All Columns by:  Date  I  Company  I  Telco Triple Play I Circuit-to-Packet  I  Metro Ethernet
For information on how to submit a column, see Editorial Calendar.
 

Guest Column

Voice over Packet Protocols
VoIP and VoATM (VoAAL1, VoAAL2)

(continued)

Bandwidth Requirements

The table below summarizes and compares the absolute bandwidth performance of the VoAAL1, VoAAL2, and VoIP protocols. The performance of legacy TDM networks, where 64 kbps bandwidth for voice is typically reserved, is also included for reference.

Bandwidth Efficiency

 

G.711

kbps/Figure of merit(1)

G.726
(Compression = 2:1)

kbps/Figure of merit(1)

Complex Vocoders
(G.723 / G.729)
(Compression = 8:1)

kbps/Figure of merit(1)

Legacy TDM
(ex: T1 circuits)

64kbps / 1 32kbps / .5 Not Applicable

VoAAL1
(One ch. per VC)

72kbps / 1.12

(6 ms packet)

36kbps / .56

(12 ms packet)

Not Applicable
VoAAL2

77kbps /1.20

(5.4 ms packet. delay – 44 byte CPS packet)

38kbps /.59

(11 ms packet. delay – 44 byte CPS packet)

12kbps / .19

(20 ms packet. delay – 20 byte CPS packet)

VoIP

80kbps / 1.25

(20 ms packet. delay - 160 byte payload IP packet) (2)

48kbps /.75

(20 ms packet. delay - 80 byte payload IP packet) (2)

24kbps /.37

(20 ms packet. delay - 20 byte payload IP packet) (2)

(1) “1” is equivalent to 64kbs. The lower the number, the better the relative efficiency.
(2) Packet payload size is restricted to limit the packetization delay to 20 milliseconds. Header is fixed at 40 bytes.

Table 2: VoP bandwidth performance

G.711/G.726 - Whenever G.711 or G.726 encoders are used, VoAAL1 theoretically offers the best possible performance. In the case of trunking applications, this is true only when all channels are active on a VoAAL1 circuit. In practice however, since normally all channels on a trunk are never fully active, VoAAL1 is actually less efficient than VoAAL2 and VoIP. In cases where silence suppression is used VoAAL1 is even less efficient, as silence suppression is a feature not supported by this protocol. As a result, VoAAL2 and VoIP are always (even theoretically) more efficient than VoAAL1. Note that typically VoAAL2 is slightly more efficient than VoIP because ATM packets have smaller headers relative to standard payloads than IP packets.

Complex Vocoders - VoAAL1 does not support complex vocoders, so it is not in the race here. As can be readily seen in Table 2, VoAAL2 provides better bandwidth performance than VoIP for otherwise equal performance (i.e. equal delays). This is because ATM packets feature smaller headers than IP packets and also because VoAAL2 supports trunking. In the end however, although the above comparisons are interesting, this performance disparity in favor of VoAAL2 will have very little bearing on which protocol/encoder pairs are retained for broad deployment. Other factors more related to the delays, echo impairment, and popularity of protocol preference will prevail.

<< Previous page         Next page >>

Page 3 of 4

 

 

 

 

 

Subscription Info  |  UnSubscribe  |  Archive  | Marketing & Advertising  |  Link2Us Events  | About Us  |  Contact Us
Copyright © 2008 Converge! Media Ventures, Inc.  All rights reserved.