1. Introduction
TDM Circuit Emulation technologies are the key to transport a TDM services along a Packet Transport Networks (PTN), directly influencing its capabilities and efficiency.
In circuit emulation, the TDM traffic is packetized, encapsulated and transmitted over the PTN using Pseudo Wire Emulation End to End (PWE3) technology and encapsulation methods such Structure Agnostic TDM over Packet (SAToP) and Circuit Emulation over Packets (CEP).
There is incorrect perception in the telecom industry regarding the bandwidth efficiency of the circuit emulation technologies. This perception is derived from the fact that circuit emulation adds additional overhead to each packet, and as a result reducing the bandwidth efficiency of the PTN, however when analyzing the bandwidth efficiency of CEP compared to native SDH, we can see that CEP is as good as SDH in terms of bandwidth efficiency!
For example, if we compare the number of VC12 that can be transmitted over 10G port we will discover that CEP can transmit 4006 VC12 compared to 4032 in SDH, In addition CEP discards unequipped and alarmed timeslots which also helps to increase the bandwidth efficiency compared to SDH.
This technical note provides information on CEP and SAToP technologies and their transport capabilities

2. Circuit Emulation technologies
2.1 CEP
2.1.1 Overview
CEP, which is described in RFC 4842, enables the transmission of TDM (SDH) circuits over an MPLS / MPLS-TP PTN. The CEP architecture is based on the PWE3 model (RFC 3985) in which a pseudo wire (PW) is established between two provider edge (PE) devices, encapsulating TDM circuit data and control information over an MPLS tunnel.

Figure 2-1: CEP MPLS encapsulation format
2.1.2 Functional Description
The following shows a typical CEP application, where a network element aggregates multiple TDM paths of the same type into PWs carrying CEP traffic (referred to as a PW-CEP). A PW-CEP transmits the payload of the TDM paths over the MPLS network, along with additional information used to maintain path timing and support OAM functions.

Figure 2-2: CEP Application Example
Various types of CEP services can be defined. The CEP service type indicates the type of TDM paths that it can transport. The following table shows the supported CEP service types and the number of TDM channels that can be transported by each.
Table 2-1: CEP Service Types
|
CEP Service Type |
TDM Channels |
|
STS-48c/ VC-4-16c |
1 |
|
STS-12c /VC-4-4c |
1 |
|
STS-3c/ VC-4 |
1 |
|
STS-1/VC-3 |
1 |
|
VT-2/TU-12 |
1-21 |
|
VT-1.5/ TU-11 |
1-28 |
CEP services require the nodes to be synchronized to a common time source, either through the packet network (for example, Synchronous Ethernet) or via external means (i.e. GPS).
Supported fault management and performance monitoring apply to both TDM and network aspects of CEP services.
Orckit-Corrigent CM-4000 PTN portfolio CEP implementation provides the operator with similar performance and experience as in native TDM equipment, including TDM-like fault management and performance monitoring. High-order (VC-3/4) and low-order (VC-12) path emulation are supported.
2.2 SAToP
2.2.1 Overview
SAToP is an encapsulation layer intended for carrying TDM circuits (E1/T1/E3/T3) over PTN in a structure-agnostic fashion. SAToP, which stands for Structure-Agnostic TDM over Packet, was defined in RFC 4553. This RFC describes a method for encapsulating TDM bit streams (T1, E1, T3, E3) as pseudo-wires over PTN. It addresses only structure agnostic transport, which means that the protocol completely disregards any structure that may possibly be imposed on these signals, in particular the structure imposed by standard TDM framing [G.704]. The SAToP solution conforms to the PWE3 architecture.

Figure 2-3: SAToP MPLS encapsulation format
2.2.2 Functional Description
2.2.2.1 SAToP Operation toward the PTN
After the PW is set up, the PTN direction operates as follows:
-
TDM data is packetized using the configured number of payload bytes per packet.
-
Sequence numbers, flags and timestamps are inserted in the SAToP headers.
-
SAToP, the PW demultiplexing layer and PTN headers are pre-pended to the packetized service data.
-
The resulting packets are transmitted over the PTN.
2.2.2.2 SAToP Operation toward the TDM Interface
After the PW is set up, the TDM direction operates as follows:
-
The SAToP toward the TDM interface includes a jitter buffer where the payload of the received SAToP packets are stored prior to transmit over the local TDM attachment circuit
-
The SAToP uses the sequence number in the SAToP PW control word for detection of lost and miss-ordered packets
-
The payload of the received SAToP packets marked with the L bit is replaced by the equivalent amount of the all-ones pattern
-
The TDM interface SAToP IWF performs Adaptive Clock Recovery (ACR), monitors PW defects and collects PW performance monitoring data
-
If the SAToP IWF detects loss of a preconfigured number of consecutive packets, it enters its packet-loss state
3. Circuit Emulation Bandwidth Efficiency – AS GOOD AS SDH
CEP technology is very efficient for high order (VC-3 and above) TDM services. Headers overhead consume approximately 5% of overall traffic, compared to SDH which increases bandwidth by 3.5%.
For low order TDM services (VC-11, VC-12), CEP enables transport of multiple LO payloads in a single CEP packet, as a result efficiency increases when large numbers of payloads go to the same destination.
The following table shows VC12/VC4 bandwidth efficiency comparison between CEP and SDH technologies over 10Gbps interface:
Table 2: CEP vs. SDH

As can be seen from the table the CEP bandwidth efficiency is good as SDH and in addition it has the following advantages over SDH:
-
CEP also supports Fractional STS-n/VC-n payloads which are mapped into shorter SPE. The use of shorter SPEs for fractional payloads maximizes the bandwidth efficiency across the network
-
CEP enables operation in a dynamic bandwidth allocation (DBA) mode where unequipped payloads and AIS signals are represented by shorter packets of only 64 bytes, saving bandwidth on the packet network when no TDM traffic is applied to the service
In SAToP technology the consumed overhead depends on the packet size and the encapsulation type, when using MEF8 encapsulation over MPLS network:
-
256 byte packets add 21% overhead, resulting in 2.59 Mbit/s for E1 transport
-
512 byte packets add 11% overhead. 2.32 Mbit/s for E1 transport
The following table shows E1 bandwidth efficiency comparison between SAToP and SDH technologies over 10Gbps interface:
Table 3: SAToP vs. CEP

In the same manner as CEP DBA function, using SAToP control word enables to save PSN bandwidth by not transferring invalid data (AIS signals).
4. Packet Delay in PTN
TDM service transported over PTN may experience packet delay variation (PDV) also known as jitter and from packets out of order.
The reasons for such packet delay are derived from the “packetization” process and from the packet network transmission delay.
To overcome the PDV and the packets out of order problems, a jitter buffer is used to smooth the jitter of the PTN and to reorder the packets. Increasing the depth of the jitter buffer helps to absorb big amount of jitter but on the other hand imposes a delay to the TDM traffic.
In addition to the jitter buffer configuration, Orckit-Corrigent CM-4000 PTN products assign always "High Priority" (HP) Class of Service (CoS) to the TDM services. With strict priority over any other CoS, TDM services are always served first, hence comply with the rigorous delay, jitter and loss requirements associated with TDM circuits.
In both technologies, CEP and SAToP, a jitter buffer is used and can be configured per service (PW), as for CoS the highest CoS is always used in CEP while in SAToP it is optional.
5. Clock Recovery and Synchronization
SAToP and CEP Circuit emulation technologies transmit PDH and SDH payloads as packets. Packets are asynchronous by nature thus special means are required to achieve synchronization for traffic and network synchronization.
Traffic synchronization over PSN is achieved by using the following clock recovery mechanisms:
-
Adaptive Clock Recovery (ACR) – recover the original clock for each circuit from the incoming rate of circuit emulation packets
-
ACR compensates for a certain amount of Jitter, typically 32msec
-
ACR is suitable for traffic quality but cannot be used to synchronize a whole network
-
-
Differential Clock Recovery (DCR) – recover the original clock for each circuit by using a network wide synchronization and marking the clock difference between each payload and the reference network clock
-
DCR is more accurate and is suitable for traffic and synchronization qualities
-
Network synchronization over PTN is achieved by one or more of the following mechanisms:
-
Synchronous Ethernet – The clock is recovered from the Ethernet interface, it supplies reference timing signal with proper frequency accuracy characteristics.
-
IEEE 1588V2 - also known as Precision Timing Protocol (PTP) is designed to synchronize nodes of a packet network, supplies Phase and frequency distribution.
-
SDH timing - Any TDM or Ethernet interface can be used as timing source / sink
Table 4: Clock recovery and synchronization

Orckit-Corrigent CM-4000 PTN products support ACR per SAToP PW, while SDH-equivalent timing is used for CEP PW.
IEEE 1588v2 and Synchronous Ethernet are also supported for packet network synchronization.
6. Standardization
Circuit emulation and packet synchronization technologies are all standard based.
ITU-T, IETF and MEF standardization bodies are defining the TDM circuit emulation requirements, how to encapsulate the TDM traffic over packet networks, how to deal with the data plane and define network synchronization and clock recovery methods.
CEP encapsulation method over PSN is defined in RFC 4842. SAToP encapsulation over PSN is defined in RFC4553. SAToP encapsulation over Ethernet is defined in MEF8.

Figure 6-1: MEF8 frame format
Orckit-Corrigent CM-4000 PTN products support all the above standards in addition to the network synchronization standards such G.8261 for synchronous Ethernet and IEEE 1588v2.
7. Applications
7.1 SDH Migration to a unified PTN

Figure 7-1: SDH migration application
As shown in figure 6-1 the SDH/PDH circuit emulation can be used to transmit the legacy TDM services over the packet network together with the Ethernet services providing multiservice solution for business customers.
7.2 Mobile Backhauling

Figure 7-2: Mobile Backhauling application
As shown in figure 6-2, the E1 and ETH Interfaces from the base stations are transported as standard Ethernet PW or TDM PW over the packet transport network, the TDM CoS is always high priority to comply with the jitter and wander requirements of the TDM services.
8. Summary
Telecommunication service providers need to support PDH and SDH services for circuit switching, cross-connections and grooming with the same bandwidth efficiency and transport capabilities like delay, jitter and wander and sub 50mSec protection.
Orckit-Corrigent CM-4000 PTN products provide a full set of SDH services over packet networks using CEP encapsulation for structured TDM services and SAToP encapsulation for structure-agnostics TDM services. These services are transported with the same bandwidth efficiency, delay, jitter and wander tolerances as in traditional SDH networks and implemented based on IETF standards.
TDM encapsulation over MPLS PW is good as traditional SDH networks.
