Welcome to TMCnet.com
TMC Launches New Sites: Cable 4G Wireless Evolution  |  Satellite  |  Green Tech  | IT | IVR |  ITEXPO East begins in:   REGISTER NOW!
Columnists:
E-mail this page to a friend Order reprints online Print this page Bookmark this page Free magazines Free newsletters RSS-XML alerts

[November 17, 2003]

CESoP Is Not VoIP – Making The Voice Connection Over A Packet Network

BY PETER MEYER


Voice over an IP-based network was once a novelty, suited only for those willing to endure poor voice quality in exchange for a few dollars in savings. Today, reality has finally caught up to the hype.

Businesses are increasingly migrating circuit-switched voice services to their data networks to realize communications as well network infrastructure and maintenance savings. Service providers are now announcing significant investments in IP telephony.  

One of the most common technologies used to transport voice over packet is Voice over IP (VoIP). However, it is not the only solution. Circuit Emulation Services over Packet (CESoP) takes a fundamentally different and less complex approach to transport voice and circuit-switched traffic over packet-based networks, while at the same time supporting a broader range of applications than VoIP

CIRCUIT TO PACKET
Adopting the same approach as circuit emulation for ATM -- now a widely deployed technique -- CESoP provides a transparent mechanism to trunk circuit-switched services across a packet network. TDM circuit traffic, including both the data and signaling, is packetized as either unstructured leased-line data or structured Nx64 Kbps voice channels. These packets are then transported across an Ethernet, IP or MPLS packet network. At the far-end the ingress packets are smoothed using a receive jitter buffer, TDM circuits are then extracted from the packets and played out onto the TDM interface.

In contrast, VoIP relies on comparatively older technology to transport TDM voice channels across the IP network, using gateways to terminate PSTN (public switched telephone network) signaling and voice traffic. The gateway then individually negotiates a single-channel connection across the IP network, while at the far end another gateway converts the voice packets back to TDM. 

Figure 1: Circuit Emulation Across A Packet Switched Network

Figure 1

KEEP IT SIMPLE
The efficiency of the two solutions can be compared based on the required software and hardware to support the VoIP or CESoP interworking function (IWF).  A generic IWF includes three main functions:

Voice Processing: Includes echo cancellation, compression, tone detection and generation, as well as voice activity detection (VAD) and comfort noise generation (CNG) for silence detection and suppression.

Packet Processing: Consists of conversion between TDM and packets (packetization), implementing the packet network protocol stack, providing a jitter buffer to compensate for packet delay variation and clock recovery.

Control and signaling: Includes telephony functions, such as channel associated signaling (CAS) and common channel signaling (CCS), and packet network call control, such as H.323 and SIP.

A generic VoIP solution demands all three functions, whereas a CESoP solution simplifies the complexity of the solution by eliminating voice processing requirements. Moreover, call control and signaling demands are greatly reduced in a CESoP solution.

A VoIP solution terminates the PSTN signaling at each IWF and performs the equivalent call control setup between the two IWFs, involving a complete signaling stack between two IWFs or gateways using control protocols such as H.323 or SIP.

In comparison, a CESoP solution generally does not terminate the PSTN signaling, but instead “tunnels” the CAS or CCS transparently through the packet switched network. The PSTN call control is still performed in the existing TDM equipment boxes, and no additional complex stacks are required between the CESoP IWFs. A CESoP connection is not visible to, and does not interfere with, the PSTN signaling.

A VoIP solution will typically perform compression to minimize packet network bandwidth. Rather than compress voice traffic, with CESoP, bandwidth savings are achieved through trunking multiple TDM channels in a single packet connection. VoIP may also implement VAD/CNG for silence detection and suppression, whereas a CESoP solution does not need to implement these bandwidth-saving mechanisms.

A VoIP solution also requires voice echo cancellation. A CESoP solution, with its low packetization delay, does not necessarily require echo cancellation when the end-to-end delays are kept under the 25 millisecond (ms) mark, as specified in ITU-T G.131 "Control of Talker Echo."

Figure 2: Comparing The Interworking Function Between CESoP And VoIP

Figure 2

MAINTAINING VOICE QUALITY
The success of any circuit-over-packet solution relies primarily on the ability to provide voice services of the same quality and reliability as a traditionally delivered service. Along with the usual quality concerns caused by delay, compression and echo, an IP-based network also introduces factors such as packet loss and jitter buffer slips. 

The most significant quality factor is the delay penalty incurred from packetizing the voice channels. The VoIP solution traditionally implements one channel per connection with a packetization size between 10 ms and 40 ms. A CESoP solution implements multiple channels per connection with a packetization size of only 125 microseconds to one ms.

Given the goal of meeting a one-way latency under 25 ms to avoid the use of echo cancellers as per G.131, or under 150 ms to provide high-quality voice service as per ITU-T G.114 "One-Way Transmission Time," the CESoP solution introduces significantly less delay than the VoIP solution.

Another issue is that with VoIP, timing synchronization is usually not treated as critical, resulting in jitter buffer slips that may yield "clicking" sounds. Timing synchronization in a CESoP solution, because it is also targeted to support data, is critical. A CESoP solution will utilize either adaptive or differential clock recovery to ensure the TDM circuits remain synchronized with no buffer slips. 

CONSIDER THE PACKET NETWORK
While a VoIP solution is targeted to use IP as the layer 3 protocol, a CESoP solution runs equally well over an IP, MPLS or Ethernet network, with the unmatched flexibility to support voice, fax, modem and data over IP or MPLS. 

A VoIP solution is best suited for a bandwidth-limited environment where compression and VAD/CNG is required to keep the packet bandwidth to an absolute minimum. A CESoP solution achieves bandwidth reduction through the use of Nx64 trunking. Compared to multiple G.729 eight Kbps VoIP connections, a CESoP connection providing G.711 Nx64 channel trunking can achieve equivalent or superior bandwidth savings.

Since it places little restriction on the quality of the network, a VoIP solution may be implemented in a managed or unmanaged network -- even the Internet -- but allows for no voice quality guarantees. A CESoP solution requires a managed packet network to ensure a performance level that is adequate for either data or voice traffic that may be present on the emulated TDM circuit.

MINIMIZE COST
A CESoP solution costs less to implement than VoIP; consider such factors as developing an IWF and the migration of existing TDM-based equipment. For CESoP IWF development, eliminating voice processing functions reduces hardware requirements and licensing royalties. Signaling transparency to CAS and CCS with CESoP reduces the amount of software signaling stacks required for the IWF when compared with a VoIP solution, which utilizes H.323 and SIP call control. 

For the end customer with a considerable network investment, this same transparency allows continued use of existing TDM-based equipment. CESoP delivers the true add-on functionality needed to migrate the TDM traffic across a packet network, without requiring replacement or a substantial upgrade of existing TDM-based equipment.

Both CESoP and VoIP are capable solutions aimed at exploiting packet switched networks to carry voice and circuit-switched traffic. However, CESoP is easier to implement, supports a wider variety of applications across any managed network, and guarantees the voice quality necessary to ensure indistinguishable services for end users.

Peter Meyer is a member of the Applications Engineering Department supporting packet processors, with Zarlink Semiconductor. For more information on CESoP, visit http://cesop.zarlink.com/

Respond to this article in our forums!

E-mail this page to a friend Order reprints online Print this page Bookmark this page Free magazines Free newsletters RSS-XML alerts


Today @ TMC
Upcoming Events
19th INTERNET TELEPHONY Conference & EXPO East
February 2-4, 2009 — Miami Beach Convention Center, Miami, FL
Digium Asterisk World Conference
February 2-4, 2009 — Miami Beach Convention Center, Miami, FL
4G Wireless Evolution Conference
February 2-4, 2009 — Miami Beach Convention Center, Miami, FL
6th Annual Communications Developer Conference
February 2-4, 2009 — Miami Beach Convention Center, Miami, FL
20th INTERNET TELEPHONY Conference & EXPO West
October 27-29, 2009 — Los Angeles Convention Center, Los Angeles, CA
Subscribe FREE to all of TMC's monthly magazines. Click here now.
TMC LOGO
Technology Marketing Corporation,
One Technology Plaza, Norwalk, CT 06854 USA
Ph: 800-243-6002, 203-852-6800; Fx: 203-866-3326
General comments: tmc@tmcnet.com. Comments about this site: webmaster@tmcnet.com.
About   Contact  Advertise
Technology Marketing Corp. 1997-2008 Copyright. Privacy Policy Sitemap
Advanced