The hopping pattern appears random but is actually, a periodicsequence tracked by sender and receiver.The 802.11 standard defines hops over channel centerfrequencies according to a periodic sequence (but it looks like arandom pattern).D S S S vs. FHSS: The 802.1 1 standard DSSS and FHSS systemshave been compared with respect to various performanceaspects. DSSS has a more robust modulation and gives a largercoverage range than FHSS even when FHSS uses twice thetransmitter power output level. FHSS gives a large number ofhop frequencies, but the adjacent channel interference behaviorlimits the number of independently operating collocatedsystems. FHSS has more transmission time overhead introducedby hop time and smaller packet size, which impacts themaximum throughput. FHSS is less robust, but it gives a moregraceful degradation in throughput and connectivity. Under poorchannel and interference conditions FHSS will continue to workover a few ho p channels a little longer than over the other hopchannels. However, for a distance at which only a very few ofall FHSS hop channels still work, DSSS still gives reliable links.For collocated networks (access points), DSSS gives, with feweraccess points, a higher potential throughput than FHSS withmore access points. This means that with DSSS, a smallernumber of access points can be used resulting in lowerinfrastructure cost.

    3. CONSIDERED WIRELESS LANIn this section WLANs based on PS SS technology as given byIEEE 802.11 is considered. The IEEE 802.11 WLAN based onDSSS is initially aimed for the 2.4 GHz band designated forISM applications as provided by the regulatory bodies worldwide [l , 3,4].Th e DSSS system provides a WLAN with 1 Mbit/s, 2 Mbit/s,5.5 Mbit/s and 11 Mbit/s data payload communicationcapability. Accord,ing to the FCC regulations, the DSSSsystem shall provide a processing gain of at least 10 dB. Thisshall be accomplished by chipping the baseband signal at 11MH z with 11-chip pseudo rand om,' PN, co de (Barkersequence).Th e DSSS system uses baseband modulations of differentialbinary phase shift keying (D BPSK) and differential quadraturephase shift keying (DQPSK) to provide the 1 and 2 Mbit/sdata rates, respectively. Complementary code keying (CCK) isused to provide 5.5 and 11 Mbit/s.

    4. SYSTEMS DESIGNThe architecture and system aspects of WLAN are discussedin this section. As systems design can vary, we will

    Authorized licensed use limited to: Amal Jyothi College of Engineering. Downloaded on July 07,2010 at 13:37:42 UTC f rom IEEE Xplore. Restrictions apply.

  • 8/7/2019 IEEE 802.11 System Design

    3/5

    492

    concentrate on the system design of Lucent Technologies802.11 complaint WLAN systems: ORiNOCO (formerlyknown a s WaveLAiV) [ 3 , 4 , 9 , 10, 1 1 1 .4.1. Network TopologiesORiNO CO can support different topologies.Stand-Alone Hub U N

    LAN Extension

    Multi-Cellular Infrastructure---+Ys-t?

    This can be a single-channel infrastructure (all cells on thesame channel) or a multiple-channel infrastructure. ORiNO COAccess Points (AP) support roaming over multi-channels.Point-to-Point Link

    Point-to-point Link with Directional Antenna

    1 I I

    4.2. CoverageThe reliable coverage analysis is based on path loss modelingfor environments like Open Plan B uilding, Semi-open Office,Closed Office with respectively path loss coefficients of 2.2,3 .3 and 4.5 above the 5 meter breakpoint (up to 5 meter freespace propagation with path loss coefficient equal to 2) ( s ~Table 2). On top of this modeling with path loss dependent onthe TX-RX (Transm it-Receive) distance there will be a marginof 10 dB required in relation to variation due to fading. Withtwo antennas and a Rayleigh fading channel the 10dB.marginreflects a reliability of 99% .

    Table2 Reliable ranges according to path loss models.~~ 1 2 5.5 11Receiver sensitivity -93 -90 -87 -84for BER (dBm)Range covered 99 % point TX ower 15 dB mOpen Pian Building 485 m 354 m 259 m 189 m1 SemiOpenOff ice 105m 8 5 m 6 9 m 5 6 m Ilosed Office 4 6 m 4 0 m 3 4 m 2 9 m

    The reliable coverage range might be influenced by multipathwhen operating at 11 Mbit/s and 5.5 Mbit/s. (in larger openspaces) and by the pres ence of co ncrete walls at all bit rates. InFigure 3 measured throughput results for different receive levelare given.

    Figure3 Throughput against receive level for 802.11PC card withARF.4.3. Cell planningThere are two basic requirements when designing a WLANnetwork throughput and coverage. On one hand a networkca n be designed and deployed with the primary requirementbeing coverage. Such network will have low aggregratethroughput and larger cells. Such network deployment willrequire lesser APs and thus w ill be cheap er. On the other handthe primary requirement can be throughput. This meanssmaller cell size (still giving full coverage), requiring moreAP s and thus more expensive.Adjacent cells can be used independently if the centerfrequency are spaced '15 MH z apart. With this requirement afrequency planning can be made fo r each of the differentworld regions having its own 2.4 - 2.5 GHz ISM bandrestrictions, Table 1.The channel reuse distance can be found using the path lossmode l and .thereceiver sensitivity. If fully o verlapped cells are,used center frequencies spaced 25 MH z apart must be used. Ofcourse besides all this the interference sources must beisolated.

    Authorized licensed use limited to: Amal Jyothi College of Engineering. Downloaded on July 07,2010 at 13:37:42 UTC from IEEE Xplore. Restrictions apply.

  • 8/7/2019 IEEE 802.11 System Design

    4/5

    493

    4.4. InterferenceRadio frequency inerference is one of the most importantissues to be addressed in the design, operation andmaintenance of w ireless communications systems. Althoughboth. inter-modulation and inter-s ymbo l interference alsoconstitute problems to account for in system plahning, awireless radio system designer is mostly concerned withadjacent channel and co-channel interference.CO-channel interference lies within the b andw idth of thevictim receiver and arises principally from the transmittersusing the same band. Adjacent channel interfer.ence arisesfrom the same sources and causes problems because thereceiver filters do not have perfec t selectivity.4.5. Power managementPower Management is aimed at the reduction of the averagepower consumption of a ORiN OCO Network Interface Card(NIC). This is particularly important for NICs operated inbattery operated poitable stations (STA).The 802.11 standard defines power management protocols thatcan be used by STAs. Power management schemes result in alower consumption of (battery) power compared to traditionaloperation where a STA is alwaJs monitoring the mediumduring idle periods. To achieve savings in power consumption,a NIC in a STA must have a special low power state ofoperation called DOZE state. In this state the NIC will notmonitor the medium and will be unable to receive a frame.This state differs from the OFF state in the sense that the cardIflust be able to make a transition from DOZE state to fullyoperational receive (AWAKE) state in a very short time (250ps). A transition from OFF o AWAKE state will take muchmore time.Power Management allows a STA to spend most of its idletime in DOZE state, while still maintaining connection to therest of the nctvvork to receive unsolicited messages. For thelatter requirement, the other STA s or the AP must temporarilybuffer the messages that are destined to a STA operating in apower management scheme, and such a ST A must wakeupon regular intervals to check if there are messa ges buffered forit.Ineach frame that is transm itted by a station, there is one so-called PM bit to indicate the mode of operation. PM-bit = PSindicates that the station is operating in the Power Savingmode, and PM-bit = A indicates that the station is in theActive mod e (continu ously active). In an access point basednetwork the access points will learn via the PM-bit whether astation is in Active or Power Saving mode and will do thebuffering of messages for such station when required. Theaccess point will also buffer multicast messages. Access poin tswill send beacon frames on a regular basis (e.g. every 100ms). In each beacon frame the access point will announce forwhich stations it has messages buffered. The stations using apower management scheme will wake up just prior to abeacon transmission with high accuracy.To support the power management schemes of 802.11, thefollowing queuing structure can be implemented in an accesspoint (Figure 4).

    . ,I FrameTransfer IFigure 4 802.11 power management scheme for an access point.

    4.6. SecurityWLANs compliant to IEEE 802.11 combats the securityproblem with open system and shared key authentication andRC4 based encryption [I , 3, 4, 9, 121. Open systemauthentica tion is essentially a null authentication in which anySTA is authen ticated by the AP. Shared key auth enticationsupports authentica tion of STAs as either a m ember of thosewho know a shared secret key or a member of those who donot. IEEE 802.1 1 shared key authentication accomplishes thiswithout the need to transmit the secret key in the clear;requiring the use of the wired equivalent privacy, WEP,mechanism. Closed system authentication, a proprietaryschem e, is implemented in ORINOCO , which provides furthersecurity. WLAN is envisaged to be used in corporate andpublic environments, the existing level of security will not beenough for these environments.In general a corporate environment has a Ethernet based LANwith OS related authentication procedure (Microsoft, Apple,Unix etc.). Such corporate environment is referred asenterprise environment. Enterprises have closed networkenvironment where reasonable security can be achieved byusing network name and shared key authentication.Reasonable because shared key and network name basedauthentication are not very secure processes. Another majorconcern in Enterp rises is the rate of chan ge in perso nnel, bothshort and long term.-Distributing keys to them and makingsure they can not misuse a key once they have left thecompany is a majcr managerial problem.Besides enterprises there are academic and other institutionswhere either OS based authen tication is used or in certaincases Kerberos is used. Kerberos is Uoix based; it includesauthentication, access control and session encryption. Theauthentication is decoupled from access control so thatresource owneis can decide who has access to their resources,In this sense, Kerberos meets the managerial needs givenabove. For such institutions, the WL AN system must becompatible to Kerberos with the wireless part giving the samelevel of security as Kerberos.The pu blic or dial in environm ent users make use of un trustedcommunications facilities to access systems of their employeror an internet service provider, ISP. Therefore bothauthentication and session security are needed. This

    ~

    Authorized licensed use limited to: Amal Jyothi College of Engineering. Downloaded on July 07,2010 at 13:37:42 UTC from IEEE Xplore. Restrictions apply.

  • 8/7/2019 IEEE 802.11 System Design

    5/5

    494

    environment is dominated by Microsoft platforms. Operatorsand service providers frequently use RADIUS (remoteauthentica tion dial In user service).RADIUS service s are usedespecially when people are mobile and require access to theircorporate network br when peop le want to access a ISP fromhome. WLAN working in such environment will requirecompatibility to RADIUS and extra security for the w irelessPart. 5. S C A L E A B L E R A D I O SYSTEMThe O RiNOCO systems are deployable in a broad variety ofsituations, posing different and often conflicting requirementson the system behavior. In particular, the situation of a stand-alone single cell network versus an infrastructure networkconsisting of multiple overlapping cells will have differentrequirements on the transmit and receive behavior. Toaccommodate these varying operational situations, theORiNOCO products have a number of built-in provisions tocreate scaleable systems, optimized for environment andnetwork usage needs. Figure 5l lus tra tes the typ ica l c w e orthe signal level in two opposite directionsbasis for the scaleability control elements. This curve is the

    I.--

    c----L

    Figure 5 Carrier detect threshold (CD T) impact on cell size.The 802.11 medium access rules (CSMNCA) are based ondefer and random backoff behavior of all stations within rangeof each other. The defer decision is based on a configurationentity called defer threshoid (DT). W hen a carrier signal levelis observed above the D T level, the O RiNO CO card holds upa pending transmission request.-

    . n s h o m l s

    Figure6 Ideal relation between defer threshold (DT ) and CDT.Th e range for the DT level has a lower boundary determinedby the sensitivity of the carrier detect circuitry. Below acertain level, the signal will not be detected and no defer willbe done. The ideal relation shown in Figure 6 cannot beachieved in the case where the CD T is set to the lowest (and

    most sensitive) level. In that case the lowest meaningful DTwill not guarantee the wanted deferral between two edgestations. This is illustrated in Figure 7.

    c--c-lnm*cFigure 7 Large cell characteristics

    6. CONCLUSIONSA key aspect of the ORiN OCO system design is the suitabilityof the radio to be upgraded to the higher data rates 5.5 and 11Mbit/s without major redesign. This allows for the class .ofhigh speed radio LANs that are fully co-existent with stationsfollowing the IEE E 802.1 1 standard, and that are interoperablewith such stations via the access points.

    REFERENCES:IEEE, 802.11, Wireless LAN Medium Access Control (MAC)and Physical Layer (PHY) specificatio ns,November 1997.K.C. Chen, Medium Access Control of Wireless LANs forMobile Computing, IEEE Network, September/October 1994,B.Tuch, Development of WaveLA N, an ISM Wireless LAN,AT&T Technical Joumal, vol. 72, w .4, July/August 1993, pp.A. Kamerman and L. Monteban, WaveLAN-11: A High-Performance Wireless LAN for the Un licensed Band, Bell LabsTechnicalJoumal, vol. 2, no. 3,1997, pp. 118-133.A. Kamennan, Spread Spectrum Schemes for Microwave-Frequency WLANs, Microwave Joumal, vol. 40, no. 2,February 199 7, pp. 80-90.K. Pahlavan and A.H. Levesque, Wireless InformationNetworks,Wiley, 1995, ISBN 0-471 -1067-0R. van Nee, G. Awater, M. Morikura, H. Takanashi, M. Websterand K. Halford, New High Rate Wireless LAN Standards, tobe published in IEEEComm unications Magazine, Dec. 1999.Richard vm Ne e and Ramjee Prasad, OFDM for MobileMultimedia Communications, Artech House, Boston-London,1999.httD://www.lucent.com/orinoco/

    pp. 50-63.

    27-37.

    [IO] A.R. rasad,*N.R. Prasad, A. Kam erman,H.Moelard and A.Eikelenboom , Indoor WirelessLANs Deploym ent, VTC 2000Spring , Tokyo, Japan, 1 5-18 May 2000.[ ll ] A.R. Prasad, A. Eikelenboom, H.Moelard, A. Kamerman &N.R. Prasad, Wireless LANs Deployment in Practice , chapterof Digital. Network Deploym ents: Issues And Chnllenges,editedby R. Ganesh and K. Pahelvan, Khwe r Publications,2000.

    [I21 A.R. Prasad, H. Moelard and J. Kruys, Security Architecturefor WirelessLANs: orporateL ublic Environment,VTC2000 Spring, Tokyo, Japan, 15-18 May 2OOO

    Authorized licensed use limited to: Amal Jyothi College of Engineering Downloaded on July 07 2010 at 13:37:42 UTC from IEEE Xplore Restrictions apply