| Rfc | 5143 |
|---|
| Title | Synchronous Optical Network/Synchronous Digital Hierarchy
(SONET/SDH) Circuit Emulation Service over MPLS (CEM) Encapsulation |
|---|
| Author | A. Malis, J. Brayley, J. Shirron, L. Martini, S. Vogelsang |
|---|
| Date | February
2008 |
|---|
| Format: | TXT, HTML |
|---|
| Obsoleted by | RFC4842 |
|---|
| Status: | HISTORIC |
|---|
|
Network Working Group A. Malis
Request for Comments: 5143 Verizon Communications
Obsoleted by: 4842 J. Brayley
Category: Historic J. Shirron
ECI Telecom Inc.
L. Martini
Cisco Systems, Inc.
S. Vogelsang
Alcatel-Lucent
February 2008
Synchronous Optical Network/Synchronous Digital Hierarchy (SONET/SDH)
Circuit Emulation Service over MPLS (CEM) Encapsulation
Status of This Memo
This memo defines a Historic Document for the Internet community. It
does not specify an Internet standard of any kind. Distribution of
this memo is unlimited.
IESG Note
The IESG thinks that this work is related to IETF work done in WG
PWE3, but this does not prevent publishing.
Abstract
This document describes a historical method for encapsulating
Synchronous Optical Network/Synchronous Digital Hierarchy (SONET/SDH)
Path signals for transport across packet-switched networks (PSNs).
The PSNs explicitly supported by this document include MPLS and IP.
Note that RFC 4842 describes the standards-track protocol for this
functionality, and new implementations must use RFC 4842 rather than
this document except when interoperability with older implementations
is desired.
Table of Contents
1. Introduction ....................................................3
2. Conventions Used in This Document ...............................3
3. Scope ...........................................................3
4. CEM Encapsulation Format ........................................4
4.1. Transport Encapsulation ....................................6
4.1.1. MPLS Transport ......................................6
4.1.2. IP Transport ........................................7
5. CEM Operation ...................................................8
5.1. Introduction and Terminology ...............................8
5.1.1. CEM Packetizer and De-Packetizer ....................9
5.1.2. CEM DBA .............................................9
5.2. Description of Normal CEM Operation .......................10
5.3. Description of CEM Operation during DBA ...................10
5.4. Packet Synchronization ....................................11
5.4.1. Acquisition of Packet Synchronization ..............11
5.4.2. Loss of Packet Synchronization .....................11
6. SONET/SDH Maintenance Signals ..................................12
6.1. SONET/SDH to PSN ..........................................12
6.1.1. AIS-P Indication ...................................13
6.1.2. STS SPE Unequipped Indication ......................14
6.1.3. CEM-RDI ............................................14
6.2. PSN to SONET/SDH ..........................................15
6.2.1. AIS-P Indication ...................................15
6.2.2. STS SPE Unequipped Indication ......................15
7. Clocking Modes .................................................16
7.1. Synchronous ...............................................16
7.1.1. Synchronous Unstructured CEM .......................16
7.1.2. Synchronous Structured CEM .........................16
7.2. Asynchronous ..............................................17
8. CEM LSP Signaling ..............................................17
9. Security Considerations ........................................18
10. IANA Considerations ...........................................18
11. References ....................................................18
11.1. Normative References .....................................18
11.2. Informative References ...................................19
Appendix A. SONET/SDH Rates and Formats ...........................20
Appendix B. ECC-6 Definition ......................................21
1. Introduction
This document describes a historical method for encapsulating
SONET/SDH Path signals for transport across packet-switched networks
(PSNs).
The native transmission system for circuit-oriented Time Division
Multiplexing (TDM) signals is the Synchronous Optical Network (SONET)
[T1.105], [GR-253]/Synchronous Digital Hierarchy (SDH) [G.707]. To
support TDM traffic (which includes voice, data, and private leased
line services), PSNs must emulate the circuit characteristics of
SONET/SDH payloads. MPLS labels and a new circuit emulation header
are used to encapsulate TDM signals and provide the Circuit Emulation
Service over MPLS (CEM) function. The MPLS encapsulation may be
further encapsulated in IP for carriage across IP PSNs [RFC4023].
This document also describes an optional extension to CEM called
Dynamic Bandwidth Allocation (DBA). This is a method for dynamically
reducing the bandwidth utilized by emulated SONET/SDH circuits in the
packet network. This bandwidth reduction is accomplished by not
sending the SONET/SDH payload through the packet network under
certain conditions, such as Alarm Indication Signal - Path (AIS-P) or
Synchronous Transport Signal Synchronous Payload Envelope (STS SPE)
Unequipped.
2. Conventions Used in This Document
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC2119].
3. Scope
This document describes how to provide CEM for the following digital
signals:
1. SONET STS-1 synchronous payload envelope (SPE)/SDH VC-3
2. STS-Nc SPE (N = 3, 12, or 48)/SDH VC-4, VC-4-4c, VC-4-16c
3. Unstructured SONET Emulation, where the entire SONET bit-stream
(including the transport overhead) is packetized and transported
across the PSN.
For the remainder of this document, these constructs will be referred
to as SONET/SDH channels.
Other SONET/SDH signals, such as virtual tributary (VT) structured
sub-rate mapping, are not explicitly discussed in this document;
however, it can be extended in the future to support VT and lower
speed non-SONET/SDH services. OC-192c SPE/VC-4-64c are also not
included at this point, since most PSNs use OC-192c or slower trunks,
and thus would not have sufficient capacity. As trunk capacities
increase in the future, the scope of this document can be accordingly
extended.
4. CEM Encapsulation Format
In order to transport SONET/SDH SPEs through a packet-oriented
network, the SPE is broken into fragments. A 32-bit CEM header is
pre-pended to each fragment. The Basic CEM packet appears in Figure
1.
+-----------------------------------+
| CEM Header |
+-----------------------------------+
| |
| |
| SONET/SDH SPE Fragment |
| |
| |
+-----------------------------------+
Figure 1. Basic CEM Packet
The 32-bit CEM header has the following format:
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|D|R|Rvd| Sequence Num | Structure Pointer |N|P| ECC-6 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 2. CEM Header Format
The above fields are defined as follows:
D-bit: This bit signals DBA Mode. It MUST be set to zero for normal
operation, and it MUST be set to one if CEM is currently in DBA mode.
DBA is an optional mode during which trivial SPEs are not transmitted
into the packet network. See Table 1 and sections 7 and 8 for
further details.
Note: for unstructured CEM, the D-bit MUST be set to zero.
R bit: CEM-RDI (Remote Defect Indicator). This bit is set to one to
signal to the remote CEM function that a loss of packet
synchronization has occurred.
Rvd: These bits are reserved for future use, and MUST be set to zero.
Sequence Number: This is a packet sequence number, which MUST
continuously cycle from 0 to 1023. It SHOULD begin at zero when a
CEM LSP (Label Switched Path) is created.
Structure Pointer: The Structure Pointer MUST contain the offset of
the J1 byte within the CEM payload. The value is from 0 to 1,022,
where 0 means the first byte after the CEM header. The Structure
Pointer MUST be set to 0x3FF (1,023) if a packet does not carry the
J1 byte. See [T1.105], [G.707], and [GR-253] for more information On
the J1 byte and the SONET/SDH payload pointer.
Note: for unstructured CEM, the Structure Pointer field MUST be
set to 0x3FF.
The N and P bits: These bits indicate negative and positive pointer
adjustment events. They are also used to relay SONET/SDH maintenance
signals, such as AIS-P. See Table 1 and sections 7 and 8 for more
details.
Note: for unstructured CEM, the N and P bits MUST both be set to
zero.
+---+---+---+----------------------------------------------+
| D | N | P | Interpretation |
+---+---+---+-------------+--------------------------------+
| 0 | 0 | 0 | Normal Mode | No Ptr Adjustment |
| 0 | 0 | 1 | Normal Mode | Positive Ptr Adjustment |
| 0 | 1 | 0 | Normal Mode | Negative Ptr Adjustment |
| 0 | 1 | 1 | Normal Mode | AIS-P |
| | | | | |
| 1 | 0 | 0 | DBA Mode | STS SPE Unequipped |
| 1 | 0 | 1 | DBA Mode | STS SPE Unequipped Pos Ptr Adj |
| 1 | 1 | 0 | DBA Mode | STS SPE Unequipped Neg Ptr Adj |
| 1 | 1 | 1 | DBA Mode | AIS-P |
+---+---+---+-------------+--------------------------------+
Table 1. Interpretation of D, N, and P bits
ECC-6: An Error Correction Code to protect the CEM header. This
offers the ability to correct single bit errors and detect up to two
bit errors. The ECC algorithm is described in Appendix B. The ECC-6
can be optionally disabled at provisioning time. If the ECC-6 is not
utilized, it MUST be set to zero.
Note: Normal CEM packets are fixed in length for all of the
packets of a particular emulated TDM stream. This length is
signaled using the CEM Payload Bytes parameter defined in
[RFC4447], or is statically provisioned for each TDM stream.
Therefore, the length of each CEM packet does not need to be
carried in the CEM header.
Note: Setting the D-bit to 0 and the R bit to 1 violates the Best
Current Practice defined in [RFC4928] when operating on MPLS
networks. In this situation, MPLS networks could mistake a CEM
payload as the first nibble of an IPv4 packet, potentially causing
mis-ordering of packets on the pseudowire in the presence of IP
Equal Cost Multi-Path (ECMP) in the MPLS network. The use of this
CEM header preceded the use of MPLS ECMP. As stated earlier,
[RFC4842] describes the standards-track protocol for this
functionality, and it does not share this violation.
4.1. Transport Encapsulation
In principle, CEM packets can be transported over any packet-oriented
network. The following sections describe specifically how CEM
packets MUST be encapsulated for transport over MPLS or IP networks.
4.1.1. MPLS Transport
To transport a CEM packet over an MPLS network, an MPLS label stack
MUST be pushed on top of the CEM packet.
The last two labels prior to the CEM header are referred to as the
Tunnel and Virtual Circuit (VC) labels.
The VC label is required, and is the last label prior to the CEM
Header. The VC label MUST be used to identify the CEM connection
within the MPLS tunnel.
The optional tunnel label is immediately above the VC label on the
label stack. If present, the tunnel label MUST be used to identify
the MPLS LSP used to tunnel the TDM packets through the MPLS network
(the tunnel LSP).
This is similar to the label stack usage defined in [RFC4447].
+-----------------------------------+
| Additional MPLS Labels (Optional) |
+-----------------------------------+
| Tunnel Label (Optional) |
+-----------------------------------+
| VC Label |
+-----------------------------------+
| CEM Header |
+-----------------------------------+
| |
| |
| SONET/SDH SPE Fragment |
| |
| |
+-----------------------------------+
Figure 3. Typical MPLS Transport Encapsulation
4.1.2. IP Transport
It is highly desirable to define a single encapsulation format that
will work for both IP and MPLS. Furthermore, it is desirable that
the encapsulation mechanism be as efficient as possible.
One way to achieve these goals is to map CEM directly onto IP by
mapping the previously described MPLS packets onto IP.
A mechanism for carrying MPLS over IP is described in [RFC4023].
Using this encapsulation scheme would result in the packet format
illustrated in Figure 4.
+-----------------------------------+
| |
| IPv6/v4 Header [RFC4023] |
| |
+-----------------------------------+
| Tunnel Label (Optional) |
+-----------------------------------+
| VC Label |
+-----------------------------------+
| CEM Header |
+-----------------------------------+
| |
| |
| SONET/SDH SPE Fragment |
| |
| |
+-----------------------------------+
Figure 4. MPLS Transport Encapsulation
5. CEM Operation
The following sections describe CEM operation.
5.1. Introduction and Terminology
There are two types of CEM: structured and unstructured.
Unstructured CEM packetizes the entire SONET/SDH bit-stream
(including transport overhead).
Structured CEM terminates the transport overhead and packetizes
individual channels (STS-1/Nc) within the SONET/SDH frame. Because
structured CEM terminates the transport overhead, structured CEM
implementations SHOULD meet the generic requirements for SONET/SDH
Line Terminating Equipment as defined in [T1.105], [G.707], and
[GR-253].
Implementations MUST support structured CEM and MAY support
unstructured CEM.
Structured CEM MUST support a normal mode of operation and MAY
support an optional extension called Dynamic Bandwidth Allocation
(DBA). During normal operation, SONET/SDH payloads are fragmented,
pre-pended with the CEM header, the VC label, and the PSN header, and
then transmitted into the packet network. During DBA mode, only the
CEM header, the VC label, and PSN header are transmitted. This is
done to conserve bandwidth when meaningful user data is not present
in the SPE, such as during AIS-P or STS SPE Unequipped.
5.1.1. CEM Packetizer and De-Packetizer
As with all adaptation functions, CEM has two distinct components:
adapting TDM SONET/SDH into a CEM packet stream, and converting the
CEM packet stream back into a TDM SONET/SDH. The first function will
be referred to as CEM packetizer and the second as CEM de-packetizer.
This terminology is illustrated in Figure 5.
+------------+ +---------------+
| | | |
SONET --> | CEM | --> PSN --> | CEM | --> SONET
SDH | Packetizer | | De-Packetizer | SDH
| | | |
+------------+ +---------------+
Figure 5. CEM Terminology
Note: the CEM de-packetizer requires a buffering mechanism to account
for delay variation in the CEM packet stream. This buffering
mechanism will be generically referred to as the CEM jitter buffer.
5.1.2. CEM DBA
DBA is an optional mode of operation for structured CEM that only
transmits the CEM header, the VC label, and PSN header into the
packet network under certain circumstances, such as AIS-P or STS SPE
Unequipped.
If DBA is supported by a CEM implementation, the user SHOULD be able
to configure if DBA will be triggered by AIS-P, STS SPE Unequipped,
both, or neither on a per channel basis.
If DBA is supported, the determination of AIS-P and STS SPE
Unequipped MUST be based on the state of SONET/SDH Section, Line, and
Path Overhead bytes. DBA based on pattern detection within the SPE
(i.e., all zeros, 7Es, or ATM idle cells) is for further study.
During AIS-P, there is no valid payload pointer, so pointer
adjustments cannot occur. During STS SPE Unequipped, the SONET/SDH
payload pointer is valid, and therefore pointer adjustments MUST be
supported even during DBA. See Table 1 for details.
5.2. Description of Normal CEM Operation
During normal operation, the CEM packetizer will receive a fixed rate
byte stream from a SONET/SDH interface. When a packet's worth of
data has been received from a SONET/SDH channel, the CEM header, the
VC Label, and PSN header are pre-pended to the SPE fragment and the
resulting CEM packet is transmitted into the packet network. Because
all normal CEM packets associated with a specific SONET/SDH channel
will have the same length, the transmission of CEM packets for that
channel SHOULD occur at regular intervals.
At the far-end of the packet network, the CEM de-packetizer will
receive packets into a jitter buffer and then play out the received
byte stream at a fixed rate onto the corresponding SONET/SDH channel.
The jitter buffer SHOULD be adjustable in length to account for
varying network delay behavior. The received packet rate from the
packet network should be exactly balanced by the transmission rate
onto the SONET/SDH channel, on average. The time over which this
average is taken corresponds to the depth of the jitter buffer for a
specific CEM channel.
The CEM sequence numbers provide a mechanism to detect lost and/or
mis-ordered packets. The CEM de-packetizer MUST detect lost or
mis-ordered packets. The CEM de-packetizer MUST play out a
programmable byte pattern in place of any dropped packets. The CEM
de-packetizer MAY re-order packets received out of order. If the CEM
de-packetizer does not support re-ordering, it MUST drop mis-ordered
packets.
5.3. Description of CEM Operation during DBA
(Note: DBA is only applicable to structured CEM.)
There are several issues that should be addressed by a workable CEM
DBA mechanism. First, when DBA is invoked, there should be a
substantial savings in bandwidth utilization in the packet network.
The second issue is that the transition in and out of DBA should be
tightly coordinated between the local CEM packetizer and CEM
de-packetizer at the far side of the packet network. A third is that
the transition in and out of DBA should be accomplished with minimal
disruption to the adapted data stream.
Another goal is that the reduction of CEM traffic due to DBA should
not be mistaken for a fault in the packet network or vice-versa.
Finally, the implementation of DBA should require minimal
modifications beyond what is necessary for the nominal CEM case. The
mechanism described below is a reasonable balance of these goals.
During DBA, packets MUST be emitted at exactly the same rate as they
would be during normal operation. This SHOULD be accomplished by
transmitting each DBA packet after a complete packet of data has been
received from the SONET/SDH channel. The only change from normal
operation is that the CEM packets during DBA MUST only carry the CEM
header, the VC label, and the PSN header.
Because some links have a minimum supported packet size, the CEM
packetizer MAY append a configurable number of bytes immediately
after the CEM header to pad out the CEM packet to reach the minimum
supported packet size. The value of the padding bytes is
implementation specific. The D-bit MUST be set to one, to indicate
that DBA is active.
The CEM de-packetizer MUST assume that each packet received with the
D-bit set represents a normal-sized packet containing an AIS-P or STS
SPE Unequipped payload as noted by N and P, (see Table 1). The CEM
de-packetizer MUST accept DBA packets with or without padding.
This allows the CEM packetization and de-packetization logic during
DBA to be similar to the nominal case. It insures that the correct
SONET/SDH indication is reliably transmitted between CEM adaptation
points. It minimizes the risk of under or over running the jitter
buffer during the transition in and out of DBA. And, it guarantees
that faults in the packet network are recognized as distinctly
different from line conditioning on the SONET/SDH interfaces.
5.4. Packet Synchronization
A key component in declaring the state of a CEM service is whether or
not the CEM de-packetizer is in or out of packet synchronization.
The following paragraphs describe how that determination is made.
5.4.1. Acquisition of Packet Synchronization
At startup, a CEM de-packetizer will be out of packet synchronization
by default. To declare packet synchronization at startup or after a
loss of packet synchronization, the CEM de-packetizer must receive a
configurable number of CEM packets with sequential sequence numbers.
5.4.2. Loss of Packet Synchronization
Once a CEM de-packetizer is in packet sync, it may encounter a set of
events that will cause it to lose packet synchronization.
As discussed in section 5.2, a CEM de-packetizer MAY support the
re-ordering of mis-ordered packets.
If a CEM de-packetizer supports re-ordering, then the determination
that packet synchronization has been lost cannot be made at the time
the packets are received from the PSN. Instead, the determination
MUST be made as the packets are being played out onto the SONET/SDH
interface. This is because it is only at play-out time that the
determination can be made as to whether the entire emulated SONET/SDH
stream was received from the PSN.
If a CEM de-packetizer does not support re-ordering, a number of
approaches may be used to minimize the impact of mis-ordered or lost
packets on the final re-assembled SONET/SDH stream. For example, ATM
Adaptation Layer 1 (AAL1) [I.363.1] uses a simple state-machine to
re-order packets in a subset of possible cases. The algorithm for
these state-machines is outside of the scope of CEM. However, the
final determination as to whether or not to declare loss of packet
synchronization MUST be based on the same criteria as for
implementations that do support re-ordering.
Whether or not a CEM implementation supports re-ordering, the
declaration of loss of packet synchronization MUST be based on the
following criteria.
As packets are played out towards the SONET/SDH interface, the CEM
de-packetizer will encounter empty packets in the place of packets
that were dropped by the PSN, or effectively dropped due to
limitations of the CEM implementation. If the CEM de-packetizer
encounters more than a configurable number of sequential dropped
packets, the CEM de-packetizer MUST declare loss of packet
synchronization.
6. SONET/SDH Maintenance Signals
There are several issues that must be considered in the mapping of
maintenance signals between SONET/SDH and a PSN. A description of
how these signals and conditions are mapped between the two domains
is given below.
For clarity, the mappings are split into two groups: SONET/SDH to PSN
and PSN to SONET/SDH.
6.1. SONET/SDH to PSN
The following sections describe how SONET/SDH Maintenance Signals and
Alarm conditions are mapped into a Packet-Switched Network.
6.1.1. AIS-P Indication
In a SONET/SDH network, SONET/SDH Path outages are signaled using
maintenance alarms, such as Path AIS (AIS-P). In particular, AIS-P
indicates that the SONET/SDH Path is not currently transmitting valid
end-user data, and the SPE contains all ones.
It should be noted that for structured CEM, nearly every type of
service-effecting section or line defect will result in an AIS-P
condition.
The SONET/SDH hierarchy is illustrated below.
+----------+
| PATH |
+----------+
^
|
AIS-P
|
|
+----------+
| LINE |
+ ---------+
^ ^
| |
AIS-L +------ LOP
|
|
+----------+
| SECTION |
+----------+
^ ^
| |
| |
LOS LOF
Figure 6. SONET/SDH Fault Hierarchy
Should the Section Layer detect a Loss of Signal (LOS) or Loss of
Frame (LOF) condition, it sends AIS-L up to the Line Layer. If the
Line Layer detects AIS-L or Loss of Path (LOP), it sends AIS-P to the
Path Layer.
In normal mode during AIS-P, structured CEM packets are generated as
usual. The N and P bits MUST be set to 11 binary to signal AIS-P
explicitly through the packet network. The D-bit MUST be set to zero
to indicate that the SPE is being carried through the packet network.
Normal CEM packets with the SPE fragment, CEM header, the VC label,
and PSN header MUST be transmitted into the packet network.
However, to conserve network bandwidth during AIS-P, DBA MAY be
employed. If DBA has been enabled for AIS-P and AIS-P is currently
occurring, the N and P bits MUST be set to 11 binary to signal AIS,
and the D-bit MUST be set to one to indicate that the SPE is not
being carried through the packet network. Only the CEM header, the
VC label, and the PSN header MUST be transmitted into the packet
network.
Also note that this differs from the outage mechanism in [RFC4447],
which withdraws the VC label as a result of an endpoint outage. TDM
circuit emulation requires the ability to distinguish between the
de-provisioning of a circuit (which causes the VC label to be
withdrawn), and temporary outages (which are signaled using AIS-P).
6.1.2. STS SPE Unequipped Indication
The STS SPE Unequipped Indication is a slightly different case than
AIS-P. When byte C2 of the Path Overhead (STS path signal label) is
00h and Byte B3 (STS Path BIP-8) is valid, it indicates that the STS
SPE is unequipped. Note: this is typically signaled by setting the
entire SPE to zeros.
For normal structured CEM operation during STS SPE Unequipped, the N
and P bits MUST be interpreted as usual. The SPE MUST be transmitted
into the packet network along with the CEM header, the VC label, and
PSN header, and the D-Bit MUST be set to zero.
If DBA has been enabled for STS SPE Unequipped and the Unequipped
condition is occurring on the SONET/SDH channel, the D-bit MUST be
set to one to indicate DBA is active. Only the CEM header, the VC
Label, and PSN header MUST be transmitted into the packet network.
The N and P bits MUST be used to signal pointer adjustments as
normal. See Table 1 and section 8 for details.
6.1.3. CEM-RDI
The CEM function MUST send CEM-RDI towards the packet network during
loss of packet synchronization. This MUST be accomplished by setting
the R bit to one in the CEM header. This applies for both structured
and unstructured CEM.
6.2. PSN to SONET/SDH
The following sections discuss how the various conditions on the
packet network are converted into SONET/SDH indications.
6.2.1. AIS-P Indication
There are several conditions in the packet network that will cause
the structured CEM de-packetization function to send an AIS-P
indication onto a SONET/SDH channel.
The first of these is the receipt of structured CEM packets with the
N and P bits set to one, and the D-bit set to zero. This is an
explicit indication of AIS-P being received at the far-end of the
packet network, with DBA disabled for AIS-P. The CEM de-packetizer
MUST play out the received SPE fragment (which will incidentally be
carrying all ones), and MUST configure the SONET/SDH Overhead to
signal AIS-P as defined in [T1.105], [G.707], and [GR-253].
The second case is the receipt of structured CEM packets with the N
and P bits set to one, and the D-bit set to one. This is an explicit
indication of AIS-P being received at the far-end of the packet
network, with DBA enabled for AIS-P. The CEM de-packetizer MUST play
out one packet's worth of all ones for each packet received, and MUST
configure the SONET/SDH Overhead to signal AIS-P as defined in
[T1.105], [G.707], and [GR-253].
A third case that will cause a structured CEM de-packetization
function to send an AIS-P indication onto a SONET/SDH channel is loss
of packet synchronization.
6.2.2. STS SPE Unequipped Indication
There are three conditions in the packet network that will cause the
CEM function to transmit STS SPE Unequipped Indications onto the
SONET/SDH channel.
The first, which is transparent to CEM, is the receipt of regular CEM
packets that happen to be carrying an SPE that contains the
appropriate Path Overhead to signal STS SPE Unequipped. This case
does not require any special processing on the part of the CEM
de-packetizer.
The second case is the receipt of structured CEM packets that have
the D-bit set to one to indicate that DBA is active and the N and P
bits set to 00 binary, 01 binary, or 10 binary to indicate STS SPE
Unequipped with or without pointer adjustments. The CEM
de-packetizer MUST use this information to transmit a packet of all
zeros onto the SONET/SDH interface, and adjust the payload pointer as
necessary.
The third case when a structured CEM de-packetizer MUST send an STS
SPE Unequipped Indication towards the SONET/SDH interface is when the
VC-label has been withdrawn due to de-provisioning of the circuit.
7. Clocking Modes
It is necessary to be able to regenerate the input service clock at
the output interface. Two clocking modes are supported: synchronous
and asynchronous. Selection of the clocking mode is made as part of
service provisioning. Both ends of the emulated circuit must be
configured with the same clocking mode.
7.1. Synchronous
When synchronous SONET/SDH timing is available at both ends of the
circuit, the issue of clock recovery becomes much simpler.
7.1.1. Synchronous Unstructured CEM
For unstructured CEM, the external clock is used to clock each bit
onto the optical carrier.
7.1.2. Synchronous Structured CEM
For structured CEM, the external clock is used to clock the SONET/SDH
carrier. The N and P bits are used to signal negative or positive
pointer adjustment events between structured CEM endpoints.
If there is a frequency offset between the frame rate of the
transport overhead and that of the SONET/SDH SPE, then the alignment
of the SPE shall periodically slip back or advance in time through
positive or negative stuffing. The N and P bits are used to replay
the pointer adjustment events and eliminate transport jitter.
During a negative pointer adjustment event, the H3 byte from the
SONET/SDH stream is incorporated into the CEM packet payload in order
with the rest of the SPE. During a positive pointer adjustment
event, the stuff byte is not included in the CEM packet payload.
The pointer adjustment event MUST be transmitted in three consecutive
packets by the packetizer. The de-packetizer MUST play out the
pointer adjustment event when the first packet of the three with the
N/P bits set is received.
The CEM de-packetizer MUST utilize the CEM sequence numbers to insure
that SONET/SDH pointer adjustment events are not played any more
frequently than once per every three CEM packets transmitted by the
remote CEM packetizer.
References [T1.105], [G.707], and [GR-253] specify that pointer
adjustment events MUST be separated by three SONET/SDH frames without
a pointer adjustment event. In order to relay all legal pointer
adjustment events, the packet size for a specific circuit MUST be no
larger than (783 * 4 * N)/3, where N is the STS-Nc multiplier.
However, some SONET/SDH equipment allows pointer adjustments to occur
in back-to-back SONET/SDH frames. In order to support this
possibility, the packet size for a particular circuit SHOULD be no
larger than (783*N)/3, where N is the STS-Nc multiplier.
Since the minimum value of N is one, CEM implementations SHOULD
support a minimum payload length of 783/3 or 261 bytes. Smaller
payload lengths MAY be supported as an option.
7.2. Asynchronous
If synchronous timing is not available, other methods MAY be employed
to regenerate the circuit timing.
For structured CEM, the CEM packetizer SHOULD generate the N and P
bits as usual. However, without external synchronization, this
information is not sufficient to reliably justify the SPE within the
SONET/SDH transport framing at the CEM de-packetizer. The
de-packetizer MAY employ an adaptive algorithm to introduce pointer
adjustment events to map the CEM SPE to the SONET/SDH transport
framing. Regardless of whether the N and P bits are used by the
de-packetizer as part of the adaptive clock recovery algorithm, they
MUST still be used in conjunction with the D-bit to signal AIS-P, STS
SPE Unequipped, and DBA.
For unstructured CEM, the CEM de-packetizer MAY use an adaptive clock
recovery technique to regenerate the SONET/SDH transport clock.
An example adaptive clock recovery method can be found in section
3.4.2 of [VTOA].
8. CEM LSP Signaling
For maximum network scaling in MPLS applications, CEM LSP signaling
may be performed using the Label Distribution Protocol (LDP) Extended
Discovery mechanism as augmented by the Pseudo-Wire id Forward Error
Correction (PWid FEC) Element defined in [RFC4447]. MPLS traffic
tunnels may be dedicated to CEM, or shared with other MPLS-based
services. The value 0x8008 is used for the PWE3 PW Type in the PWid
FEC Element in order to signify that the LSP being signaled is to
carry CEM. Note that the generic control word defined in [GR-253] is
not used, as its functionality is included in the CEM encapsulation
header.
Alternatively, static label assignment may be used, or a dedicated
traffic engineered LSP may be used for each CEM service.
Normal CEM packets are fixed in length for all of the packets of a
particular emulated TDM stream. This length is signaled using the
CEM Payload Bytes parameter defined in [RFC4447], or it is statically
provisioned for each CEM service.
At this time, other aspects of the CEM service MUST be statically
provisioned.
9. Security Considerations
The CEM encapsulation is subject to all of the general security
considerations discussed in [RFC3985] and [RFC4447]. In addition,
this document specifies only encapsulations, and not the protocols
used to carry the encapsulated packets across the PSN. Each such
protocol may have its own set of security issues, but those issues
are not affected by the encapsulations specified herein. Note that
the security of the transported CEM service will only be as good as
the security of the PSN. This level of security may be less rigorous
then that available from a native TDM service due to the inherent
differences between circuit-switched and packet-switched public
networks.
10. IANA Considerations
IANA has already allocated the PWE3 PW Type value 0x0008 for this
specification. No further actions are required.
11. References
11.1. Normative References
[G.707] ITU Recommendation G.707, "Network Node Interface For The
Synchronous Digital Hierarchy", 1996.
[GR-253] Telcordia Technologies, "Synchronous Optical Network
(SONET) Transport Systems: Common Generic Criteria", GR-
253-CORE, Issue 3, September 2000.
[I.363.1] ITU-T, "Recommendation I.363.1, B-ISDN Adaptation Layer
Specification: Type AAL1", Appendix III, August 1996.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC4023] Worster, T., Rekhter, Y., and E. Rosen, Ed.,
"Encapsulating MPLS in IP or Generic Routing
Encapsulation (GRE)", RFC 4023, March 2005.
[RFC4447] Martini, L., Ed., Rosen, E., El-Aawar, N., Smith, T., and
G. Heron, "Pseudowire Setup and Maintenance Using the
Label Distribution Protocol (LDP)", RFC 4447, April 2006.
[RFC4842] Malis, A., Pate, P., Cohen, R., Ed., and D. Zelig,
"Synchronous Optical Network/Synchronous Digital
Hierarchy (SONET/SDH) Circuit Emulation over Packet
(CEP)", RFC 4842, April 2007.
[RFC4928] Swallow, G., Bryant, S., and L. Andersson, "Avoiding
Equal Cost Multipath Treatment in MPLS Networks", BCP
128, RFC 4928, June 2007.
[T1.105] American National Standards Institute, "Synchronous
Optical Network (SONET) - Basic Description including
Multiplex Structure, Rates and Formats," ANSI T1.105-
1995.
[VTOA] ATM Forum, "Circuit Emulation Service Interoperability
Specification Version 2.0", af-vtoa-0078.000, January
1997.
11.2. Informative References
[RFC3985] Bryant, S., Ed., and P. Pate, Ed., "Pseudo Wire Emulation
Edge-to-Edge (PWE3) Architecture", RFC 3985, March 2005.
Appendix A. SONET/SDH Rates and Formats
For simplicity, the discussion in this section uses SONET
terminology, but it applies equally to SDH as well. SDH-equivalent
terminology is shown in the tables.
The basic SONET modular signal is the synchronous transport
signal-level 1 (STS-1). A number of STS-1s may be multiplexed into
higher-level signals denoted as STS-N, with N synchronous payload
envelopes (SPEs). The optical counterpart of the STS-N is the
Optical Carrier-level N, or OC-N. Table 2 lists standard SONET line
rates discussed in this document.
OC Level OC-1 OC-3 OC-12 OC-48 OC-192
SDH Term - STM-1 STM-4 STM-16 STM-64
Line Rate(Mb/s) 51.840 155.520 622.080 2,488.320 9,953.280
Table 2. Standard SONET Line Rates
Each SONET frame is 125 us and consists of nine rows. An STS-N frame
has nine rows and N*90 columns. Of the N*90 columns, the first N*3
columns are transport overhead and the other N*87 columns are SPEs.
A number of STS-1s may also be linked together to form a super-rate
signal with only one SPE. The optical super-rate signal is denoted
as OC-Nc, which has a higher payload capacity than OC-N.
The first 9-byte column of each SPE is the Path Overhead (POH) and
the remaining columns form the payload capacity with fixed stuff
(STS-Nc only). The fixed stuff, which is purely overhead, is N/3-1
columns for STS-Nc. Thus, STS-1 and STS-3c do not have any fixed
stuff, STS-12c has three columns of fixed stuff, and so on.
The POH of an STS-1 or STS-Nc is always nine bytes in nine rows. The
payload capacity of an STS-1 is 86 columns (774 bytes) per frame.
The payload capacity of an STS-Nc is (N*87)-(N/3) columns per frame.
Thus, the payload capacity of an STS-3c is (3*87 - 1)*9 = 2,340 bytes
per frame. As another example, the payload capacity of an STS-192c
is 149,760 bytes, which is exactly 64 times larger than the STS-3c.
There are 8,000 SONET frames per second. Therefore, the SPE size,
(POH plus payload capacity) of an STS-1 is 783*8*8,000 = 50.112 Mb/s.
The SPE size of a concatenated STS-3c is 2,349 bytes per frame or
150.336 Mb/s. The payload capacity of an STS-192c is 149,760 bytes
per frame, which is equivalent to 9,584.640 Mb/s. Table 3 lists the
SPE and payload rates supported.
SONET STS Level STS-1 STS-3c STS-12c STS-48c STS-192c
SDH VC Level - VC-4 VC-4-4c VC-4-16c VC-4-64c
Payload Size(Bytes) 774 2,340 9,360 37,440 149,760
Payload Rate(Mb/s) 49.536 149.760 599.040 2,396.160 9,584.640
SPE Size(Bytes) 783 2,349 9,396 37,584 150,336
SPE Rate(Mb/s) 50.112 150.336 601.344 2,405.376 9,621.504
Table 3. Payload Size and Rate
To support circuit emulation, the entire SPE of a SONET STS or SDH VC
level is encapsulated into packets, using the encapsulation defined
in section 5, for carriage across packet-switched networks.
Appendix B. ECC-6 Definition
ECC-6 is an Error Correction Code to protect the CEM header. This
provides single bit correction and the ability to detect up to two
bit errors.
Error Correction Code:
|---------------Header bits 0-25 -------------------| ECC-6 code|
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|1 1 1 1 1 0 0 0 1 0 0 0 1 1 1 1 1 0 1 0 0 0 1 0 1 1|1 0 0 0 0 0|
|1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 1 0 1 1 1 1 1 1 1|0 1 0 0 0 0|
|1 0 0 0 1 1 1 1 0 0 1 0 1 1 1 0 0 0 1 1 1 1 0 0 1 1|0 0 1 0 0 0|
|0 1 0 0 1 1 1 1 0 0 0 1 1 0 0 1 1 1 1 1 0 0 1 1 0 1|0 0 0 1 0 0|
|0 0 1 0 0 0 1 0 1 1 1 1 1 1 0 0 1 1 1 1 1 0 1 0 1 0|0 0 0 0 1 0|
|0 0 0 1 0 0 0 1 1 1 1 1 0 0 1 1 0 0 1 1 0 1 1 1 1 1|0 0 0 0 0 1|
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 7. ECC-6 Check Matrix X
The ECC-6 code protects the 32-bit CEM header as follows:
The encoder generates the 6-bit ECC using the matrix shown in Figure
7. In brief, the encoder builds another 26 column by 6 row matrix
and calculates even parity over the rows. The matrix columns
represent CEM header bits 0 through 25.
Denote each column of the ECC-6 check matrix by X[], and each column
of the intermediate encoder matrix as Y[]. CEM[] denotes the CEM
header and ECC[] is the error correction code that is inserted into
CEM header bits 26 through 31.
for i = 0 to 25 {
if CEM[i] = 0 {
Y[i] = 0;
} else {
Y[i] = X[i];
}
}
In other words, for each CEM header bit (i) set to one, set the
resulting matrix column Y[i] according to Figure 7.
The final ECC-6 code is calculated as even parity of each row in Y
(i.e., ECC[k]=CEM[25+k]=even parity of row k).
The receiver also uses matrix X to calculate an intermediate matrix
Y' based on all 32 bits of the CEM header. Therefore, Y' is 32
columns wide and includes the ECC-6 code.
for i = 0 to 31 {
if CEM[i] = 0 {
Y'[i] = 0;
} else {
Y'[i] = X[i];
}
}
The receiver then appends the incoming ECC-6 code to Y as column 32
(ECC[0] should align with row 0) and calculates even parity for each
row. The result is a single 6-bit column Z. If all 6 bits are 0,
there are no bit errors (or at least no detectable errors).
Otherwise, it uses Z to perform a reverse lookup on X[] from Figure
7. If Z matches column X[i], then there is a single bit error. The
receiver should invert bit CEM[i] to correct the header. If Z fails
to match any column of X, then the CEM header contains more than one
bit error and the CEM packet MUST be discarded.
Note that the ECC-6 code provides single-bit correction and 2-bit
detection of errors within the received ECC-6 code itself.
Acknowledgments
The authors would like to thank Mitri Halabi, Bob Colvin, and Ken
Hsu, all formerly of Vivace Networks and Tellabs; Tom Johnson,
Marlene Drost, Ed Hallman, and Dave Danenberg, all formerly of
Litchfield Communications, for their contributions to the document.
Authors' Addresses
Andrew G. Malis
Verizon Communications
40 Sylvan Road
Waltham, MA 02451
EMail: andrew.g.malis@verizon.com
Jeremy Brayley
ECI Telecom Inc.
Omega Corporate Center
1300 Omega Drive
Pittsburgh, PA 15205
EMail: jeremy.brayley@ecitele.com
John Shirron
ECI Telecom Inc.
Omega Corporate Center
1300 Omega Drive
Pittsburgh, PA 15205
EMail: john.shirron@ecitele.com
Luca Martini
Cisco Systems, Inc.
9155 East Nichols Avenue, Suite 400
Englewood, CO, 80112
EMail: lmartini@cisco.com
Steve Vogelsang
Alcatel-Lucent
600 March Road
Kanata, ON K2K 2T6
Canada
EMail: steve.vogelsang@alcatel-lucent.com
Full Copyright Statement
Copyright (C) The IETF Trust (2008).
This document is subject to the rights, licenses and restrictions
contained in BCP 78 and at www.rfc-editor.org/copyright.html, and
except as set forth therein, the authors retain all their rights.
This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND
THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS
OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF
THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
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
http://www.ietf.org/ipr.
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
ietf-ipr@ietf.org.