PWE3                                                          Y(J) Stein
Internet-Draft                                               R. Shashoua
Expires: June 8, 2007                                          R. Insler
                                                                M. Anavi
                                                 RAD Data Communications
                                                        December 5, 2006

                              TDM over IP

Status of this Memo

   By submitting this Internet-Draft, each author represents that any
   applicable patent or other IPR claims of which he or she is aware
   have been or will be disclosed, and any of which he or she becomes
   aware will be disclosed, in accordance with Section 6 of BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups.  Note that
   other groups may also distribute working documents as Internet-

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   The list of current Internet-Drafts can be accessed at

   The list of Internet-Draft Shadow Directories can be accessed at

   This Internet-Draft will expire on June 8, 2007.

Copyright Notice

   Copyright (C) The Internet Society (2006).


   TDM over IP (TDMoIP) is a structure-aware method for transporting
   Time Division Multiplexed (TDM) signals as pseudowires (PWs).  By
   ensuring TDM structure integrity, TDMoIP is able to better withstand
   network degradations.  As TDM signaling and individual channels are
   visible, it is possible to use mechanisms that exploit or manipulate
   these to conceal packet loss or conserve bandwidth.

Stein, et al.             Expires June 8, 2007                  [Page 1]

Internet-Draft                   TDMoIP                    December 2006

Table of Contents

   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  3
   2.  TDM Structure and Structure-aware Transport  . . . . . . . . .  4
   3.  TDMoIP Encapsulation . . . . . . . . . . . . . . . . . . . . .  6
   4.  Encapsulation Details for Specific PSNs  . . . . . . . . . . .  9
     4.1   UDP/IPv4 . . . . . . . . . . . . . . . . . . . . . . . . .  9
     4.2   MPLS . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
     4.3   L2TPv3 . . . . . . . . . . . . . . . . . . . . . . . . . . 13
     4.4   Ethernet . . . . . . . . . . . . . . . . . . . . . . . . . 15
   5.  TDMoIP Payload types . . . . . . . . . . . . . . . . . . . . . 17
     5.1   AAL1 Format Payload  . . . . . . . . . . . . . . . . . . . 18
     5.2   AAL2 Format Payload  . . . . . . . . . . . . . . . . . . . 19
     5.3   HDLC Format Payload  . . . . . . . . . . . . . . . . . . . 20
   6.  TDMoIP Defect Handling . . . . . . . . . . . . . . . . . . . . 21
   7.  Implementation Issues  . . . . . . . . . . . . . . . . . . . . 24
     7.1   Jitter and Packet Loss . . . . . . . . . . . . . . . . . . 24
     7.2   Timing Recovery  . . . . . . . . . . . . . . . . . . . . . 24
     7.3   Congestion Control . . . . . . . . . . . . . . . . . . . . 25
   8.  Security Considerations  . . . . . . . . . . . . . . . . . . . 27
   9.  IANA Considerations  . . . . . . . . . . . . . . . . . . . . . 28
   10.   Applicability Statement  . . . . . . . . . . . . . . . . . . 28
   11.   Acknowledgments  . . . . . . . . . . . . . . . . . . . . . . 28
       Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 29
   A.  Sequence Number Processing (informative) . . . . . . . . . . . 30
   B.  AAL1 Review (informative)  . . . . . . . . . . . . . . . . . . 32
   C.  AAL2 Review (informative)  . . . . . . . . . . . . . . . . . . 36
   D.  Performance Monitoring Mechanisms (informative)  . . . . . . . 38
     D.1   TDMoIP Connectivity Verification . . . . . . . . . . . . . 38
     D.2   OAM Packet Format  . . . . . . . . . . . . . . . . . . . . 39
   E.  Capabilities, Configuration and Statistics (informative) . . . 42
   F.  References . . . . . . . . . . . . . . . . . . . . . . . . . . 45
     F.1   Normative References . . . . . . . . . . . . . . . . . . . 45
     F.2   Informative References . . . . . . . . . . . . . . . . . . 47
       Intellectual Property and Copyright Statements . . . . . . . . 49

Stein, et al.             Expires June 8, 2007                  [Page 2]

Internet-Draft                   TDMoIP                    December 2006

1.  Introduction

   Telephony traffic is conventionally carried over connection-oriented
   synchronous or plesiochronous links (loosely called TDM circuits
   herein).  With the proliferation of Packet Switched Networks (PSNs),
   transport of TDM services over PSN infrastructures has become
   desirable.  Emulation of TDM circuits over the PSN can be carried out
   using pseudowires (PWs), as described in the PWE3 architecture
   [RFC3985].  This emulation must maintain service quality of native
   TDM; in particular voice quality, latency, timing, and signaling
   features must be similar to those of existing TDM networks, as
   described in the TDM PW requirements document [RFC4197].

   Structure Agnostic TDM over Packet (SAToP) [RFC4553] is a structure-
   agnostic protocol for transporting TDM over PSNs.  The present
   document details TDM over IP (TDMoIP), a structure-aware method for
   TDM transport.  In contrast to SAToP, structure-aware methods such as
   TDMoIP ensure the integrity of TDM structure and thus enable the PW
   to better withstand network degradations.  Individual multiplexed
   channels become visible, enabling the use of per channel mechanisms
   for packet loss concealment and bandwidth conservation.  TDM
   signaling also becomes accessible, facilitating mechanisms that
   exploit or manipulate this signaling.

   Despite its name, the TDMoIP(R) protocol herein described may operate
   over several types of PSN, including UDP over IPv4 or IPv6, MPLS,
   L2TPv3 over IP, or pure Ethernet.  Implementation specifics for
   particular PSNs are discussed in Section 4.  Although the protocol
   should be more generally called TDMoPW and its specific
   implementations TDMoIP, TDMoMPLS, etc. we retain the nomenclature
   TDMoIP for consistency with earlier usage.

   The interworking function that connects between the TDM and PSN
   worlds will be called a TDMoIP interworking function (IWF), and it
   may be situated at the provider edge (PE) or at the customer edge
   (CE).  The IWF that encapsulates TDM and injects packets into the PSN
   will be called the PSN-bound interworking function, while the IWF
   that extracts TDM data from packets and generates traffic on a TDM
   network will be called the TDM-bound interworking function.  Emulated
   TDM circuits are always point-to-point, bidirectional, and transport
   the same TDM rate in both directions.

   As with all PWs, TDMoIP PWs may be manually configured or set up
   using the PWE3 control protocol [RFC4447].  Extensions to the PWE3
   control protocol required specifically for setup and maintenance of
   TDMoIP pseudowires are described in [TDM-CONTROL].

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",

Stein, et al.             Expires June 8, 2007                  [Page 3]

Internet-Draft                   TDMoIP                    December 2006

   document are to be interpreted as described in [RFC2119].

2.  TDM Structure and Structure-aware Transport

   Although TDM circuits can be used to carry arbitrary bit-streams,
   there are standardized methods for carrying constant-length blocks of
   data called "structures".  Familiar structures are the T1 or E1
   frames [G704] of length 193 and 256 bits, respectively.  By
   concatenation of consecutive T1 or E1 frames we can build higher
   level structures called superframes or multiframes.  T3 and E3 frames
   [G704][G751] are much larger than those of T1 and E1, and even larger
   structures are used in the GSM Abis channel described in [TRAU].  TDM
   structures contain TDM data plus structure overhead; for example, the
   193-bit T1 frame contains a single bit of structure overhead and 24
   bytes of data, while the 32-byte E1 frame contains a byte of overhead
   and 31 data bytes.

   Structured TDM circuits are frequently used to transport multiplexed
   channels.  A single byte in the TDM frame (called a timeslot) is
   allocated to each channel.  A frame of a channelized T1 carries 24
   byte-sized channels, while an E1 frame consists of 32 channels.
   Since TDM frames are sent 8000 times per second, a single byte-sized
   channel carries 64 kbps.

   TDM structures are universally delimited by placing an easily
   detectable periodic bit pattern, called the Frame Alignment Signal
   (FAS), in the structure overhead.  The structure overhead may
   additionally contain error monitoring and defect indications.  We
   will use the term "structured TDM" to refer to TDM with any level of
   structure imposed by an FAS.  Unstructured TDM signifies a bit stream
   upon which no structure has been imposed, implying that all bits are
   available for user data.

   SAToP [RFC4553] is a structure-agnostic protocol for transporting TDM
   using PWs.  SAToP treats the TDM input as an arbitrary bit-stream,
   completely disregarding any structure that may exist in the TDM bit-
   stream.  Hence SAToP is ideal for transport of truly unstructured
   TDM, but is also suitable for transport of structured TDM when there
   is no need to protect structure integrity nor interpret or manipulate
   individual channels during transport.  In particular, SAToP is the
   technique of choice for PSNs with negligible packet loss, and for
   applications that do not require discrimination between channels nor
   intervention in TDM signaling.

   As described in [RFC4553], when a single SAToP packet is lost an "all
   ones" pattern is played out to the TDM interface.  This pattern is

Stein, et al.             Expires June 8, 2007                  [Page 4]

Internet-Draft                   TDMoIP                    December 2006

   interpreted by the TDM end equipment as an AIS indication, which,
   according to TDM standards [G826], immediately triggers a "severely
   errored second" event.  As such events are considered highly
   undesirable, the suitability of SAToP is limited to extremely
   reliable and overprovisioned PSNs.

   When structure-aware TDM transport is employed, it is possible to
   explicitly safeguard TDM structure during transport over the PSN,
   thus making possible to effectively conceal packet loss events.
   Structure-aware transport exploits at least some level of the TDM
   structure to enhance robustness to packet loss or other PSN
   shortcomings.  Structure-aware TDM PWs are not required to transport
   structure overhead across the PSN; in particular, the FAS MAY be
   stripped by the PSN-bound IWF and MUST be regenerated by the TDM-
   bound IWF.  However, structure overhead MAY be transported over the
   PSN, since it may contain information other than FAS.

   In addition to guaranteeing maintenance of TDM synchronization,
   structure-aware TDM transport can also distinguish individual
   timeslots of channelized TDM, thus enabling sophisticated packet loss
   concealment at the channel level.  TDM signaling also becomes
   visible, facilitating mechanisms that maintain or exploit this
   information.  Finally, by taking advantage of TDM signaling and/or
   voice activity detection, structure-aware TDM transport makes
   bandwidth conservation possible.

   There are three conceptually distinct methods of ensuring TDM
   structure integrity, namely structure-locking, structure-indication,
   and structure-reassembly.  Structure-locking requires each packet to
   commence at the start of a TDM structure, and to contain an entire
   structure or integral multiples thereof.  Structure-indication allows
   packets to contain arbitrary fragments of basic structures, but
   employs pointers to indicate where each structure commences.
   Structure-reassembly is only defined for channelized TDM; the PSN-
   bound IWF extracts and buffers individual channels, and the original
   structure is reassembled from the received constituents by the TDM-
   bound IWF.

   All three methods of TDM structure preservation have their
   advantages.  Structure-locking is described in [CESoPSN], while the
   present document specifies both structure-indication (see
   Section 5.1) and structure-reassembly (see Section 5.2) approaches.
   Structure-indication is used when channels may be allocated
   statically, and/or when it is required to interwork with existing
   circuit emulation systems (CES) based on AAL1.  Structure-reassembly
   is used when dynamic allocation of channels is desirable and/or when
   it is required to interwork with existing loop emulation systems
   (LES) based on AAL2.

Stein, et al.             Expires June 8, 2007                  [Page 5]

Internet-Draft                   TDMoIP                    December 2006

   Operation, administration, and maintenance (OAM) mechanisms are vital
   for proper TDM deployments.  As aforementioned, structure-aware
   mechanisms may refrain from transporting structure overhead across
   the PSN, disrupting OAM functionality.  It is beneficial to
   distinguish between two OAM cases, the "trail terminated" and the
   "trail extended" scenarios.  A trail is defined to be the combination
   of data and associated OAM information transfer.  When the TDM trail
   is terminated, OAM information such as error monitoring and defect
   indications are not transported over the PSN, and the TDM networks
   function as separate OAM domains.  In the trail extended case we
   transfer the OAM information over the PSN (although not necessarily
   in its native format).  OAM will be discussed further in Section 6.

3.  TDMoIP Encapsulation

   The overall format of TDMoIP packets is shown in Figure 1.

                            |    PSN Headers      |
                            | TDMoIP Control Word |
                            |   Adapted Payload   |

                   Figure 1.  Basic TDMoIP Packet Format

   The PSN-specific headers are those of UDP/IP, L2TPv3/IP, MPLS or
   layer 2 Ethernet, and contain all information necessary for
   forwarding the packet from the PSN-bound IWF to the TDM-bound one.
   The PSN is assumed to be reliable enough and of sufficient bandwidth
   to enable transport of the required TDM data.

   A TDMoIP IWF may simultaneously support multiple TDM PWs, and the
   TDMoIP IWF MUST maintain context information for each TDM PW.
   Distinct PWs are differentiated based on PW labels, which are carried
   in the PSN-specific layers.  Since TDM is inherently bidirectional,
   the association of two PWs in opposite directions is required.  The
   PW labels of the two directions MAY take different values.

   In addition to the aforementioned headers, an OPTIONAL 12-byte RTP
   header may appear in order to enable explicit transfer of timing
   information.  This usage is a purely formal reuse of the header
   format of [RFC3550].  RTP mechanisms, such as header extensions, CSRC
   list, padding, RTCP, RTP header compression, SRTP, etc. are not

Stein, et al.             Expires June 8, 2007                  [Page 6]

Internet-Draft                   TDMoIP                    December 2006

   The RTP timestamp indicates the packet creation time in units of a
   common clock available to both communicating TDMoIP IWFs.  When no
   common clock is available, or when the TDMoIP IWFs have sufficiently
   accurate local clocks or can derive sufficiently accurate timing
   without explicit timestamps, the RTP header SHOULD be omitted.

   If RTP is used, the fixed RTP header described in [RFC3550] MUST
   immediately follow the control word for all PSN types except UDP/IP,
   for which it MUST precede the control word.  The version number MUST
   be set to 2, the P (padding), X (header extension), CC (CSRC count),
   and M (marker) fields in the RTP header MUST be set to zero, and the
   PT values MUST be allocated from the range of dynamic values.  The
   RTP sequence number MUST be identical to the sequence number in the
   TDMoIP control word (see below).  The RTP timestamp MUST be generated
   in accordance with the rules established in [RFC3550]; the clock
   frequency MUST be an integer multiple of 8 kHz, and MUST be chosen to
   enable timing recovery that conforms with the appropriate standards
   (see Section 7.2).

   The 32-bit control word MUST appear in every TDMoIP packet.  Its
   format, in conformity with [RFC4385], is depicted in Figure 2.

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       |  RES  |L|R| M |RES|  Length   |         Sequence Number       |

              Figure 2.  Structure of the TDMoIP Control Word

   RES The first nibble of the control word MUST be set to zero when the
      PSN is MPLS, in order to ensure that the packet does not alias an
      IP packet when forwarding devices perform deep packet inspection.
      For PSNs other than MPLS the first nibble MAY be set to zero;
      however, in earlier versions of TDMoIP this field contained a
      format identifier that was optionally used to specify the payload

   L Local Failure (1 bit) The L flag is set when the IWF has detected
      or has been informed of a TDM physical layer fault impacting the
      TDM data being forwarded.  In the "trail extended" OAM scenario
      the L flag MUST be set when the IWF detects loss of signal, loss
      of frame synchronization, or AIS.  When the L flag is set the
      contents of the packet may not be meaningful, and the payload MAY
      be suppressed in order to conserve bandwidth.  Once set, if the
      TDM fault is rectified the L flag MUST be cleared.  Use of the L
      flag is further explained in Section 6.

Stein, et al.             Expires June 8, 2007                  [Page 7]

Internet-Draft                   TDMoIP                    December 2006

   R Remote Failure (1 bit) The R flag is set when the IWF has detected
      or has been informed, that TDM data is not being received from the
      remote TDM network, indicating failure of the reverse direction of
      the bidirectional connection.  An IWF SHOULD generate TDM RDI upon
      receipt of an R flag indication.  In the "trail extended" OAM
      scenario the R flag MUST be set when the IWF detects RDI.  Use of
      the R flag is further explained in Section 6.

   M Defect Modifier (2 bits) Use of the M field is optional, and when
      used supplements the meaning of the L flag.

      When L is cleared (indicating valid TDM data) the M field is used
      as follows:

       0 0  indicates no local defect modification.
       0 1  reserved.
       1 0  reserved.
       1 1  reserved.

      When L is set (invalid TDM data) the M field is used as follows:

       0 0  indicates a TDM defect that should trigger conditioning
            or AIS generation by the TDM-bound IWF.
       0 1  indicates idle TDM data that should not trigger any alarm.
            If the payload has been suppressed then the preconfigured
            idle code should be generated at egress.
       1 0  indicates corrupted but potentially recoverable TDM data.
       1 1  reserved.

      Use of the M field is further explained in Section 6.

   RES (2 bits) These bits are reserved and MUST be set to zero.

   Length (6 bits) is used to indicate the length of the TDMoIP packet
      (control word and payload), in case padding is employed to meet
      minimum transmission unit requirements of the PSN.  It MUST be
      used if the total packet length (including PSN, optional RTP,
      control word, and payload) is less than 64 bytes, and MUST be set
      to zero when not used.

   Sequence number (16 bits) The TDMoIP sequence number provides the
      common PW sequencing function described in [RFC3985], and enables
      detection of lost and misordered packets.  The sequence number
      space is a 16-bit, unsigned circular space; the initial value of
      the sequence number SHOULD be random (unpredictable) for security
      purposes, and its value is incremented modulo 2^16 separately for
      PW.  Pseudocode for a sequence number processing algorithm that
      could be used by a TDM-bound IWF is provided in Appendix A.

Stein, et al.             Expires June 8, 2007                  [Page 8]

Internet-Draft                   TDMoIP                    December 2006

   In order to form the TDMoIP payload, the PSN-bound IWF extracts bytes
   from the continuous TDM stream, filling each byte from its most
   significant bit.  The extracted bytes are then adapted using one of
   two adaptation algorithms (see Section 5), and the resulting adapted
   payload is placed into the packet.

4.  Encapsulation Details for Specific PSNs

   TDMoIP PWs may exploit various PSNs, including UDP/IP (both IPv4 and
   IPv6), L2TPv3 over IP (with no intervening UDP), MPLS, and layer-2
   Ethernet.  In the following subsections we depict the packet format
   for these cases.

   For MPLS PSNs the format is aligned with those specified in [Y1413]
   and [Y1414].  For UDP/IP PSNs the format is aligned with those
   specified in [Y1453] and [Y1452].  For transport over layer 2
   Ethernet the format is aligned with [MEF8].

4.1  UDP/IPv4

   ITU-T recommendation Y.1453 [Y1453] describes structure-agnostic and
   structure-aware mechanisms for transporting TDM over IP networks.
   Similarly, ITU-T recommendation Y.1452 [Y1452] defines structure-
   reassembly mechanisms for this purpose.  Although the terminology
   used here differs slightly from that of the ITU, implementations of
   TDMoIP for UDP/IP PSNs as described herein will interoperate with
   implementations designed to comply with Y.1453 subclause 9.2.2 or
   Y.1452 clause 10.

   For UDP/IPv4, the headers as described in [RFC768] and [RFC791] are
   prefixed to the TDMoIP data.  The format is similar for UDP/IPv6,
   except the IP header described in [RFC2460 ] is used.  The TDMoIP
   packet structure is depicted in Figure 3.

Stein, et al.             Expires June 8, 2007                  [Page 9]

Internet-Draft                   TDMoIP                    December 2006

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       | IPVER |  IHL  |    IP TOS     |          Total Length         |
       |         Identification        |Flags|      Fragment Offset    |
       |  Time to Live |    Protocol   |      IP Header Checksum       |
       |                     Source IP Address                         |
       |                  Destination IP Address                       |
       |      Source Port Number       |    Destination Port Number    |
       |           UDP Length          |         UDP Checksum          |
    opt|RTV|P|X|  CC   |M|     PT      |      RTP Sequence Number      |
    opt|                            Timestamp                          |
    opt|                         SSRC identifier                       |
       |  RES  |L|R| M |RES|  Length   |         Sequence Number       |
       |                                                               |
       |                        Adapted Payload                        |
       |                                                               |

                Figure 3.  TDMoIP Packet Format for UDP/IP

   The first five rows are the IP header, the sixth and seventh rows are
   the UDP header.  Rows 8 through 10 are the optional RTP header.  Row
   11 is the TDMoIP control word.

   IPVER (4 bits) is the IP version number, e.g. for IPv4 IPVER=4.

   IHL (4 bits) is the length in 32-bit words of the IP header, IHL=5.

   IP TOS (8 bits) is the IP type of service.

   Total Length (16 bits) is the length in bytes of header and data.

   Identification (16 bits) is the IP fragmentation identification

Stein, et al.             Expires June 8, 2007                 [Page 10]

Internet-Draft                   TDMoIP                    December 2006

   Flags (3 bits) are the IP control flags and MUST be set to Flags=010
      to avoid fragmentation.

   Fragment Offset (13 bits) indicates where in the datagram the
      fragment belongs and is not used for TDMoIP.

   Time to Live (8 bits) is the IP time to live field.  Datagrams with
      zero in this field are to be discarded.

   Protocol (8 bits) MUST be set to 11 hex = 17 dec to signify UDP.

   IP Header Checksum (16 bits) is a checksum for the IP header.

   Source IP Address (32 bits) is the IP address of the source.

   Destination IP Address (32 bits) is the IP address of the

   Source and Destination Port Numbers (16 bits each)

      Either the source UDP port or destination UDP port MAY be used to
      multiplex and demultiplex individual PWs between nodes.
      Architecturally [RFC3985], this makes the UDP port act as the PW
      Label.  PW endpoints MUST agree upon use of either the source UDP
      or destination UDP port as the PW Label.

      UDP ports MUST be manually configured by both endpoints of the PW.
      The configured source or destination port (one or the other, but
      not both) together with both the source and destination IP
      addresses uniquely identify the PW.  When the source UDP port is
      used as the PW label, the destination UDP port number MUST be set
      to the IANA assigned value of 2142 (0x085E).  All UDP port values
      that function as PW labels SHOULD be in the range of dynamically
      allocated UDP port numbers (49152 through 65535).

      While many UDP-based protocols are able to traverse middleboxes
      without dire consequences, the use of UDP ports as PW labels makes
      middlebox traversal more difficult.  Hence, it is NOT RECOMMENDED
      to use UDP-based PWs where port-translating middleboxes are
      present between PW endpoints.

   UDP Length (16 bits) is the length in bytes of UDP header and data.

   UDP Checksum (16 bits) is the checksum of UDP/IP header and data.  If
      not computed it must be set to zero.

Stein, et al.             Expires June 8, 2007                 [Page 11]

Internet-Draft                   TDMoIP                    December 2006

4.2  MPLS

   ITU-T recommendation Y.1413 [Y1413] describes structure-agnostic and
   structure-aware mechanisms for transporting TDM over MPLS networks.
   Similarly, ITU-T recommendation Y.1414 [Y1413] defines structure-
   reassembly mechanisms for this purpose.  Although the terminology
   used here differs slightly from that of the ITU, implementations of
   TDMoIP for MPLS PSNs as described herein will interoperate with
   implementations designed to comply with Y.1413 subclause 9.2.2 or
   Y.1414 clause 10.

   The MPLS header as described in [RFC3032] is prefixed to the control
   word and TDM payload.  The packet structure is depicted in Figure 4.

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       |            Tunnel Label               | EXP |S|     TTL       |
       |              PW label                 | EXP |1|     TTL       |
       |  RES  |L|R| M |RES|  Length   |         Sequence Number       |
    opt|RTV|P|X|  CC   |M|     PT      |      RTP Sequence Number      |
    opt|                            Timestamp                          |
    opt|                         SSRC identifier                       |
       |                                                               |
       |                        Adapted Payload                        |
       |                                                               |

                 Figure 4.  TDMoIP Packet Format for MPLS

   The first two rows depicted above are the MPLS header; the third is
   the TDMoIP control word.  Fields not previously described will now be

   Tunnel Label (20 bits) is the MPLS label that identifies the MPLS LSP
      used to tunnel the TDM packets through the MPLS network.  The
      label can be assigned either by manual provisioning or via an MPLS
      control protocol.  While transiting the MPLS network there may be
      zero, one or several tunnel label rows.  For label stack usage see

Stein, et al.             Expires June 8, 2007                 [Page 12]

Internet-Draft                   TDMoIP                    December 2006

   EXP (3 bits) experimental field, may be used to carry DiffServ
      classification for tunnel labels.

   S  (1 bit)  the stacking bit indicates MPLS stack bottom.  S=0 for
      all tunnel labels, and S=1 for the PW label.

   TTL (8 bits) MPLS Time to live.

   PW Label (20 bits) This label MUST be a valid MPLS label, and MAY be
      configured or signaled.

4.3  L2TPv3

   The L2TPv3 header defined in [RFC3931] is prefixed to the TDMoIP
   data.  The packet structure is depicted in Figure 5.

Stein, et al.             Expires June 8, 2007                 [Page 13]

Internet-Draft                   TDMoIP                    December 2006

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       | IPVER |  IHL  |    IP TOS     |          Total Length         |
       |         Identification        |Flags|      Fragment Offset    |
       |  Time to Live |    Protocol   |      IP Header Checksum       |
       |                     Source IP Address                         |
       |                  Destination IP Address                       |
       |                     Session ID = PW label                     |
       |                      cookie 1 (optional)                      |
       |                      cookie 2 (optional)                      |
       |  RES  |L|R| M |RES|  Length   |         Sequence Number       |
    opt|RTV|P|X|  CC   |M|     PT      |      RTP Sequence Number      |
    opt|                            Timestamp                          |
    opt|                         SSRC identifier                       |
       |                                                               |
       |                        Adapted Payload                        |
       |                                                               |

                Figure 5.  TDMoIP Packet Format for L2TPv3

   Rows 6 through 8 are the L2TPv3 header.  Fields not previously
   described will now be explained.

   Protocol the IP protocol field must be set to 73 hex = 115 dec, the
      user port number that has been assigned to L2TP by IANA.

   Session ID (32 bits) is the locally significant L2TP session
      identifier, and contains the PW label.  The value 0 is reserved.

   Cookie (32 or 64 bits) is an optional field that contains a randomly
      selected value that can be used to validate association of the
      received frame with the expected PW.

Stein, et al.             Expires June 8, 2007                 [Page 14]

Internet-Draft                   TDMoIP                    December 2006

4.4  Ethernet

   MEF Implementation Agreement 8 [MEF8] describes structure-agnostic
   and structure-aware mechanisms for transporting TDM over Ethernet
   networks.  Implementations of structure-indicated TDMoIP as described
   herein will interoperate with implementations designed to comply with
   MEF 8 section 6.3.3.

   The TDMoIP payload is encapsulated in an Ethernet frame by prefixing
   the Ethernet destination and source MAC addresses, optional VLAN
   header, and Ethertype, and suffixing the four-byte frame check
   sequence.  TDMoIP implementations MUST be able to receive both
   industry standard (DIX) Ethernet and IEEE 802.3 [IEEE802.3] frames
   and SHOULD transmit Ethernet frames.

   Ethernet encapsulation introduces restrictions on both minimum and
   maximum packet size.  Whenever the entire TDMoIP packet is less than
   64 bytes, padding is introduced and the true length indicated by
   using the Length field in the control word.  In order to avoid
   fragmentation the TDMoIP packet MUST be restricted to the maximum
   payload size.  For example, the length of the Ethernet payload for a
   UDP/IP encapsulation of AAL1 format payload with 30 PDUs per packet
   is 1472 bytes, which falls below the maximal permitted payload size
   of 1500 bytes.

   Ethernet frames MAY be used for TDMoIP transport without intervening
   IP or MPLS layers, however, an MPLS-style label MUST always be
   present.  In this four-byte header S=1, and all other non-label bits
   are reserved (set to zero in the PSN-bound direction and ignored in
   the TDM-bound direction).  The Ethertype SHOULD be set to 0x88D8
   (35032), the value allocated for this purpose by the IEEE, but MAY be
   set to 0x8847 (34887), the Ethertype of MPLS.  The overall frame
   structure is as follows:

Stein, et al.             Expires June 8, 2007                 [Page 15]

Internet-Draft                   TDMoIP                    December 2006

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
                                       |  Destination MAC Address
                           Destination MAC Address (cont)              |
       |                     Source MAC Address
           Source MAC Address  (cont)  |   VLAN Ethertype (opt)        |
       |VLP|C|      VLAN ID (opt)      |         Ethertype             |
       |              PW label                 | RES |1|    RES        |
       |  RES  |L|R| M |RES|  Length   |         Sequence Number       |
    opt|RTV|P|X|  CC   |M|     PT      |      RTP Sequence Number      |
    opt|                            Timestamp                          |
    opt|                         SSRC identifier                       |
       |                                                               |
       |                        Adapted Payload                        |
       |                                                               |
       |                     Frame Check Sequence                      |

               Figure 6.  TDMoIP Packet Format for Ethernet

   Rows 1 through 6 are the (DIX) Ethernet header; for 802.3 there may
   be additional fields, depending on the value of the length field, see
   [IEEE802.3].  Fields not previously described will now be explained.

   Destination MAC Address (48 bits) is the globally unique address of a
      single station that is to receive the packet.  The format is
      defined in [IEEE802.3].

   Source MAC Address (48 bits) is the globally unique address of the
      station that originated the packet.  The format is defined in

   VLAN Ethertype (16 bits) a 8100 hex in this position indicates that
      optional VLAN tagging according to [IEEE802.1Q] is employed, and
      that the next two bytes contain the VLP, C and VLAN ID fields.
      VLAN tags may be stacked, in which case the two-byte field

Stein, et al.             Expires June 8, 2007                 [Page 16]

Internet-Draft                   TDMoIP                    December 2006

      following the VLAN ID is once again a VLAN Ethertype.

   VLP (3 bits) is the VLAN priority, see [IEEE802.1Q].

   C  (1 bit) the "canonical format indicator" being set, indicates that
      route descriptors appear; see [IEEE802.1Q].

   VLAN ID (12 bits) the VLAN identifier uniquely identifies the VLAN to
      which the frame belongs.  If zero only the VLP information is
      meaningful.  Values 1 and FFF are reserved.  The other 4193 values
      are valid VLAN identifiers.

   Ethertype (16 bits) is the protocol identifier, as allocated by the
      IEEE.  The Ethertype SHOULD be set to 0x88D8 (35032), but MAY be
      set to 0x8847 (34887).

   PW Label (20 bits) This label MUST be manually configured.  The
      remainder of this row is formatted to resemble an MPLS label.

   Frame Check Sequence (32 bits) is a CRC error detection field,
      calculated per [IEEE802.3].

5.  TDMoIP Payload types

   As discussed at the end of Section 3, TDMoIP transports real-time
   streams by first extracting bytes from the stream, and then adapting
   these bytes.  TDMoIP offers two different adaptation algorithms, one
   for constant rate real-time traffic, and one for variable rate real-
   time traffic.

   For unstructured TDM, or structured but unchannelized TDM, or
   structured channelized TDM with all channels active all the time, a
   constant rate adaptation is needed.  In such cases TDMoIP uses
   structure-indication to emulate the native TDM circuit, and the
   adaptation is known as "circuit emulation".  However, for channelized
   TDM wherein the individual channels (corresponding to "loops" in
   telephony terminology) are frequently inactive, bandwidth may be
   conserved by transporting only active channels.  This results in
   variable rate real-time traffic, for which TDMoIP uses structure-
   reassembly to emulate the individual loops, and the adaptation is
   known as "loop emulation".

   TDMoIP uses constant-rate AAL1 [AAL1,CES] for circuit emulation,
   while variable-rate AAL2 [AAL2] is employed for loop emulation.  The
   AAL1 mode MUST be used for structured transport of unchannelized data
   and SHOULD be used for circuits with relatively constant usage.  In

Stein, et al.             Expires June 8, 2007                 [Page 17]

Internet-Draft                   TDMoIP                    December 2006

   addition, AAL1 MUST be used when the TDM-bound IWF is required to
   maintain a high timing accuracy (e.g. when its timing is further
   distributed) and SHOULD be used when high reliability is required.
   AAL2 SHOULD be used for channelized TDM when bandwidth needs to be
   conserved, and MAY be used whenever usage of voice-carrying channels
   is expected to be highly variable.

   Additionally, a third mode is defined specifically for efficient
   transport of HDLC-based CCS signaling carried in TDM channels.

   The AAL family of protocols is a natural choice for TDM emulation.
   Although originally developed to adapt various types of application
   data to the rigid format of ATM, the mechanisms are general solutions
   to the problem of transporting constant or variable rate real-time
   streams over a packet network.

   Since the AAL mechanisms are extensively deployed within and on the
   edge of the public telephony system, they have been demonstrated to
   reliably transfer voice-grade channels, data and telephony signaling.
   These mechanisms are mature and well understood, and implementations
   are readily available.

   Finally, simplified service interworking with legacy networks is a
   major design goal of TDMoIP.  Re-use of AAL technologies simplifies
   interworking with existing AAL1- and AAL2-based networks.

5.1  AAL1 Format Payload

   For the prevalent cases of unchannelized TDM, or channelized TDM for
   which the channel allocation is static, the payload can be
   efficiently encoded using constant rate AAL1 adaptation.  The AAL1
   format is described in [AAL1] and its use for circuit emulation over
   ATM in [CES].  We briefly review highlights of AAL1 technology in
   Appendix B.  In this section we describe the use of AAL1 in the
   context of TDMoIP.

                        |control word |    AAL1 PDU    |

               Figure 7a.  Single AAL1 PDU per TDMoIP Packet

             +-------------+----------------+   +----------------+
             |control word |    AAL1 PDU    |---|    AAL1 PDU    |
             +-------------+----------------+   +----------------+

             Figure 7b.  Multiple AAL1 PDUs per TDMoIP Packet

Stein, et al.             Expires June 8, 2007                 [Page 18]

Internet-Draft                   TDMoIP                    December 2006

   In AAL1 mode the TDMoIP payload consists of at least one, and perhaps
   many, 48-byte "AAL1 PDUs", see Figures 7a and 7b.  The number of PDUs
   MUST be pre-configured and MUST be chosen such that the overall
   packet size does not exceed the maximum allowed by the PSN (e.g. 30
   for UDP/IP over Ethernet).  The precise number of PDUs per packet is
   typically chosen taking latency and bandwidth constraints into
   account.  Using a single PDU delivers minimal latency, but incurs the
   highest overhead.  All TDMoIP implementations MUST support between 1
   and 8 PDUs per packet for E1 and T1 circuits, and between 5 and 15
   PDUs per packet for E3 and T3 circuits.

   AAL1 differentiates between unstructured and structured data
   transfer, which correspond to structure-agnostic and structure-aware
   transport.  For structure-agnostic transport, AAL1 provides no
   inherent advantage as compared to SAToP; however, there may be
   scenarios for which its use is desirable.  For example, when it is
   necessary to interwork with an existing AAL1 ATM circuit emulation
   system, or when clock recovery based on AAL1-specific mechanisms is

   For structure-aware transport, [CES] defines two modes, structured
   and structured with CAS.  Structured AAL1 maintains TDM frame
   synchronization by embedding a pointer to the beginning of the next
   frame in the AAL1 PDU header.  Similarly, structured AAL1 with CAS
   maintains TDM frame and multiframe synchronization by embedding a
   pointer to the beginning of the next multiframe.  Furthermore,
   structured AAL1 with CAS contains a substructure including the CAS
   signaling bits.

5.2  AAL2 Format Payload

   Although AAL1 may be configured to transport fractional E1 or T1
   circuits, the allocation of channels to be transported must be static
   due to the fact that AAL1 transports constant rate bit-streams.  It
   is often the case that not all the channels in a TDM circuit are
   simultaneously active ("off-hook"), and by observation of the TDM
   signaling channel activity status may be determined.  Moreover, even
   during active calls about half the time is silence that can be
   identified using voice activity detection (VAD).  Using the variable
   rate AAL2 mode we may dynamically allocate channels to be
   transported, thus conserving bandwidth.

   The AAL2 format is described in [AAL2] and its use for loop emulation
   over ATM is explained in [SSCS,LES].  We briefly review highlights of
   AAL2 technology in Appendix C.  In this section we describe the use
   of AAL2 in the context of TDMoIP.

Stein, et al.             Expires June 8, 2007                 [Page 19]

Internet-Draft                   TDMoIP                    December 2006

             +-------------+----------------+   +----------------+
             |control word |    AAL2 PDU    |---|    AAL2 PDU    |
             +-------------+----------------+   +----------------+

         Figure 8.  Concatenation of AAL2 PDUs in a TDMoIP packet

   In AAL2 mode the TDMoIP payload consists of one or more variable-
   length "AAL2 PDUs", see Figure 8.  Each AAL2 PDU contains 3 bytes of
   overhead and between 1 and 64 bytes of payload.  A packet may be
   constructed by inserting PDUs corresponding to all active channels,
   by appending PDUs ready at a certain time, or by any other means.
   Hence, more than one PDU belonging to a single channel may appear in
   a packet.

   [RFC3985] denotes as Native Service Processing (NSP) functions all
   processing of the TDM data before its use as payload.  Since AAL2 is
   inherently variable rate, arbitrary NSP functions MAY be performed
   before the channel is placed in the AAL2 loop emulation payload.
   These include testing for on-hook/off-hook status, voice activity
   detection, speech compression, fax/modem/tone relay, etc.

   All mechanisms described in [AAL2,SSCS,LES] may be used for TDMoIP.
   In particular, CID encoding and use of PAD octets according to
   [AAL2], encoding formats defined in [SSCS], and transport of CAS and
   CCS signaling as described in [LES] MAY all be used in the PSN-bound
   direction, and MUST be supported in the TDM-bound direction.  The
   overlap functionality and AAL-CU timer and related functionalities
   may not be required, and the STF field is NOT used.  Computation of
   error detection codes, namely the HEC in the AAL2 PDU header and the
   CRC in the CAS packet, is superfluous if an appropriate error
   detection mechanism is provided by the PSN.  In such cases these
   fields MAY be set to zero.

5.3  HDLC Format Payload

   The motivation for handling HDLC in TDMoIP is to efficiently
   transport common channel signaling (CCS) such as SS7 [SS7] or ISDN
   PRI signaling [ISDN-PRI], embedded in the TDM stream.  This mechanism
   is not intended for general HDLC payloads, and assumes that the HDLC
   messages are always shorter than the maximum packet size.

   The HDLC mode should only be used when the majority of the bandwidth
   of the input HDLC stream is expected to be occupied by idle flags.
   Otherwise the CCS channel should be treated as an ordinary channel.

   The HDLC format is intended to operate in port mode, transparently
   passing all HDLC data and control messages over a separate PW.  The

Stein, et al.             Expires June 8, 2007                 [Page 20]

Internet-Draft                   TDMoIP                    December 2006

   encapsulation is compatible with that of [RFC4618], however the
   sequence number generation and processing SHOULD be according to
   section 3 above.

   The PSN-bound IWF monitors flags until a frame is detected.  The
   contents of the frame are collected and the FCS tested.  If the FCS
   is incorrect the frame is discarded, otherwise the frame is sent
   after initial or final flags and FCS have been discarded and zero
   removal has been performed.  When an TDMoIP- HDLC frame is received
   its FCS is recalculated, and the original HDLC frame reconstituted.

6.  TDMoIP Defect Handling

   Native TDM networks signify network faults by carrying indications of
   forward defects (AIS) and reverse defects (RDI) in the TDM bit
   stream.  Structure-agnostic TDM transport transparently carries all
   such indications; however, for structure-aware mechanisms where the
   PSN-bound IWF may remove TDM structure overhead carrying defect
   indications, explicit signaling of TDM defect conditions is required.

   We saw in Section 3 that defects can be indicated by setting flags in
   the control word.  This insertion of defect reporting into the packet
   rather than in a separate stream mimics the behavior of native TDM
   OAM mechanisms that carry such indications as bit patterns embedded
   in the TDM stream.  The flags are designed to address the urgent
   messaging, i.e. messages whose contents must not be significantly
   delayed with respect to the TDM data that they potentially impact.
   Mechanisms for slow OAM messaging are discussed in Appendix D.

    +---+   +-----+   +------+   +-----+   +------+   +-----+   +---+
    |TDM|->-|     |->-|TDMoIP|->-|     |->-|TDMoIP|->-|     |->-|TDM|
    |   |   |TDM 1|   |      |   | PSN |   |      |   |TDM 2|   |   |
    |ES1|-<-|     |-<-| IWF1 |-<-|     |-<-| IWF2 |-<-|     |-<-|ES2|
    +---+   +-----+   +------+   +-----+   +------+   +-----+   +---+

              Figure 9.  Typical TDMoIP Network Configuration

   The operation of TDMoIP defect handling is best understood by
   considering the downstream TDM flow from TDM end system 1 (ES1)
   through TDM network 1, through TDMoIP IWF 1 (IWF1), through the PSN,
   through TDMoIP IWF 2 (IWF2), through TDM network 2, towards TDM end
   system 2 (ES2), as depicted in the figure.  We wish not only to
   detect defects in TDM network 1, the PSN, and TDM network 2, but to
   localize such defects in order to raise alarms only in the
   appropriate network.

   In the "trail terminated" OAM scenario, only user data is exchanged

Stein, et al.             Expires June 8, 2007                 [Page 21]

Internet-Draft                   TDMoIP                    December 2006

   between TDM network 1 and TDM network 2.  The IWF functions as a TDM
   trail termination function, and defects detected in TDM network 1 are
   not relayed to network 2, or vice versa.

   In the "trail extended" OAM scenario, if there is a defect (e.g. loss
   of signal or loss of frame synchronization) anywhere in TDM network 1
   before the ultimate link, the following TDM node will generate AIS
   downstream (towards TDMoIP IWF1).  If a break occurs in the ultimate
   link, the IWF itself will detect the loss of signal.  In either case,
   IWF1 having directly detected lack of validity of the TDM signal, or
   having been informed of an earlier problem, raises the local ("L")
   defect flag in the control word of the packets it sends across the
   PSN.  In this way the trail is extended to TDM network 2 across the

   Unlike forward defect indications that are generated by all network
   elements, reverse defect indications are only generated by trail
   termination functions.  In the trail terminated scenario, IWF1 serves
   as a trail termination function for TDM network 1, and thus when IWF1
   directly detects lack of validity of the TDM signal, or is informed
   of an earlier problem, it MAY generate TDM RDI towards TDM ES1.  In
   the trail extended scenario IWF1 is not a trail termination, and
   hence MUST NOT generate TDM RDI, but rather, as we have seen, sets
   the "L" defect flag.  As we shall see, this will cause the AIS
   indication to reach ES2, which is the trail termination, and which
   MAY generate TDM RDI.

   When the "L" flag is set there are four possibilities for treatment
   of payload content.  The default is for IWF1 to fill the payload with
   the appropriate amount of AIS (usually all-ones) data.  If the AIS
   has been generated before the IWF this can be accomplished by copying
   the received TDM data; if the penultimate TDM link fails and the IWF
   needs to generate the AIS itself.  Alternatively, with structure-
   aware transport of channelized TDM one SHOULD fill the payload with
   "trunk conditioning"; this involves placing a preconfigured "out of
   service" code in each individual channel (the "out of service" code
   may differ between voice and data channels).  Trunk conditioning MUST
   be used when channels taken from several TDM PWs are combined by the
   TDM-bound IWF into a single TDM circuit.  The third possibility is to
   suppress the payload altogether.  Finally, if IWF1 believes that the
   TDM defect is minor or correctable (e.g. loss of multiframe
   synchronization, or initial phases of detection of incorrect frame
   sync), it MAY place the TDM data it has received into the payload
   field, and specify in the defect modification field ("M") that the
   TDM data is corrupted, but potentially recoverable.

   When IWF2 receives a local defect indication without "M"-field
   modification, it forwards (or generates if the payload has been

Stein, et al.             Expires June 8, 2007                 [Page 22]

Internet-Draft                   TDMoIP                    December 2006

   suppressed) AIS or trunk conditioning towards ES2 (the choice between
   AIS and conditioning being preconfigured).  Thus AIS has been
   properly delivered to ES2 emulating the TDM scenario from the TDM end
   system's point of view.  In addition, IWF2 receiving the "L"
   indication uniquely specifies that the defect was in TDM network 1
   and not in TDM network 2, thus suppressing alarms in the correctly
   functioning network.

   If the M field indicates that the TDM has been marked as potentially
   recoverable, then implementation specific algorithms (not herein
   specified) may optionally be utilized to minimize the impact of
   transient defects on the overall network performance.  If the "M"
   field indicates that the TDM is "idle", no alarms should be raised
   and IWF2 treats the payload contents as regular TDM data.  If the
   payload has been suppressed, trunk conditioning and not AIS MUST be
   generated by IWF2.

   The second case is when the defect is in TDM network 2.  Such defects
   cause AIS generation towards ES2, which may respond by sending TDM
   RDI in the reverse direction.  In the trail terminated scenario this
   RDI is restricted to network 2.  In the trail extended scenario, IWF2
   upon observing this RDI inserted into valid TDM data, MUST indicate
   this by setting the "R" flag in packets sent back across the PSN
   towards IWF1.  IWF1, upon receiving this indication, generates RDI
   towards ES1, thus emulating a single conventional TDM network.

   The final possibility is that of a unidirectional defect in the PSN.
   In such a case TDMoIP IWF1 sends packets toward IWF2, but these are
   not received.  IWF2 MUST inform the PSN's management system of this
   problem, and furthermore generate TDM AIS towards ES2.  ES2 may
   respond with TDM RDI, and as before, in the trail extended scenario,
   when IWF2 detects RDI it MUST raise the "R" flag indication.  When
   IWF1 receives packets with the "R" flag set it has been informed of a
   reverse defect, and MUST generate TDM RDI towards ES1.

   In all cases, if any of the above defects persist for a preconfigured
   period (default value of 2.5 seconds) a service failure is declared.
   Since TDM PWs are inherently bidirectional, a persistent defect in
   either directional results in a bidirectional service failure.  In
   addition, if signaling is sent over a distinct PW as per Section 5.3,
   both PWs are considered to have failed when persistent defects are
   detected in either.

   When failure is declared the PW MUST be withdrawn, and both TDMoIP
   IWFs commence sending AIS (and not trunk conditioning) to their
   respective TDM networks.  The IWFs then engage in connectivity
   testing using VCCV or TDMoIP OAM as described in Appendix D until
   connectivity is restored.

Stein, et al.             Expires June 8, 2007                 [Page 23]

Internet-Draft                   TDMoIP                    December 2006

7.  Implementation Issues

   General requirements for transport of TDM over pseudo-wires are
   detailed in [RFC4197].  In the following subsections we review
   additional aspects essential to successful TDMoIP implementation.

7.1  Jitter and Packet Loss

   In order to compensate for packet delay variation that exists in any
   PSN, a jitter buffer MUST be provided.  A jitter buffer is a block of
   memory into which the data from the PSN is written at its variable
   arrival rate, and data is read out and sent to the destination TDM
   equipment at a constant rate.  Use of a jitter buffer partially hides
   the fact that a PSN has been traversed rather than a conventional
   synchronous TDM network, except for the additional latency.
   Customary practice is to operate with the jitter buffer approximately
   half full, thus minimizing the probability of its overflow or
   underflow.  Hence the additional delay equals half the jitter buffer
   size.  The length of the jitter buffer SHOULD be configurable and MAY
   be dynamic (i.e. grow and shrink in length according to the
   statistics of the PDV).

   In order to handle (infrequent) packet loss and misordering a packet
   sequence integrity mechanism MUST be provided.  This mechanism MUST
   track the serial numbers of arriving packets and MUST take
   appropriate action when anomalies are detected.  When missing
   packet(s) are detected the mechanism MUST output filler packet(s) in
   order to retain TDM timing.  Packets arriving in incorrect order
   SHOULD be reordered.  Processing of filler packets SHOULD ensure that
   proper FAS is sent to the TDM network.  An example sequence number
   processing algorithm is provided in Appendix A.

   While the insertion of arbitrary filler packets may be sufficient to
   maintain the TDM timing, for telephony traffic it may lead to gaps or
   artifacts that result in choppy, annoying or even unintelligible
   audio.  An implementation MAY blindly insert a preconfigured constant
   value in place of any lost samples, and this value SHOULD be chosen
   to minimize the perceptual effect.  Alternatively one MAY replay the
   previously received packet.  When computational resources are
   available, implementations SHOULD conceal the packet loss event by
   properly estimating missing sample values in such fashion as to
   minimize the perceptual error.

7.2  Timing Recovery

   TDM networks are inherently synchronous; somewhere in the network
   there will always be at least one extremely accurate primary
   reference clock, with long-term accuracy of one part in 1E-11.  This

Stein, et al.             Expires June 8, 2007                 [Page 24]

Internet-Draft                   TDMoIP                    December 2006

   node provides reference timing to secondary nodes with somewhat lower
   accuracy, and these in turn distribute timing information further.
   This hierarchy of time synchronization is essential for the proper
   functioning of the network as a whole; for details see [G823][G824].

   Packets in PSNs reach their destination with delay that has a random
   component, known as packet delay variation (PDV).  When emulating TDM
   on a PSN, extracting data from the jitter buffer at a constant rate
   overcomes much of the high frequency component of this randomness
   ("jitter").  The rate at which we extract data from the jitter buffer
   is determined by the destination clock, and were this to be precisely
   matched to the source clock proper timing would be maintained.
   Unfortunately the source clock information is not disseminated
   through a PSN, and the destination clock frequency will only
   nominally equal the source clock frequency, leading to low frequency
   ("wander") timing inaccuracies.

   In broadest terms there are four methods of overcoming this
   difficulty.  In the first and second methods timing information is
   provided by some means independent of the PSN.  This timing may be
   provided to the TDM end systems (method 1) or to the IWFs (method 2).
   In a third method a common clock is assumed available to both IWFs,
   and the relationship between the TDM source clock and this clock is
   encoded in the packet.  This encoding may be take the form of RTP
   timestamps or may utilizing the SRTS bits in the AAL1 overhead.  In
   the final method (adaptive clock recovery) the timing must be deduced
   solely based on the packet arrival times.  Example scenarios are
   detailed in [RFC4197] and in [Y1413].

   Adaptive clock recovery utilizes only observable characteristics of
   the packets arriving from the PSN, such as the precise time of
   arrival of the packet at the TDM-bound IWF, or the fill-level of the
   jitter buffer as a function of time.  Due to the packet delay
   variation in the PSN, filtering processes that combat the statistical
   nature of the observable characteristics must be employed.  Frequency
   Locked Loops (FLL) and Phase Locked Loops (PLL) are well suited for
   this task.

   Whatever timing recovery mechanism is employed, the output of the
   TDM-bound IWF MUST conform to the jitter and wander specifications of
   TDM traffic interfaces, as defined in [G823][G824].  For some
   applications, more stringent jitter and wander tolerances MAY be

7.3  Congestion Control

   As explained in [RFC3985], the underlying PSN may be subject to
   congestion.  Unless appropriate precautions are taken, undiminished

Stein, et al.             Expires June 8, 2007                 [Page 25]

Internet-Draft                   TDMoIP                    December 2006

   demand of bandwidth by TDMoIP can contribute to network congestion
   that may impact network control protocols.

   The AAL1 mode of TDMoIP is an inelastic constant bit-rate (CBR) flow
   and cannot respond to congestion in a TCP-friendly manner prescribed
   by [RFC2914], although the percentage of total bandwidth they consume
   remains constant.  The AAL2 mode of TDMoIP is variable bit-rate
   (VBR), and it is often possible to reduce the bandwidth consumed by
   employing mechanisms that are beyond the scope of this document.

   Whenever possible, TDMoIP SHOULD be carried across traffic-
   engineered PSNs that provide either bandwidth reservation and
   admission control or forwarding prioritization and boundary traffic
   conditioning mechanisms.  IntServ-enabled domains supporting
   Guaranteed Service (GS) [RFC2212] and DiffServ-enabled domains
   [RFC2475] supporting Expedited Forwarding (EF) [RFC3246] provide
   examples of such PSNs.  Such mechanisms will negate, to some degree,
   the effect of TDMoIP on neighboring streams.  In order to facilitate
   boundary traffic conditioning of TDMoIP traffic over IP PSNs, the
   TDMoIP packets SHOULD NOT use the DiffServ Code Point (DSCP) value
   reserved for the Default Per-Hop Behavior (PHB) [RFC2474].

   If TDMoIP run over a PSN providing best-effort service, packet loss
   SHOULD be monitored in order to detect congestion.  If congestion is
   detected and bandwidth reduction is possible, then such reduction
   SHOULD be enacted.  If bandwidth reduction is not possible the TDMoIP
   PW SHOULD shut down bi-directionally for some period of time as
   described in Section 6.5 of [RFC3985].

   Note that:

      1.  In AAL1 mode TDMoIP can inherently provide packet loss
      measurement since the expected rate of packet arrival is fixed and

      2.  The results of the packet loss measurement may not be a
      reliable indication of presence or absence of severe congestion if
      the PSN provides enhanced delivery.  For example, if TDMoIP
      traffic takes precedence over other traffic, severe congestion may
      not significantly affect TDMoIP packet loss.

      3.  The TDM services emulated by TDMoIP have high availability
      objectives (see [G.826]) that MUST be taken into account when
      deciding on temporary shutdown.

   This specification does not define exact criteria for detecting
   severe congestion or specific methods for TDMoIP shutdown or
   subsequent re-start.  However, the following considerations may be

Stein, et al.             Expires June 8, 2007                 [Page 26]

Internet-Draft                   TDMoIP                    December 2006

   used as guidelines for implementing the shutdown mechanism:

      1.  If the TDMoIP PW has been set up using the PWE3 control
      protocol [RFC4447], the regular PW teardown procedures of these
      protocols SHOULD be used.

      2.  If one of the TDMoIP IWFs stops transmission of packets for a
      sufficiently long period, its peer (observing 100% packet loss)
      will necessarily detect "severe congestion" and also stop
      transmission, thus achieving bi-directional PW shutdown.

   TDMoIP does not provide mechanisms to ensure timely delivery or
   provide other quality-of-service guarantees; hence it is required
   that the lower-layer services do so.  Layer 2 priority can be
   bestowed upon a TDMoIP stream by using the VLAN priority field, MPLS
   priority can be provided by using EXP bits, and layer 3 priority is
   controllable by using TOS.  Switches and routers which the TDMoIP
   stream must traverse should be configured to respect these

8.  Security Considerations

   TDMoIP does not enhance or detract from the security performance of
   the underlying PSN, rather it relies upon the PSN's mechanisms for
   encryption, integrity, and authentication whenever required.  The
   level of security provided may be less than that of a native TDM

   When MPLS is the PSN, PW-specific security mechanisms will generally
   be required, while for IP-based PSNs IPsec MAY be used.  TDMoIP using
   L2TPv3 is subject to the security considerations discussed in section
   4.1.3 of [RFC3931].

   TDMoIP shares susceptibility to a number of pseudowire-layer attacks
   (see [RFC3985]) and implementations SHOULD use whatever mechanisms
   for confidentiality, integrity, and authentication are developed for
   general PWs.  These methods are beyond the scope of this document.

   Random initialization of sequence numbers, in both the control word
   and the optional RTP header, makes known-plaintext attacks on
   encrypted TDMoIP more difficult.  Encryption of PWs is beyond the
   scope of this document.

   PW labels SHOULD be selected in an unpredictable manner rather than
   sequentially or otherwise in order to deter session hijacking.  When
   using L2TPv3, randomly selected cookies MAY be used to validate
   circuit origin.

Stein, et al.             Expires June 8, 2007                 [Page 27]

Internet-Draft                   TDMoIP                    December 2006

   Although TDMoIP MAY employ an RTP header when explicit transfer of
   timing information is required, SRTP (see [RFC3711]) mechanisms are
   not applicable.

9.  IANA Considerations

   For MPLS PSNs, PW Types for TDMoIP PWs are allocated in [RFC4446].

   For UDP/IP PSNs, when the source port is used to identify the TDM
   stream, the destination port number MUST be set to 0x085E (2142), the
   user port number assigned by IANA to TDMoIP.

10.  Applicability Statement

   It must be recognized that the emulation provided by TDMoIP may be
   imperfect, and the service may differ from of the native TDM circuit
   in the following ways.

   The end-to-end delay of a TDM circuit emulated using TDMoIP may
   exceed that of a native TDM circuit.

   When using adaptive clock recovery, the timing performance of the
   emulated TDM circuit depends on characteristics of the PSN, and thus
   may be inferior to that of a native TDM circuit.

   If the TDM structure overhead is not transported over the PSN, then
   non-FAS data in the overhead will be lost.

   When packets are lost in the PSN TDMoIP mechanisms ensure that frame
   synchronization will be maintained, and when packet loss events are
   properly concealed the effect on telephony channels will be
   perceptually minimized.  However, the bit error rate will be degraded
   as compared to the native service.

   Data in inactive channels is not transported in AAL2 mode, and thus
   this data will differ from that of the native service.

   Native TDM connections are point-to-point, while PSNs are shared
   infrastructures.  Hence the level of security of the emulated service
   may be less than that of the native service.

11.  Acknowledgments

   The authors would like to thank Hugo Silberman, Shimon HaLevy, Tuvia
   Segal, and Eitan Schwartz of RAD Data Communications for their
   invaluable contributions to the technology described herein.

Stein, et al.             Expires June 8, 2007                 [Page 28]

Internet-Draft                   TDMoIP                    December 2006

Authors' Addresses

   Yaakov (Jonathan) Stein
   RAD Data Communications
   24 Raoul Wallenberg St., Bldg C
   Tel Aviv  69719

   Phone: +972 3 645-5389

   Ronen Shashoua
   RAD Data Communications
   24 Raoul Wallenberg St., Bldg C
   Tel Aviv  69719

   Phone: +972 3 645-5447

   Ron Insler
   RAD Data Communications
   24 Raoul Wallenberg St., Bldg C
   Tel Aviv  69719

   Phone: +972 3 645-5445

   Motty (Mordechai) Anavi
   RAD Data Communications
   900 Corporate Drive
   Mahwah, NJ  07430

   Phone: +1 201 529-1100 Ext. 213

Stein, et al.             Expires June 8, 2007                 [Page 29]

Internet-Draft                   TDMoIP                    December 2006

Appendix A.  Sequence Number Processing (informative)

   The sequence number field in the control word enables detection of
   lost and misordered packets.  Here we give pseudocode for an example
   algorithm in order to clarify the issues involved.  These issues are
   implementation specific and no single explanation can capture all the

   In order to simplify the description modulo arithmetic is
   consistently used in lieu of ad-hoc treatment of the cyclicity.  All
   differences between indexes are explicitly converted to the range
   [-2^15 ... +2^15 - 1] to ensure that simple checking of the
   difference's sign correctly predicts the packet arrival order.

   Furthermore, we introduce the notion of a playout buffer in order to
   unambiguously define packet lateness.  When a packet arrives after
   having previously having been assumed lost, the TDM-bound IWF may
   discard it, and continue to treat it as lost.  Alternatively if the
   filler data that had been inserted in its place has not yet been
   played out, the option remains to insert the true data into the
   playout buffer.  Of course, the filler data may be generated upon
   initial detection of a missing packet or upon playout.  This
   description is stated in terms of a packet-oriented playout buffer
   rather than a TDM byte oriented one; however this is not a true
   requirement for re-ordering implementations since the latter could be
   used along with pointers to packet commencement points.

   Having introduced the playout buffer we explicitly treat over-run and
   under-run of this buffer.  Over-run occurs when packets arrive so
   quickly that they can not be stored for playout.  This is usually an
   indication of gross timing inaccuracy or misconfiguration, and we can
   do little but discard such early packets.  Under-run is usually a
   sign of network starvation, resulting from congestion or network

   The external variables used by the pseudocode are:

      received:  sequence number of packet received
      played:    sequence number of the packet being played out (Note 1)
      over-run:  is the playout buffer full? (Note 3)
      under-run: has the playout buffer been exhausted? (Note 3)

   The internal variables used by the pseudocode are:

      expected: sequence number we expect to receive next
      D: difference between expected and received (Note 2)
      L: difference between sequence numbers of packet being played out
         and that just received (Notes 1 and 2)

Stein, et al.             Expires June 8, 2007                 [Page 30]

Internet-Draft                   TDMoIP                    December 2006

   In addition, the algorithm requires one parameter:

      R: maximum lateness of packet recoverable (Note 1).

     Note 1: this is only required for the optional re-ordering
     Note 2: this number is always in the range -2^15 ... +2^15 - 1
     Note 3: the playout buffer is emptied by the TDM playout process,
             which runs asynchronously to the packet arrival processing,
             and which is not herein specified

   Sequence Number Processing Algorithm

   Upon receipt of a packet
     if received = expected
       { treat packet as in-order }
       if not over-run then
         place packet contents into playout buffer
         discard packet contents
       set expected = (received + 1) mod 2^16
       calculate D = ( (expected-received) mod 2^16 ) - 2^15
       if D > 0 then
         { packets expected, expected+1, ... received-1 are lost }
         while not over-run
           place filler (all-ones or interpolation) into playout buffer
           if not over-run then
             place packet contents into playout buffer
             discard packet contents
           set expected = (received + 1) mod 2^16
       else  { late packet arrived }
         declare "received" to be a late packet
         do NOT update "expected"
           discard packet
           if not under-run then
             calculate L = ( (played-received) mod 2^16 ) - 2^15
             if 0 < L <= R then
               replace data from packet previously marked as lost
               discard packet
   Note: by choosing R=0 we always discard the late packet

Stein, et al.             Expires June 8, 2007                 [Page 31]

Internet-Draft                   TDMoIP                    December 2006

Appendix B.  AAL1 Review (informative)

   The first byte of the 48-byte AAL1 PDU always contains an error-
   protected three-bit sequence number.

                    1 2 3 4 5 6 7 8
                   |C| SN  | CRC |P| 47 bytes of payload

   C  (1 bit) convergence sublayer indication, its use here is limited
      to indication of the existence of a pointer (see below); C=0 means
      no pointer, C=1 means a pointer is present.

   SN (3 bits) The AAL1 sequence number increments from PDU to PDU.

   CRC (3 bits) is a 3 bit error cyclic redundancy code on C and SN.

   P  (1 bit) even byte parity.

   As can be readily inferred, incrementing the sequence number forms an
   eight PDU sequence number cycle, the importance of which will become
   clear shortly.

   The structure of the remaining 47 bytes in the AAL1 PDU depends on
   the PDU type, of which there are three, corresponding to the three
   types of AAL1 circuit emulation service defined in [CES].  These are
   known as namely unstructured circuit emulation, structured circuit
   emulation and structured circuit emulation with CAS.

   The simplest PDU is the unstructured one, which is used for
   transparent transfer of whole circuits (T1,E1,T3,E3).  Although AAL1
   provides no inherent advantage as compared to SAToP for unstructured
   transport, in certain cases AAL1 may be required or desirable.  For
   example, when it is necessary to interwork with an existing AAL1-
   based network, or when clock recovery based on AAL1-specific
   mechanisms is favored.

   For unstructured AAL1 the 47 bytes after the sequence number byte
   contain the full 376 bits from the TDM bit stream.  No frame
   synchronization is supplied or implied, and framing is the sole
   responsibility of the end-user equipment.  Hence the unstructured
   mode can be used to carry data, and for circuits with nonstandard
   frame synchronization.  For the T1 case the raw frame consists of 193
   bits, and hence 1 183/193 T1 frames fit into each AAL1 PDU.  The E1
   frame consists of 256 bits, and so 1 15/32 E1 frames fit into each

Stein, et al.             Expires June 8, 2007                 [Page 32]

Internet-Draft                   TDMoIP                    December 2006

   When the TDM circuit is channelized according to [G704], and in
   particular when it is desired to fractional E1 or T1, it is
   advantageous to use one of the structured AAL1 circuit emulation
   services.  Structured AAL1 views the data not merely as a bit stream,
   but as a bundle of channels.  Furthermore, when CAS signaling is used
   it can be formatted so that it can be readily detected and

   In the structured circuit emulation mode without CAS, N bytes from
   the N channels to be transported are first arranged in order of
   channel number.  Thus if channels 2, 3, 5, 7 and 11 are to be
   transported the corresponding five bytes are placed in the PDU
   immediately after the sequence number byte.  This placement is
   repeated until all 47 bytes in the PDU are taken;

        byte     1  2  3  4  5  6  7  8  9 10 --- 41 42 43 44 45 46 47
        channel  2  3  5  7 11  2  3  5  7 11 ---  2  3  5  7 11  2  3

   the next PDU commences where the present PDU left off

        byte     1  2  3  4  5  6  7  8  9 10 --- 41 42 43 44 45 46 47
        channel  5  7 11  2  3  5  7 11  2  3 ---  5  7 11  2  3  5  7

   and so forth.  The set of channels 2,3,5,7,11 is the basic structure
   and the point where one structure ends and the next commences is the
   structure boundary.

   The problem with this arrangement is the lack of explicit indication
   of the byte identities.  As can be seen in the above example, each
   AAL1 PDU starts with a different channel, so a single lost packet
   will result in misidentifying channels from that point onwards,
   without possibility of recovery.  The solution to this deficiency is
   the periodic introduction of a pointer to the next structure
   boundary.  This pointer need not be used too frequently, as the
   channel identifications are uniquely inferable unless packets are

   The particular method used in AAL1 is to insert a pointer once every
   sequence number cycle of eight PDUs.  The pointer is seven bits and
   protected by an even parity MSB, and so occupies a single byte.
   Since seven bits are sufficient to represent offsets larger than 47,
   we can limit the placement of the pointer byte to PDUs with even
   sequence number.  Unlike most AAL1 PDUs that contain 47 TDM bytes,
   PDUs that contain a pointer (P-format PDUs) have the following

Stein, et al.             Expires June 8, 2007                 [Page 33]

Internet-Draft                   TDMoIP                    December 2006

            0                 1
            1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6
           |C| SN  | CRC |P|E|   pointer   | 46 bytes of payload


   C  (1 bit) convergence sublayer indication, C=1 for P-format PDUs.

   SN (3 bits) is an even AAL1 sequence number.

   CRC (3 bits) is a 3 bit error cyclic redundancy code on C and SN.

   P  (1 bit) even byte parity LSB for sequence number byte.

   E  (1 bit) even byte parity MSB for pointer byte.

   pointer (7 bits) pointer to next structure boundary.

   Since P-format PDUs have 46 bytes of payload and the next PDU has 47
   bytes, viewed as a single entity the pointer needs to indicate one of
   93 bytes.  If P=0 it is understood that the structure commences with
   the following byte (i.e. the first byte in the payload belongs to the
   lowest numbered channel).  P=93 means that the last byte of the
   second PDU is the final byte of the structure, and the following PDU
   commences with a new structure.  The special value P=127 indicates
   that there is no structure boundary to be indicated (needed when
   extremely large structures are being transported).

   The P-format PDU is always placed at the first possible position in
   the sequence number cycle that a structure boundary occurs, and can
   only occur once per cycle.

   The only difference between the structured circuit emulation format
   and structured circuit emulation with CAS is the definition of the
   structure.  Whereas in structured circuit emulation the structure is
   composed of the N channels, in structured circuit emulation with CAS
   the structure encompasses the superframe consisting of multiple
   repetitions of the N channels and then the CAS signaling bits.  The
   CAS bits are tightly packed into bytes and the final byte is padded
   with zeros if required.

   For example, for E1 circuits the CAS signaling bits are updated once
   per superframe of 16 frames.  Hence the structure for N*64 derived
   from an E1 with CAS signaling consists of 16 repetitions of N bytes,
   followed by N sets of the four ABCD bits, and finally four zero bits
   if N is odd.  For example, the structure for channels 2,3 and 5 will

Stein, et al.             Expires June 8, 2007                 [Page 34]

Internet-Draft                   TDMoIP                    December 2006

   be as follows

       2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 2 3 5
       2 3 5 2 3 5 2 3 5 2 3 5 2 3 5 [ABCD2 ABCD3] [ABCD5 0000]

   Similarly for T1 ESF circuits the superframe is 24 frames, and the
   structure consists of 24 repetitions of N bytes, followed by the ABCD
   bits as before.  For the T1 case the signaling bits will in general
   appear twice, in their regular (bit-robbed) positions and at the end
   of the structure.

Stein, et al.             Expires June 8, 2007                 [Page 35]

Internet-Draft                   TDMoIP                    December 2006

Appendix C.  AAL2 Review (informative)

   The basic AAL2 PDU is :

         |    Byte  1    |    Byte  2    |    Byte  3    |
          0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
         |      CID      |     LI    |   UUI   |   HEC   |   PAYLOAD

   CID (8 bits) channel identifier is an identifier that must be unique
      for the PW.  The values 0-7 are reserved for special purposes,
      (and if interworking with VoDSL is required, so are values 8
      through 15 as specified in [LES]), thus leaving 248 (240) CIDs per
      PW.  The mapping of CID values to channels MAY be manually
      configured manually or signaled.

   LI (6 bits) length indicator is one less than the length of the
      payload in bytes.  Note that the payload is limited to 64 bytes.

   UUI (5 bits) user-to-user indication is the higher layer
      (application) identifier and counter.  For voice data the UUI will
      always be in the range 0-15, and SHOULD be incremented modulo 16
      each time a channel buffer is sent.  The receiver MAY monitor this
      sequence.  UUI is set to 24 for CAS signaling packets.

   HEC (5 bits) the header error control

   Payload - voice A block of length indicated by LI of voice samples
      are placed as- is into the AAL2 packet.

   Payload - CAS signaling For CAS signaling the payload is formatted as
      an AAL2 "fully protected" (type 3) packet (see [AAL2]) in order to
      ensure error protection.  The signaling is sent with the same CID
      as the corresponding voice channel.  Signaling MUST be sent
      whenever the state of the ABCD bits changes, and SHOULD be sent
      with triple redundancy, i.e. sent three times spaced 5
      milliseconds apart.  In addition, the entire set of the signaling
      bits SHOULD be sent periodically to ensure reliability.

                       |RED|       timestamp           |
                       |  RES  | ABCD  |    type   | CRC
                           CRC (cont)  |

Stein, et al.             Expires June 8, 2007                 [Page 36]

Internet-Draft                   TDMoIP                    December 2006

   RED (2 bits) is the triple redundancy counter.  For the first packet
      it takes the value 00, for the second 01 and for the third 10.
      RED=11 means non-redundant information, and is used when triple
      redundancy is not employed, and for periodic refresh messages.

   Timestamp (14 bits) The timestamp is optional and in particular is
      not needed if RTP is employed.  If not used the timestamp MUST be
      set to zero.  When used with triple redundancy it MUST be the same
      for all three redundant transmissions.

   RES (4 bits) is reserved and MUST be set to zero.

   ABCD (4 bits) are the CAS signaling bits.

   type (6 bits) for CAS signaling this is 000011.

   CRC-10 (10 bits) is a 10 bit CRC error detection code.

Stein, et al.             Expires June 8, 2007                 [Page 37]

Internet-Draft                   TDMoIP                    December 2006

Appendix D.  Performance Monitoring Mechanisms (informative)

   PWs require OAM mechanisms to monitor performance measures that
   impact the emulated service.  Performance measures, such as packet
   loss ratio and packet delay variation, may be used to set various
   parameters and thresholds; for TDMoIP PWs adaptive timing recovery
   and packet loss concealment algorithms may benefit from such
   information.  In addition, OAM mechanisms may be used to collect
   statistics relating to the underlying PSN [RFC2330], and its
   suitability for carrying TDM services.

   TDMoIP IWFs may benefit from knowledge of PSN performance metrics,
   such as round trip time (RTT), packet delay variation (PDV) and
   packet loss ratio (PLR).  These measurements are conventionally
   performed by a separate flow of packets designed for this purpose,
   e.g.  ICMP packets [RFC792] or MPLS LSP ping packets [RFC4379] with
   multiple timestamps.  For AAL1 mode TDMoIP sends packets across the
   PSN at a constant rate, and hence no additional OAM flow is required
   for measurement of PDV or PLR.  However, separate OAM flows are
   required for RTT measurement, for AAL2 mode PWs, for measurement of
   parameters at setup, for monitoring of inactive backup PWs, and for
   low-rate monitoring of PSNs after PWs have been withdrawn due to
   service failures.

   If the underlying PSN has appropriate maintenance mechanisms that
   provide connectivity verification, RTT, PDV, and PLR measurements
   that correlate well with those of the PW, then these mechanisms
   SHOULD be used.  If such mechanisms are not available, either of two
   similar OAM signaling mechanisms may be used.  The first is internal
   to the PW and based on inband VCCV [VCCV], and the second is defined
   only for UDP/IP PSNs, and is based on a separate PW.  The latter is
   particularly efficient for a large number of fate-sharing TDM PWs.

D.1  TDMoIP Connectivity Verification

   In most conventional IP applications a server sends some finite
   amount of information over the network after explicit request from a
   client.  With TDMoIP PWs the PSN-bound IWF could send a continuous
   stream of packets towards the destination without knowing whether the
   TDM-bound IWF is ready to accept them.  For layer-2 networks this may
   lead to flooding of the PSN with stray packets.

   This problem may occur when a TDMoIP IWF is first brought up, when
   the TDM-bound IWF fails or is disconnected from the PSN, or the PW is
   broken.  After an aging time the destination IWF becomes unknown, and
   intermediate switches may flood the network with the TDMoIP packets
   in an attempt to find a new path.

Stein, et al.             Expires June 8, 2007                 [Page 38]

Internet-Draft                   TDMoIP                    December 2006

   The solution to this problem is to significantly reduce the number of
   TDMoIP packets transmitted per second when PW failure is detected,
   and to return to full rate only when the PW is available.  The
   detection of failure and restoration is made possible by the periodic
   exchange of one-way connectivity-verification messages.

   Connectivity is tested by periodically sending OAM messages from the
   source IWF to the destination IWF, and having the destination reply
   to each message.  The connectivity verification mechanism SHOULD be
   used during setup and configuration.  Without OAM signaling one must
   ensure that the destination IWF is ready to receive packets before
   starting to send them.  Since TDMoIP IWFs operate full-duplex, both
   would need to be set up and properly configured simultaneously if
   flooding is to be avoided.  When using connectivity verification, a
   configured IWF may wait until it detects its peer before transmitting
   at full rate.  In addition, configuration errors may be readily
   discovered by using the service specific field of the OAM PW packets.

   In addition to one way connectivity, OAM signaling mechanisms can be
   used to request and report on various PSN metrics, such as one way
   delay, round trip delay, packet delay variation, etc.  They may also
   be used for remote diagnostics, and for unsolicited reporting of
   potential problems (e.g. dying gasp messages).

D.2  OAM Packet Format

   When using inband performance monitoring, additional packets are sent
   using the same PW label.  These packets are identified by having
   their first nibble equal to 0001, and must be separated from TDM data
   packets before further processing of the control word.

   When using a separate OAM PW, all OAM messages MUST use the PW label
   preconfigured to indicate OAM.  All PSN layer parameters MUST remain
   those of the PW being monitored.

   The format of an inband OAM PW message packet for UDP/IP PSNs is
   based on [RFC2679].  The PSN-specific layers are identical to those
   defined in Section 4.1 with the PW label set to the value
   preconfigured or assigned for PW OAM.

Stein, et al.             Expires June 8, 2007                 [Page 39]

Internet-Draft                   TDMoIP                    December 2006

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       |         PSN-specific layers  (with preconfigured PW label)    |
       |0 0 0 0|L|R| M |RES| Length    |     OAM Sequence Number       |
       | OAM Msg Type  | OAM Msg Code  | Service specific information  |
       |       Forward PW label        |      Reverse PW label         |
       |                   Source Transmit Timestamp                   |
       |                 Destination Receive Timestamp                 |
       |                Destination Transmit Timestamp                 |

   L, R, and M are identical to those of the PW being tested.

   Length is the length in bytes of the OAM message packet.

   OAM Sequence Number (16 bits) is used to uniquely identify the
      message.  Its value is unrelated to the sequence number of the
      TDMoIP data packets for the PW in question.  It is incremented in
      query messages, and replicated without change in replies.

   OAM Msg Type (8 bits) indicates the function of the message.  At
      present the following are defined:

             0 for one way connectivity query message
             8 for one way connectivity reply message.

   OAM Msg Code (8 bits) is used to carry information related to the
      message, and its interpretation depends on the message type.  For
      type 0 (connectivity query) messages the following codes are

             0 validate connection.
             1 do not validate connection

      for type 8 (connectivity reply) messages the available codes are:

             0 acknowledge valid query
             1 invalid query (configuration mismatch).

Stein, et al.             Expires June 8, 2007                 [Page 40]

Internet-Draft                   TDMoIP                    December 2006

   Service specific information (16 bits) is a field that can be used to
      exchange configuration information between IWFs.  If it is not
      used this field MUST contain zero.  Its interpretation depends on
      the payload type.  At present the following is defined for AAL1

                        0                   1
                        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
                       | Number of TSs | Number of SFs |

   Number of TSs (8 bits) is the number of channels being transported,
      e.g. 24 for full T1.

   Number of SFs (8 bits) is the number of 48-byte AAL1 PDUs per packet,
      e.g. 8 when packing 8 PDUs per packet.

   Forward PW label (16 bits) is the PW label used for TDMoIP traffic
      from the source to destination IWF.

   Reverse PW label (16 bits) is the PW label used for TDMoIP traffic
      from the destination to source IWF.

   Source Transmit Timestamp (32 bits) represents the time the PSN-bound
      IWF transmitted the query message.  This field and the following
      ones only appear if delay is being measured.  All time units are
      derived from a clock of preconfigured frequency, the default being
      100 microseconds.

   Destination Receive Timestamp (32 bits) represents the time the
      destination IWF received the query message.

   Destination Transmit Timestamp (32 bits) represents the time the
      destination IWF transmitted the reply message.

Stein, et al.             Expires June 8, 2007                 [Page 41]

Internet-Draft                   TDMoIP                    December 2006

Appendix E.  Capabilities, Configuration and Statistics (informative)

   Every TDMoIP IWF will support some number of physical TDM
   connections, certain types of PSN, and some subset of the modes
   defined above.  The following capabilities SHOULD be able to be
   queried by the management system:

      AAL1 capable

      AAL2 capable (and AAL2 parameters, e.g. support for VAD and

      HDLC capable

      Supported PSN types (UDP/IPv4, UDP/IPv6, L2TPv3/IPv4, L2TPv3/IPv6,
      MPLS, Ethernet)

      OAM support (none, separate PW, VCCV) and capabilities (CV, delay
      measurement, etc.)

      maximum packet size supported.

   For every TDM PW the following parameters MUST be provisioned or

      PW label (for UDP and Ethernet the label MUST be manually

      TDM type  (E1, T1, E3, T3, fractional E1, fractional T1)

         for fractional links: number of timeslots

      TDMoIP mode (AAL1, AAL2, HDLC)

      for AAL1 mode:

         AAL1 type (unstructured, structured, structured with CAS)

         number of AAL1 PDUs per packet

      for AAL2 mode:

         CID mapping

         creation time of full minicell (units of 125 microsecond)

Stein, et al.             Expires June 8, 2007                 [Page 42]

Internet-Draft                   TDMoIP                    December 2006

      size of jitter buffer (in 32-bit words)

      clock recovery method (local, loop-back timing, adaptive, common

      use of RTP (if used: frequency of common clock, PT and SSRC

   During operation the following statistics and impairment indications
   SHOULD be collected for each TDM PW, and can be queried by the
   management system.

      average round-trip delay

      packet delay variation (maximum delay - minimum delay)

      number of potentially lost packets

      indication of misordered packets (succesfully reordered or

      for AAL1 mode PWs:

         indication of malformed PDUs (incorrect CRC, bad C, P or E)

         indication of cells with pointer mismatch

         number of seconds with jitter buffer over-run events

         number of seconds with jitter buffer under-run events

      for AAL2 mode PWs:

         number of malformed minicells (incorrect HEC)

         indication of misordered minicells (unexpected UUI)

         indication of stray minicells (CID unknown, illegal UUI)

         indication of mis-sized minicells (unexpected LI)

         for each CID: number of seconds with jitter buffer over-run

      for HDLC mode PWs:

         number of discarded frames from TDM (e.g.  CRC error, illegal
         packet size)

Stein, et al.             Expires June 8, 2007                 [Page 43]

Internet-Draft                   TDMoIP                    December 2006

         number of seconds with jitter buffer over-run events.

   During operation the following statistics MAY be collected for each
   TDM PW.

      number of packets sent to PSN

      number of packets received from PSN

      number of seconds during which packets were received with L flag

      number of seconds during which packets were received with R flag

Stein, et al.             Expires June 8, 2007                 [Page 44]

Internet-Draft                   TDMoIP                    December 2006

Appendix F.  References

F.1  Normative References

   [AAL1] ITU-T Recommendation I.363.1 (08/96) - B-ISDN ATM Adaptation
      Layer (AAL) specification: Type 1

   [AAL2] ITU-T Recommendation I.363.2 (11/00) - B-ISDN ATM Adaptation
      Layer (AAL) specification: Type 2

   [CES] ATM forum specification atm-vtoa-0078 (CES 2.0) Circuit
      Emulation Service Interoperability Specification Ver. 2.0

   [G704] ITU-T Recommendation G.704 (10/98) - Synchronous frame
      structures used at 1544, 6312, 2048, 8448 and 44736 kbit/s
      hierarchical levels

   [G751] ITU-T Recommendation G.751 (11/88) - Digital multiplex
      equipments operating at the third order bit rate of 34368 kbit/s
      and the fourth order bit rate of 139264 kbit/s and using positive

   [G823] ITU-T Recommendation G.823 (03/00) - The control of jitter and
      wander within digital networks which are based on the 2048 Kbit/s

   [G824] ITU-T Recommendation G.824 (03/00) - The control of jitter and
      wander within digital networks which are based on the 1544 Kbit/s

   [G826] ITU-T Recommendation G.826 (13/02) - End-to-end error
      performance parameters and objectives for international, constant
      bit-rate digital paths and connections

   [IEEE802.1Q] IEEE 802.1Q, IEEE Standards for Local and Metropolitan
      Area Networks -- Virtual Bridged Local Area Networks (2003)

   [IEEE802.3] IEEE 802.3, IEEE Standard Local and Metropolitan Area
      Networks - Carrier Sense Multiple Access with Collision Detection
      (CSMA/CD) Access Method and Physical Layer Specifications (2002)

   [LES] ATM forum specification atm-vmoa-0145 (LES) Voice and
      Multimedia over ATM - Loop Emulation Service Using AAL2

   [MEF8] Metro Ethernet Forum, "Implementation Agreement for the
      Emulation of PDH Circuits over Metro Ethernet Networks, October

Stein, et al.             Expires June 8, 2007                 [Page 45]

Internet-Draft                   TDMoIP                    December 2006

   [RFC768] Postel, J., "User Datagram Protocol (UDP)", STD 6, RFC 768,
      August 1980.

   [RFC791] Postel, J., "Internet Protocol (IP)", STD 5, RFC 791,
      September 1981.

   [RFC2119] Bradner, S., "Key Words in RFCs to Indicate Requirement
      Levels", RFC 2119, March 1997.

   [RFC3032] Rosen, E., Tappan, D., Fedorkow, G., Rekhter, Y.,
      Farinacci, D., Li, T., and A. Conta, "MPLS Label Stack Encoding",
      RFC 3032, January 2001.

   [RFC3931] Lau, J., Townsley, M., Goyret, I., "Layer Two Tunneling
      Protocol - Version 3 (L2TPv3)", RFC 3931, March 2005.

   [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and Jacobson,
      V., "RTP: A Transport Protocol for Real-Time Applications", STD
      64, RFC 3550, July 2003.

   [RFC4446] Martini, L., "IANA Allocations for Pseudowire Edge to Edge
      Emulation (PWE3)", BCP 116, RFC 4446, April 2006.

   [RFC4447] Martini, L., Rosen, E., El-Aawar, N., Smith, T., and G.
      Heron, "Pseudowire Setup and Maintenance Using the Label
      Distribution Protocol (LDP)", RFC 4447, April 2006.

   [RFC4553] Vainshtein A., and Stein YJ., "Structure-Agnostic TDM over
      Packet (SAToP)", RFC 4553, June 2006.

   [RFC4618] Martini L., Rosen E., Heron G., and Malis A.,
      "Encapsulation Methods for Transport of PPP/High-Level Data Link
      Control (HDLC) over MPLS Networks", RFC 4618, September 2006.

   [SSCS] ITU-T Recommendation I.366.2 (11/00) - AAL type 2 service
      specific convergence sublayer for narrow-band services.

   [VCCV] draft-ietf-pwe3-vccv-06.txt (08/05) - Pseudo Wire Virtual
      Circuit Connectivity Verification, T. Nadeau and R. Aggarwal, work
      in progress.

   [Y1413] ITU-T Recommendation Y.1413 (03/04) - TDM-MPLS network
      interworking - User plane interworking

   [Y1414] ITU-T Recommendation Y.1414 (07/04) - Voice services - MPLS
      network interworking.

Stein, et al.             Expires June 8, 2007                 [Page 46]

Internet-Draft                   TDMoIP                    December 2006

   [Y1452] ITU-T Recommendation Y.1452 (03/06) - Voice trunking over IP

   [Y1453] ITU-T Recommendation Y.1453 (03/06) - TDM-IP interworking -
      User plane interworking.

F.2  Informative References

   [CESoPSN] draft-ietf-cesopsn-07.txt, TDM Circuit Emulation Service
      over Packet Switched Network, A. Vainshtein et al, work in
      progress, May 2006.

   [ISDN-PRI] ITU-T Recommendation Q.931 (05/98) - ISDN user-network
      interface layer 3 specification for basic call control.

   [RFC792] Postel J., "Internet Control Message Protocol", STD 5, RFC
      792, September 1981.

   [RFC2212] Shenker, S., Partridge, C., and R. Guerin, "Specification
      of Guaranteed Quality of Service", RFC 2212, September 1997.

   [RFC2330] Paxson, V., Almes, G., Mahdavi, J., Mathis M., "Framework
      for IP Performance Metrics", RFC 2330, May 1998.

   [RFC2474] Nichols, K., Blake, S., Baker, F., and D. Black,
      "Definition of the Differentiated Services Field (DS Field) in the
      IPv4 and IPv6 Headers", RFC 2474, December 1998.

   [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z.,
      and W. Weiss, "An Architecture for Differentiated Service", RFC
      2475, December 1998.

   [RFC2679] Almes, G., Kalidindi, S., and M. Zekauskas, "A One-way
      Delay Metric for IPPM", RFC 2679, September 1999.

   [RFC2914] Floyd, S., "Congestion Control Principles", BCP 41, RFC
      2914, September 2000.

   [RFC3246] Davie, B., Charny, A., Bennet, J.C., Benson, K., Le Boudec,
      J., Courtney, W., Davari, S., Firoiu, V., and D. Stiliadis, "An
      Expedited Forwarding PHB (Per-Hop Behavior)", RFC 3246, March

   [RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K.
      Norrman, "The Secure Real-time Transport Protocol (SRTP)", RFC
      3711, March 2004.

Stein, et al.             Expires June 8, 2007                 [Page 47]

Internet-Draft                   TDMoIP                    December 2006

   [RFC3985] Bryant, S. and P. Pate, "Pseudo Wire Emulation Edge- to-
      Edge (PWE3) Architecture", RFC 3985, March 2005.

   [RFC4197] Riegel, M., "Requirements for Edge-to-Edge Emulation of
      Time Division Multiplexed (TDM) Circuits over Packet Switching
      Networks", RFC 4197, October 2005.

   [RFC4379] Kompella, K. and Swallow, G., "Detecting Multi-Protocol
      Label Switched (MPLS) Data Plane Failures", RFC 4379, February

   [RFC4385] Bryant, S., Swallow, G., Martini, L., and D. McPherson,
      "Pseudowire Emulation Edge-to-Edge (PWE3) Control Word for Use
      over an MPLS PSN", RFC 4385, February 2006.

   [SS7] ITU-T Recommendation Q.700 (03/93) - Introduction to CCITT
      Signalling System No. 7.

   [TDM-CONTROL] Vainshtein, A. and Stein, Y(J), "Control Protocol
      Extensions for Setup of TDM Pseudowires", Work in Progress, July

   [TRAU] GSM 08.60 (10/01) - Digital cellular telecommunications system
      (Phase 2+); Inband control of remote transcoders and rate adaptors
      for Enhanced Full Rate (EFR) and full rate traffic channels.

Stein, et al.             Expires June 8, 2007                 [Page 48]

Internet-Draft                   TDMoIP                    December 2006

Intellectual Property Statement

   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at

Disclaimer of Validity

   This document and the information contained herein are provided on an

Copyright Statement

   Copyright (C) The Internet Society (2006).  This document is subject
   to the rights, licenses and restrictions contained in BCP 78, and
   except as set forth therein, the authors retain all their rights.


   Funding for the RFC Editor function is currently provided by the
   Internet Society.

Stein, et al.             Expires June 8, 2007                 [Page 49]