Voice IP

  • Upload
    jbond07

  • View
    219

  • Download
    0

Embed Size (px)

Citation preview

  • 7/28/2019 Voice IP

    1/8

  • 7/28/2019 Voice IP

    2/8

    WIRELESS NETWORKS, CHALMERS 2011 2

    Fig. 1. VoIP over WLAN [3]

    made for real time application because the QoS which is

    big issue with WLAN is important and main portion of

    VoIP applications. Challenges VoWLAN: there are many

    challenge in VoWLAN like Quality of Services (QoS),

    security etc QoS in VoWLAN consist of these three thing

    which is discussed below.

    A. Packet Loss

    The total number of packet transmit over the network

    is not receive to the end point or destination, so it means

    the some data or packet loos or not received by the desti-

    nation. There are two main sources of packet losses:one

    is network packet losses, mainly due to network conges-

    tion (router buffer overflow), link failures and rerouting,

    transmission errors, etc; and the other is discarded packet

    losses for packets experienced excessive delay.

    B. Delay

    The time taken by a packet to reach from a source

    to destination, delay can be occurred from different

    sources like delay at source, delay at receiver, delay

    in network. Delay at source and receiver is due to

    coding like changing analog to digital and digital to

    analog and packetization, while network delay is due

    to transmission, queuing and propagation.

    C. JitterThe variation of time between packet transmit from

    source to reach destination, means one packet reach in

    100 ms and one reach in 125 ms to handle this problem

    jet-buffer is used at receiving end and it has two type

    static jet-buffer which hardware base and dynamic base

    which is software base and can be handle by administra-

    tor .but should take care about jet-buffer because some

    time it is also becoming reason for delay like memory

    over-flows etc. The following are the some measurement

    and recommendation of ITU-T G.114 for a VoIP call for

    the three attribute which is define in tab. I [5]. Factorssuch as packet delay, jitter, packet loss and network

    latency can noticeably affect the quality of UDP- based

    TABLE I

    QOS

    Packet delay Packet loss Jitter

    150 ms 1% 25 ms

    services such as VoIP and video streaming. Contrary to

    TCP-based services such as HTTP, SMTP, etc, a steady

    stream of data packets is crucial for VoIP connections,

    where even slight connectivity problems can cause noise

    or echo. Quality of any service depends on the traffic

    flow as well as the network of terminating partners.

    Following are some issues should be considered to

    provide better-quality service. Number of calls managed

    simultaneously by the network The alternate way to

    transfer the call to it desired destination in case of any

    fault/failure occurred in the network CODECs for codingand encoding purposes. Overall setup of the network

    [6].

    D. Original IEEE 802.11 MAC layer

    The original IEEE802.11 has no idea QoS especially

    for voice data application have no sensitivity about Delay

    jitter. The basic MAC layer use distributed coordina-

    tion function (DCF) and Point coordination function

    (PCF ) to share medium with station both have several

    limitation [2]. DCF relies on CSMA/CA and optional802.11 RTS/CTS to share the medium between station.

    The problem in DCF is that if many station want to

    communicate at the same time there is always a collision

    occur and it is based on collision avoidance means it has

    to wait the medium to be free which produce delay and

    if collision occur it is waste of the bandwidth and make

    communication slow. some problem in DCF:

    there is no QoS guaranty and priority between data

    traffic like voice and data;

    if a station sense medium and it is free and

    get medium to communicate no other cantcommunicate until it didnt let free the medium

    if a station has slow bit rate it will capture the

    medium for along time.

    PCF is the other coordination which is define by ba-

    sic IEEE802.11. It is optional. PCF is used only in

    infrastructure mod in which all station are connected

    by one center object called Access Point (AP). PCF

    define two frame Contention Free Period (CFP) and

    Contention Period (CP). In the CP DCP is used. To

    give the right of communicate over the medium the CFPsend Contention-Free-Poll (CF-Poll) to station at time

    one packet each The AP is coordinator PCF has a little

  • 7/28/2019 Voice IP

    3/8

    WIRELESS NETWORKS, CHALMERS 2011 3

    bit QoS management but have no idea of the different

    class of traffic.

    E. IEEE 802.11e

    This standard define enhancement in the original

    IEEE802.11 Mac layer DCF and ECF with new coordi-

    nation function Hybrid Coordination Function (HCF). It

    proposed priority and class based traffic means the voice

    and multimedia application data class will have high

    priority during transmission compare with other data like

    email data class in a shared wireless medium etc. There

    are two method to access the channel to communicate

    like original IEEE802.11 MAC. HCF Controlled Chan-

    nel Access (HCCA) and Enhanced Distributed Channel

    Access (EDCA)[6]. With the EDCA the data which

    have high priority will have have chance to send early

    then the low priority data and station having EDCA

    implemented will have to wait less to send data. It work

    mostly like PCF. In PCF scenario the interval between

    to beacon frame is divide into two period CFP and an

    CP, the HCCA is allowing the CFP to initiated almost

    any time during CP. This kind of CFP is called Access

    Phase (CAP) in 802.11e. The AP will initiate CAP any

    time which it want and can receive frame from other

    contention-free manner. The CAC is a method which

    will decide whether a new connection will be allow

    to established or not, it will be decide on the basisof capacity of WLAN means if the new connection is

    allowed what will be MOS or quality of over all call

    which would be specified. So the CAC will maintain the

    over all quality of Voice of VoWLAN. For infrastructure

    mode of VoWLAN the CAC can be implemented in AP.

    Codec is used to convert voice signal to digitally encoded

    version compress it on the sender end and then reverse

    the processes on the receiver end. These codec are

    standardized by International Telecommunication (ITU-

    T). There are many codec technique which is used in

    VoIP for Encoding and Decoding. Some coding and itdifferent result are mentioned in below table fig.2 which

    has been calculated with different IEEE 802.11 standard

    with sample period 20, and voice activity detection active

    [8].

    The above data int the table is calculated by [9]

    Connect 802 VoIP Bandwidth Provisioning Calculator

    : from the table we got different result from different

    code which different bit rate per kbps and with have

    different WLAN IEEE802.11 like a,b,g and got different

    MOS and found how much simultaneously connection or

    calls can be established at time on per AP. Among weobserve that Codec G.711 have high MOS rate and with

    reasonable simultaneously calls at time But it should

    Fig. 2. VoIP over WLAN

    be care about the bandwidth of connection is ok for

    requirement of codec selected like G.711 require at least

    128 bit for both way communication. MOS is Inter-

    national Telecommunications Union Telecommunication

    Standardization sector (ITU-T) approved which gives

    a numerical indication of the perceived quality of the

    media received after being transmitted and eventually

    compressed using codec. The WLAN are working on

    radio wave which are open which can eavesdrops and

    some one can manage to use it illegally like crack the

    secret key.

    F. IEEE 802.11i

    The IEEE802.11e enhance the security issue of orig-

    inal WLAN and put forward the WAP2, it using Ad-

    vanced Encryption Standard (AES) block cipher. TheWEP and WAP were using RC4 stream cipher. The

    IEEE802.11e replace the issue of Authentication and

    privacy issue with more detail and security adjustment

    [9]. Different VLAN can be used to separate Voice traffic

    and data traffic: it will solve the space problem and

    voice device can be protected from external network.

    Separate VLAN will have private addresses which will

    hide phone device from directly connected to public

    network; QoS trust boundary extension to voice devices-

    QoS trust boundaries can be extended to voice devices

    without extending these trust boundaries and, in turn,QoS features to PCs and other data devices; protection

    from malicious network attacks-Subnet access control,

    can provide protection for voice devices from malicious

    internal and external network attacks such as worms,

    denial of service (DoS) attacks, and attempts by data

    devices to gain access to priority queues; ease of man-

    agement and configuration-Separate VLANs for voice

    and data devices at the access layer provide ease of

    management and simplified QoS configuration.

    III. VOIP OVER 3GTraditionally, real-time services (e.g. voice) are

    transported over dedicated channels because of their

  • 7/28/2019 Voice IP

    4/8

    WIRELESS NETWORKS, CHALMERS 2011 4

    Fig. 3. Transport of speech in IP

    delay sensitivity while data is transported over shared

    channels because of its transmitted in short, uneven

    spurts. In order to carry voice on IP networks,

    appropriate protocols must be used. The main protocolsare Real Time Protocol (RTP), User Datagram Protocol

    (UDP) and Internet Protocol (IP) [11]. In Fig. 1, the

    voice frames are generated in the application layer,

    encoded and encapsulated within payload of an RTP

    SDU. The RTP PDU is encapsulated into an UDP SDU,

    which is delivered to the IP layer.

    Adaptive Multi-Rate Speech Codec (AMR) is a codec

    with 8 narrow-band speech encoding modes with bit

    rates between 4.75 and 12.2 kbps. If the data rate

    is 12.2 kbps, the AMR codec generates packets of244 bits which represent voice frames of 20 ms [12].

    Since the AMR codec encodes and decodes digital

    speech data with an optimum power and bandwidth

    consumption, the Internet Engineering Task Force

    (IETF) has approved the RTP payload format for AMR.

    Real Time Protocol (RTP) is an end to end transport

    protocol, used to transport multimedia traffic in IP

    networks, supporting unicast and multicast traffic. In the

    case of VoIP service, it is implemented together with

    UDP/IP [11]. Since RTP does not provide any reliabilitymechanisms and other layers should be implemented.

    AMR and RTP the main performance parameters for

    VoIP quality that are described earlier, can be measured

    by the RTP protocol. RTP

    AMR and RTP The main performance parameters for

    VoIP quality that are described earlier, can be measured

    by the RTP protocol.

    User Datagram Protocol (UDP) as a transport layer

    protocol for VoIP over Internet Protocol (IP), UDP is

    used to avoid any retransmission delays. On the otherhand, it provides no reliability on datagram delivery.

    The UDP header size is standardized in 8 bytes and 20

    Fig. 4. FER for several loads and channel error for Simulation 1[11]

    bytes for IPv4 or 40 bytes for IPv6.

    Header Compression (HC) in 3G networks it is

    important to use bandwidth efficiently. On the other

    hand, large headers of the protocols used when voice

    data is sent over the wireless network where a high bit

    error rate (BER) due to fading and mobility is present.

    Robust Header Compression (ROHC) protocol has been

    developed for this problem. The effective compression

    makes use of the fact that majority of the fields in

    the combined IP, UDP and RTP header either remain

    constant or introduce constant change throughout a

    session.

    A. QoS Analysis

    One main parameter for assessing packet loss is FER

    (Frame Error Rate). Although packet loss is undesired

    some loss can be tolerated since error-concealment

    techniques can be used. Buffer length can also cause

    packet loss due to discarding of delayed packets. On

    the other hand buffer length also may also increase the

    delay where for acceptable conversational quality, the

    maximum end-to-end delay should be around 250-300

    ms [13]. Therefore buffer length takes an important roleshort buffering time will risk buffer underflows causing

    jitter, and long buffering time causes long delay and

    buffer overflows. Too short buffering time may also

    cause increased packet loss due to loss of segmented

    packets. Simulation with parameters specified for 2 dif-

    ferent simulations can be seen in tab. II.

    From the first simulation it can be seen that for

    different error probabilities, ranging from 1% to 10%,

    packet loss is directly related to the load on the wireless

    network. With the increase number of network users,

    applied packet switching technique is not feasible.Therefore it can be said that delay and delay jitter

    mainly depends on both Round Robin switching

  • 7/28/2019 Voice IP

    5/8

    WIRELESS NETWORKS, CHALMERS 2011 5

    TABLE II

    SIMULATION PARAMETERS

    Parameter Value (Simulation 1) Value (Simulation 2)

    Simulation runs 10000 6min30s

    of speech

    Load Variable One user

    Channel error Variable Variable

    probability

    AMR source 12.2kbps 12.2kbps

    data rate

    AMR voice 20ms 20ms

    frame duration

    Call duration 120s 390s

    Sil ence Voice on/off Sil ence Descri pt or (SID)

    p er io ds ( mean d ur at io n 3 s) ( 16 0 ms i nt er val s)

    AMR voice 244bits 244bits

    packet payload

    size

    Protocol Stack RTP + UDP + IP v4 RTP + UDP + IP v6

    size = 40 byte = 60 byte

    Header Robust HC Robust HC

    Compression (HC)RLC mode Unacknowledged Unacknowledged

    Mode Mode

    Maximum number 3 None

    of MAC-hs

    retransmissions

    Number of 4 None

    MAC-hs H- ARQ

    parallel processes

    Packet scheduling Round-Robin None

    algorithm

    Delay budget 100ms Predefined jitter

    buffer (FIFO algorithm)

    Fig. 5. Mean packet delay for several loads and channel for

    Simulation 1[11]

    technique and Hybrid-ARQ mechanism where the

    main features of MAC-hs (Medium Access Control-

    high speed) protocol of HSDPA are retransmission

    of erroneous packets which is handled by H-ARQ

    and sequential delivery of the packets to the upper

    layer [14]. This reasoning can also be seen in the PDF

    of delay jitter for 5 fixed users on the network in figure 7.

    In simulation 2, a predetermined buffer is imple-mented; therefore average network delay is constant

    for different error probabilities On the other hand as

    Fig. 6. PDF of the mean packet delay jitter for several channel error

    for Simulation 1[11]

    Fig. 7. Simulation Results for Simulation 2[1]

    packet loss ratio increases on the wireless channel, total

    packet loss rate increases. The reason for occurrence of

    erroneous packets in loss-free simulation is due to the

    packet segmentation at RLC (Radio Link Control Layer),where packets larger than one TTI (Time Transfer Inter-

    val) are segmented over several TTIs, introducing longer

    transmission delays and packet drops.

    IV. VOIP OVER LTE

    There are two important conditions must be met to

    ensure an adequate VoIP quality:

    1) delay from sender to receiver must be as low as

    possible;

    2) packet loss must be between 1% to 3%.

    So, in LTE, end-to-end Quality of Service is based

    on two parameters that formalize these two conditions.

    First, Layer 2 Packet Delay Budget is specified for

    every connection and for every User Equipment (UE).

    Second, Layer 2 Packet Loss Ratio is defined in order

    to guarantee the above specification. Hence, if a VoIP

    connection has a L2PDB of 100 ms and a L2PLR

    of 2% it mens that the QoS level for a subscriber is

    satisfactory. [16]

    In wireless networks, like LTE, the principal cause

    of issues is the path between the radio base-station and

  • 7/28/2019 Voice IP

    6/8

  • 7/28/2019 Voice IP

    7/8

    WIRELESS NETWORKS, CHALMERS 2011 7

    their own average channel conditions but also the

    delay bound. Therefore, it is able to meet the

    different QoS requirements of real time users [15].

    MAC-hs protocol enables retransmissions which

    causes decrease on QoS. On the other hand intro-

    ducing predetermined TTIs also causes segmenta-

    tion problems related to RLC creating delay and

    packet loss. Therefore an adaptive TTI and buffer-

    ing should be simulated for future work. Only

    then MAC-hs protocols retransmission can be fully

    effective since RLC does not guarantee delivery.

    Current Packet Loss Concealment techniques are

    effective only for small numbers of consecutive lost

    packets, for example a total of 20-30 milliseconds

    of speech, and for low packet loss rates. Therefore

    a further study on intelligent PLC where a learning

    technique can overcome packet loss issues.

    VoIP over LTE, QoS analysis is mainly based on variants

    of H-ARQ to improve Eb/No over a low packet errorrate which is usually 103 for voice and 106 for

    data transmissions. The trade-offs in this assessment

    was between memory usage and SNR. Standard H-ARQ

    needs almost no memory but provides very little SNR

    improvement on the other hand Type-II Full Incremental

    Redundancy requires high memory but provides more

    than 10 dB improvement compared to the standard H-

    ARQ. Therefore Type-III Partial IR where the retransmit-ted packet can be chase combined with previous packets

    to increase the diversity gain, is the main candidate for

    future work.

    REFERENCES

    [1] Renaud Cuny, Ari Lakaniemi, VoIP in 3G Networks: An End-

    to-End Quality of Service Analysis, Nokia Research Center.

    [2] INTERNATIONAL JOURNAL OF COMMUNICATION SYS-

    TEMS Int. J. Commun. Syst. 2006; 19:491?508.

    [3] http://uwanted.blogspot.com/2006/09/wireless-voip.html

    [4] Recommendation of ITU-T G.114

    [5] 1998 - 2011 Paessler AG.[6] www.advancedvoip.com

    [7] IInt. J. Commun. Syst. 2006; 19:491?508 Published on-

    line in Wiley InterScience (www.interscience.wiley.com). DOI:

    10.1002/dac.801

    [8] http://www.ozvoip.com/voip-codecs/

    [9] http://www.connect802.com/voipbandwidth.php

    [10] Voice over WLAN Campus Test Architecture Cisco

    [11] Leonardo Ramon N. Sousa, Marcone L. Carvalho, Emanuel B.

    Rodrigues, Leonardo Sampaio and Francisco R. P. Cavalcanti,

    Quality of Service Evaluation of VoIP over HSDPA, Wireless

    Telecommunications Research Group - GTEL, Department of

    Teleinformatics Engineering - DETI, Federal University of

    Ceara, 2006.

    [12] 3GPP, Mandatory speech codec speech processing func-

    tions; amr speech codec; error concealment of lost frames,

    3rd GenerationPartnership Project, 1999. [Online]. Available:

    http://www 3gpp org

    [13] IETF Differentiated Services (DiffServ) Working Group,

    http://www.ieft.org/html.charters/diffserv-charter.html.

    [14] Robert Bestak, Performance Analysis of MAC-hs Protocol,

    Czech Technical University in Prague, Department of Telecom-

    munications Engineering, 2005.

    [15] Matthias Malkowski, Andreas Kemper, Xiaohua Wang, Perfor-

    mance of Scheduling Algorithms for HSDPA, CommunicationNetworks, RWTH Aachen University.

    [16] Capacity Enhancement of VoIP over LTE by Stochastic Adap-

    tive Modulation and Coding, K.Homayounfar and B. Rohani,

    10-10 Cendex Center , Singapore.

    [17] E. Dahlman. 3G Evolution: HSPA and LTE for Mobile Broad-

    band. Academic Press, 2008.

    [18] Kian Chung Beh, Angela Doufexi, Simon Armour, PER-

    FORMANCE EVALUATION OF HYBRID ARQ SCHEMES

    OF 3GPP LTE OFDMA SYSTEM, The 18th Annual IEEE

    International Symposium on Personal, Indoor and Mobile Radio

    Communications (PIMRC?07).

    [19] D. Chase, Code combining; A maximum likelihood decoding

    approach for combining an arbitrary number of noisy packets,

    IEEE Trans. Commun., vol. 33, pp. 385 to 393, May 1985.[20] S. Kallel, Analysis of Type II Hybrid ARQ Schemes with code

    combining, IEEE Trans. on Commun., vol. 38, No. 8, Aug.

    1990.

    [21] Kingsley Oteng-Amoako, Jinhong Yuan, Saeid Nooshabadi,

    Selective Hybrid-ARQ turbo schemes with various Combining

    methods in Fading Channels, Dept. of Electrical Eng. and

    Telecomm, University of NSW, Sydney 2052, Australia.

  • 7/28/2019 Voice IP

    8/8

    Group-13 Voice over IP - WLAN, 3G and LTE issues

    Question:

    What are the Quality of Service issues for voice communications over different wireless technologies?

    Answer:

    - Delay- Packet Loss- Jitter (Delay Variation)