| Rfc | 6310 |
|---|
| Title | Pseudowire (PW) Operations, Administration, and Maintenance (OAM)
Message Mapping |
|---|
| Author | M. Aissaoui, P. Busschbach, L. Martini, M. Morrow,
T. Nadeau, Y(J). Stein |
|---|
| Date | July 2011 |
|---|
| Format: | TXT, HTML |
|---|
| Status: | PROPOSED STANDARD |
|---|
|
Internet Engineering Task Force (IETF) M. Aissaoui
Request for Comments: 6310 P. Busschbach
Category: Standards Track Alcatel-Lucent
ISSN: 2070-1721 L. Martini
M. Morrow
Cisco Systems, Inc.
T. Nadeau
CA Technologies
Y(J). Stein
RAD Data Communications
July 2011
Pseudowire (PW) Operations, Administration, and Maintenance (OAM)
Message Mapping
Abstract
This document specifies the mapping and notification of defect states
between a pseudowire (PW) and the Attachment Circuits (ACs) of the
end-to-end emulated service. It standardizes the behavior of
Provider Edges (PEs) with respect to PW and AC defects. It addresses
ATM, Frame Relay, Time Division Multiplexing (TDM), and Synchronous
Optical Network / Synchronous Digital Hierarchy (SONET/SDH) PW
services, carried over MPLS, MPLS/IP, and Layer 2 Tunneling Protocol
version 3/IP (L2TPv3/IP) Packet Switched Networks (PSNs).
Status of This Memo
This is an Internet Standards Track document.
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by the
Internet Engineering Steering Group (IESG). Further information on
Internet Standards is available in Section 2 of RFC 5741.
Information about the current status of this document, any errata,
and how to provide feedback on it may be obtained at
http://www.rfc-editor.org/info/rfc6310.
Copyright Notice
Copyright (c) 2011 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
This document may contain material from IETF Documents or IETF
Contributions published or made publicly available before November
10, 2008. The person(s) controlling the copyright in some of this
material may not have granted the IETF Trust the right to allow
modifications of such material outside the IETF Standards Process.
Without obtaining an adequate license from the person(s) controlling
the copyright in such materials, this document may not be modified
outside the IETF Standards Process, and derivative works of it may
not be created outside the IETF Standards Process, except to format
it for publication as an RFC or to translate it into languages other
than English.
10.2. Consequent Actions .......................................30
10.2.1. PW Receive Defect State Entry/Exit ................30
10.2.2. PW Transmit Defect State Entry/Exit ...............30
10.2.3. AC Receive Defect State Entry/Exit ................30
11. Security Considerations .......................................31
12. Contributors and Acknowledgments ..............................31
13. References ....................................................32
13.1. Normative References .....................................32
13.2. Informative References ...................................34
Appendix A. Native Service Management (Informative) ...............36
A.1. Frame Relay Management .....................................36
A.2. ATM Management .............................................37
Appendix B. PW Defects and Detection Tools ........................38
B.1. PW Defects .................................................38
B.2. Packet Loss ................................................38
B.3. PW Defect Detection Tools ..................................38
B.4. PW Specific Defect Detection Mechanisms ....................39
1. Introduction
This document specifies the mapping and notification of defect states
between a pseudowire and the Attachment Circuits (AC) of the end-to-
end emulated service. It covers the case where the ACs and the PWs
are of the same type in accordance to the Pseudowire Emulation Edge-
to-Edge (PWE3) architecture [RFC3985] such that a homogeneous PW
service can be constructed.
This document is motivated by the requirements put forth in [RFC4377]
and [RFC3916]. Its objective is to standardize the behavior of PEs
with respect to defects on PWs and ACs, so that there is no ambiguity
about the alarms generated and consequent actions undertaken by PEs
in response to specific failure conditions.
This document addresses PWs over MPLS, MPLS/IP, L2TPv3/IP PSNs, ATM,
Frame Relay, TDM, and SONET/SDH PW native services. Due to its
unique characteristics, the Ethernet PW service is covered in a
separate document [Eth-OAM-Inter].
This document provides procedures for PWs set up using Label
Distribution Protocol (LDP) [RFC4447] or L2TPv3 [RFC3931] control
protocols. While we mention fault reporting options for PWs
established by other means (e.g., by static configuration or via
BGP), we do not provide detailed procedures for such cases.
This document is scoped only to single segment PWs. The mechanisms
described in this document could also be applied to terminating PEs
(T-PEs) for multi-segment PWs (MS-PWs) ([RFC5254]). Section 10 of
[RFC6073] details procedures for generating or relaying PW status by
a switching PE (S-PE).
2. Abbreviations and Conventions
2.1. Abbreviations
AAL5 ATM Adaptation Layer 5
AIS Alarm Indication Signal
AC Attachment Circuit
ATM Asynchronous Transfer Mode
AVP Attribute Value Pair
BFD Bidirectional Forwarding Detection
CC Continuity Check
CDN Call Disconnect Notify
CE Customer Edge
CV Connectivity Verification
DBA Dynamic Bandwidth Allocation
DLC Data Link Connection
FDI Forward Defect Indication
FR Frame Relay
FRBS Frame Relay Bearer Service
ICMP Internet Control Message Protocol
LB Loopback
LCCE L2TP Control Connection Endpoint
LDP Label Distribution Protocol
LSP Label Switched Path
L2TP Layer 2 Tunneling Protocol
MPLS Multiprotocol Label Switching
NE Network Element
NS Native Service
OAM Operations, Administration, and Maintenance
PE Provider Edge
PSN Packet Switched Network
PW Pseudowire
RDI Reverse Defect Indication
PDU Protocol Data Unit
SDH Synchronous Digital Hierarchy
SDU Service Data Unit
SONET Synchronous Optical Network
TDM Time Division Multiplexing
TLV Type Length Value
VCC Virtual Channel Connection
VCCV Virtual Connection Connectivity Verification
VPC Virtual Path Connection
2.2. Conventions
The words "defect" and "fault" are used interchangeably to mean any
condition that negatively impacts forwarding of user traffic between
the CE endpoints of the PW service.
The words "defect notification" and "defect indication" are used
interchangeably to mean any OAM message generated by a PE and sent to
other nodes in the network to convey the defect state local to this
PE.
The PW can be carried over three types of Packet Switched Networks
(PSNs). An "MPLS PSN" makes use of MPLS Label Switched Paths
[RFC3031] as the tunneling technology to forward the PW packets. An
"MPLS/IP PSN" makes use of MPLS-in-IP tunneling [RFC4023], with an
MPLS shim header used as PW demultiplexer. An "L2TPv3/IP PSN" makes
use of L2TPv3/IP [RFC3931] as the tunneling technology with the
L2TPv3/IP Session ID as the PW demultiplexer.
If LSP-Ping [RFC4379] is run over a PW as described in [RFC5085], it
will be referred to as "VCCV-Ping". If BFD is run over a PW as
described in [RFC5885], it will be referred to as "VCCV-BFD".
While PWs are inherently bidirectional entities, defects and OAM
messaging are related to a specific traffic direction. We use the
terms "upstream" and "downstream" to identify PEs in relation to the
traffic direction. A PE is upstream for the traffic it is forwarding
and is downstream for the traffic it is receiving.
We use the terms "local" and "remote" to identify native service
networks and ACs in relation to a specific PE. The local AC is
attached to the PE in question, while the remote AC is attached to
the PE at the other end of the PW.
A "transmit defect" is any defect that uniquely impacts traffic sent
or relayed by the observing PE. A "receive defect" is any defect
that impacts information transfer to the observing PE. Note that a
receive defect also impacts traffic meant to be relayed, and thus can
be considered to incorporate two defect states. Thus, when a PE
enters both receive and transmit defect states of a PW service, the
receive defect takes precedence over the transmit defect in terms of
the consequent actions.
A "forward defect indication" (FDI) is sent in the same direction as
the user traffic impacted by the defect. A "reverse defect
indication" (RDI) is sent in the direction opposite to that of the
impacted traffic.
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. Reference Model and Defect Locations
Figure 1 illustrates the PWE3 network reference model with an
indication of the possible defect locations. This model will be
referenced in the remainder of this document for describing the OAM
procedures.
ACs PSN tunnel ACs
+----+ +----+
+----+ | PE1|==================| PE2| +----+
| |---(a)---(b)..(c)......PW1..(d)..(e)..(f)---(g)---| |
| CE1| (N1) | | | | (N2) |CE2 |
| |----------|............PW2.............|----------| |
+----+ | |==================| | +----+
^ +----+ +----+ ^
| Provider Edge 1 Provider Edge 2 |
| |
|<-------------- Emulated Service ---------------->|
Customer Customer
Edge 1 Edge 2
Figure 1: PWE3 Network Defect Locations
The procedures will be described in this document from the viewpoint
of PE1, so that N1 is the local native service network and N2 is the
remote native service network. PE2 will typically implement the same
functionality. Note that PE1 is the upstream PE for traffic
originating in the local NS network N1, while it is the downstream PE
for traffic originating in the remote NS network N2.
The following is a brief description of the defect locations:
a. Defect in NS network N1. This covers any defect in network N1
(including any CE1 defect) that impacts all or some ACs attached
to PE1, and is thus a local AC defect. The defect is conveyed to
PE1 and to NS network N2 using NS specific OAM defect indications.
b. Defect on a PE1 AC interface (another local AC defect).
c. Defect on a PE1 PSN interface.
d. Defect in the PSN network. This covers any defect in the PSN that
impacts all or some PWs between PE1 and PE2. The defect is
conveyed to the PE using a PSN and/or a PW specific OAM defect
indication. Note that both data plane defects and control plane
defects must be taken into consideration. Although control
messages may follow a different path than PW data plane traffic, a
control plane defect may affect the PW status.
e. Defect on a PE2 PSN interface.
f. Defect on a PE2 AC interface (a remote AC defect).
g. Defect in NS network N2 (another remote AC defect). This covers
any defect in N2 (including any CE2 defect) that impacts all or a
subset of ACs attached to PE2. The defect is conveyed to PE2 and
to NS network N1 using the NS OAM defect indication.
4. Abstract Defect States
PE1 must track four defect states that reflect the observed states of
both directions of the PW service on both the AC and the PW sides.
Defects may impact one or both directions of the PW service.
The observed state is a combination of defects directly detected by
PE1 and defects of which it has been made aware via notifications.
+-----+
----AC receive---->| |-----PW transmit---->
CE1 | PE1 | PE2/CE2
<---AC transmit----| |<----PW receive-----
+-----+
(arrows indicate direction of user traffic impacted by a defect)
Figure 2: Receive and Transmit Defect States
PE1 will directly detect or be notified of AC receive or PW receive
defects as they occur upstream of PE1 and impact traffic being sent
to PE1. As a result, PE1 enters the AC or PW receive defect state.
In Figure 2, PE1 may be notified of a receive defect in the AC by
receiving a forward defect indication, e.g., ATM AIS, from CE1 or an
intervening network. This defect notification indicates that user
traffic sent by CE1 may not be received by PE1 due to a defect. PE1
can also directly detect an AC receive defect if it resulted from a
failure of the receive side in the local port or link over which the
AC is configured.
Similarly, PE1 may detect or be notified of a receive defect in the
PW by receiving a forward defect indication from PE2. If the PW
status TLV is used for fault notification, this message will indicate
a Local PSN-facing PW (egress) Transmit Fault or a Local AC (ingress)
Receive Fault at PE2, as described in Section 6.1.1. This defect
notification indicates that user traffic sent by CE2 may not be
received by PE1 due to a defect. As a result, PE1 enters the PW
receive defect state.
Note that a forward defect indication is sent in the same direction
as the user traffic impacted by the defect.
Generally, a PE cannot detect transmit defects by itself and will
therefore need to be notified of AC transmit or PW transmit defects
by other devices.
In Figure 2, PE1 may be notified of a transmit defect in the AC by
receiving a reverse defect indication, e.g., ATM RDI, from CE1. This
defect relates to the traffic sent by PE1 to CE1 on the AC.
Similarly, PE1 may be notified of a transmit defect in the PW by
receiving a reverse defect indication from PE2. If PW status is used
for fault notification, this message will indicate a Local PSN-
facing PW (ingress) Receive Fault or a Local Attachment Circuit
(egress) Transmit Fault at PE2, as described in Section 6.1.1. This
defect impacts the traffic sent by PE1 to CE2. As a result, PE1
enters the PW transmit defect state.
Note that a reverse defect indication is sent in the reverse
direction to the user traffic impacted by the defect.
The procedures outlined in this document define the entry and exit
criteria for each of the four states with respect to the set of PW
services within the document scope and the consequent actions that
PE1 must perform.
When a PE enters both receive and transmit defect states related to
the same PW service, then the receive defect takes precedence over
transmit defect in terms of the consequent actions.
5. OAM Modes
A homogeneous PW service forwards packets between an AC and a PW of
the same type. It thus implements both NS OAM and PW OAM mechanisms.
PW OAM defect notification messages are described in Section 6.1. NS
OAM messages are described in Appendix A.
This document defines two different OAM modes, the distinction being
the method of mapping between the NS and PW OAM defect notification
messages.
The first mode, illustrated in Figure 3, is called the "single
emulated OAM loop" mode. Here, a single end-to-end NS OAM loop is
emulated by transparently passing NS OAM messages over the PW. Note
that the PW OAM is shown outside the PW in Figure 3, as it is
transported in LDP messages or in the associated channel, not inside
the PW itself.
+-----+ +-----+
+-----+ | |=================| | +-----+
| CE1 |-=NS-OAM=>| PE1 |----=NS-OAM=>----| PE2 |-=NS-OAM=>| CE2 |
+-----+ | |=================| | +-----+
+-----+ +-----+
\ /
-------=PW-OAM=>-------
Figure 3: Single Emulated OAM Loop Mode
The single emulated OAM loop mode implements the following behavior:
a. The upstream PE (PE1) MUST transparently relay NS OAM messages
over the PW.
b. The upstream PE MUST signal local defects affecting the AC using a
NS defect notification message sent over the PW. In the case that
it is not possible to generate NS OAM messages (e.g., because the
defect interferes with NS OAM message generation), the PE MUST
signal local defects affecting the AC using a PW defect
notification message.
c. The upstream PE MUST signal local defects affecting the PW using a
PW defect notification message.
d. The downstream PE (PE2) MUST insert NS defect notification
messages into its local AC when it detects or is notified of a
defect in the PW or remote AC. This includes translating received
PW defect notification messages into NS defect notification
messages for defects signaled by the upstream PE.
The single emulated OAM loop mode is suitable for PW services that
have a widely deployed NS OAM mechanism. This document specifies the
use of this mode for ATM PW, TDM PW, and Circuit Emulation over
Packet (CEP) PW services. It is the default mode of operation for
all ATM cell mode PW services and the only mode specified for CEP and
Structure-Agnostic TDM over Packets / Circuit Emulation Service over
Packet Switched Network (SAToP/CESoPSN) TDM PW services. It is
optional for AAL5 PDU transport and AAL5 SDU transport modes.
The second OAM mode operates three OAM loops joined at the AC/PW
boundaries of the PEs. This is referred to as the "coupled OAM
loops" mode and is illustrated in Figure 4. Note that in contrast to
Figure 3, NS OAM messages are never carried over the PW.
+-----+ +-----+
+-----+ | |=================| | +-----+
| CE1 |-=NS-OAM=>| PE1 | | PE2 |-=NS-OAM=>| CE2 |
+-----+ | |=================| | +-----+
+-----+ +-----+
\ /
-------=PW-OAM=>-------
Figure 4: Coupled OAM Loops Mode
The coupled OAM loops mode implements the following behavior:
a. The upstream PE (PE1) MUST terminate and translate a received NS
defect notification message into a PW defect notification message.
b. The upstream PE MUST signal local failures affecting its local AC
using PW defect notification messages to the downstream PE.
c. The upstream PE MUST signal local failures affecting the PW using
PW defect notification messages.
d. The downstream PE (PE2) MUST insert NS defect notification
messages into the AC when it detects or is notified of defects in
the PW or remote AC. This includes translating received PW defect
notification messages into NS defect notification messages.
This document specifies the coupled OAM loops mode as the default
mode for the Frame Relay, ATM AAL5 PDU transport, and AAL5 SDU
transport services. It is an optional mode for ATM VCC cell mode
services. This mode is not specified for TDM, CEP, or ATM VPC cell
mode PW services. RFC 5087 defines a similar but distinct mode, as
will be explained in Section 9. For the ATM VPC cell mode case a
pure coupled OAM loops mode is not possible as a PE MUST
transparently pass VC-level (F5) ATM OAM cells over the PW while
terminating and translating VP-level (F4) OAM cells.
6. PW Defect States and Defect Notifications
6.1. PW Defect Notification Mechanisms
For MPLS and MPLS/IP PSNs, a PE that establishes a PW using the Label
Distribution Protocol [RFC5036], and that has negotiated use of the
LDP status TLV per Section 5.4.3 of [RFC4447], MUST use the PW status
TLV mechanism for AC and PW status and defect notification.
Additionally, such a PE MAY use VCCV-BFD Connectivity Verification
(CV) for fault detection only (CV types 0x04 and 0x10 [RFC5885]).
A PE that establishes an MPLS PW using means other than LDP, e.g., by
static configuration or by use of BGP, MUST support some alternative
method of status reporting. The design of a suitable mechanism to
carry the aforementioned status TLV in the PW associated channel is
work in progress [Static-PW-Status]. Additionally, such a PE MAY use
VCCV-BFD CV for both fault detection and status notification (CV
types 0x08 and 0x20 [RFC5885]).
For a L2TPv3/IP PSN, a PE SHOULD use the Circuit Status Attribute
Value Pair (AVP) as the mechanism for AC and PW status and defect
notification. In its most basic form, the Circuit Status AVP
[RFC3931] in a Set-Link-Info (SLI) message can signal active/inactive
AC status. The Circuit Status AVP as described in [RFC5641] is
proposed to be extended to convey status and defects in the AC and
the PSN-facing PW in both ingress and egress directions, i.e., four
independent status bits, without the need to tear down the sessions
or control connection.
When a PE does not support the Circuit Status AVP, it MAY use the
Stop-Control-Connection-Notification (StopCCN) and the Call-
Disconnect-Notify (CDN) messages to tear down L2TP sessions in a
fashion similar to LDP's use of Label Withdrawal to tear down a PW.
A PE may use the StopCCN to shut down the L2TP control connection,
and implicitly all L2TP sessions associated with that control
connection, without any explicit session control messages. This is
useful for the case of a failure which impacts all L2TP sessions (all
PWs) managed by the control connection. It MAY use CDN to disconnect
a specific L2TP session when a failure only affects a specific PW.
Additionally, a PE MAY use VCCV-BFD CV types 0x04 and 0x10 for fault
detection only, but SHOULD notify the remote PE using the Circuit
Status AVP. A PE that establishes a PW using means other than the
L2TP control plane, e.g., by static configuration or by use of BGP,
MAY use VCCV-BFD CV types 0x08 and 0x20 for AC and PW status and
defect notification. These CV types SHOULD NOT be used when the PW
is established via the L2TP control plane.
The CV types are defined in Section 6.1.3 of this document.
6.1.1. LDP Status TLV
[RFC4446] defines the following PW status code points:
0x00000000 - Pseudowire forwarding (clear all failures)
0x00000001 - Pseudowire Not Forwarding
0x00000002 - Local Attachment Circuit (ingress) Receive Fault
0x00000004 - Local Attachment Circuit (egress) Transmit Fault
0x00000008 - Local PSN-facing PW (ingress) Receive Fault
0x00000010 - Local PSN-facing PW (egress) Transmit Fault
[RFC4447] specifies that the "Pseudowire forwarding" code point is
used to indicate that all faults are to be cleared. It also
specifies that the "Pseudowire Not Forwarding" code point means that
a defect has been detected that is not represented by the defined
code points.
The code points used in the LDP status TLV in a PW status
notification message report defects from the viewpoint of the
originating PE. The originating PE conveys this state in the form of
a forward defect or a reverse defect indication.
The forward and reverse defect indication definitions used in this
document map to the LDP Status TLV codes as follows:
Forward defect indication corresponds to the logical OR of:
* Local Attachment Circuit (ingress) Receive Fault,
* Local PSN-facing PW (egress) Transmit Fault, and
* PW Not Forwarding.
Reverse defect indication corresponds to the logical OR of:
* Local Attachment Circuit (egress) Transmit Fault and
* Local PSN-facing PW (ingress) Receive Fault.
A PE MUST use PW status notification messages to report all defects
affecting the PW service including, but not restricted to, the
following:
o defects detected through fault detection mechanisms in the MPLS
and MPLS/IP PSN,
o defects detected through VCCV-Ping or VCCV-BFD CV types 0x04 and
0x10 for fault detection only,
o defects within the PE that result in an inability to forward
traffic between the AC and the PW,
o defects of the AC or in the Layer 2 network affecting the AC as
per the rules detailed in Section 5 for the "single emulated OAM
loop" mode and "coupled OAM loops" modes.
Note that there are two situations that require PW label withdrawal
as opposed to a PW status notification by the PE. The first one is
when the PW is taken down administratively in accordance with
[RFC4447]. The second one is when the Target LDP session established
between the two PEs is lost. In the latter case, the PW labels will
need to be re-signaled when the Targeted LDP session is re-
established.
6.1.2. L2TP Circuit Status AVP
[RFC3931] defines the Circuit Status AVP in the Set-Link-Info (SLI)
message to exchange initial status and status changes in the circuit
to which the pseudowire is bound. [RFC5641] defines extensions to
the Circuit Status AVP that are analogous to the PW Status TLV
defined for LDP. Consequently, for L2TPv3/IP, the Circuit Status AVP
is used in the same fashion as the PW Status described in the
previous section. Extended circuit status for L2TPv3/IP is described
in [RFC5641].
If the extended Circuit Status bits are not supported, and instead
only the "A bit" (Active) is used as described in [RFC3931], a PE MAY
use CDN messages to clear L2TPv3/IP sessions in the presence of
session-level failures detected in the L2TPv3/IP PSN.
A PE MUST set the Active bit in the Circuit Status to clear all
faults, and it MUST clear the Active bit in the Circuit Status to
convey any defect that cannot be represented explicitly with specific
Circuit Status flags from [RFC3931] or [RFC5641].
The forward and reverse defect indication definitions used in this
document map to the L2TP Circuit Status AVP as follows:
Forward defect indication corresponds to the logical OR of:
* Local Attachment Circuit (ingress) Receive Fault,
* Local PSN-facing PW (egress) Transmit Fault, and
* PW Not Forwarding.
Reverse defect indication corresponds to the logical OR of:
* Local Attachment Circuit (egress) Transmit Fault and
* Local PSN-facing PW (ingress) Receive Fault.
The status notification conveys defects from the viewpoint of the
originating LCCE (PE).
When the extended Circuit Status definition of [RFC5641] is
supported, a PE SHALL use the Circuit Status to report all failures
affecting the PW service including, but not restricted to, the
following:
o defects detected through defect detection mechanisms in the
L2TPv3/IP PSN,
o defects detected through VCCV-Ping or VCCV-BFD CV types 0x04 (BFD
IP/UDP-encapsulated, for PW Fault Detection only) and 0x10 (BFD
PW-ACH-encapsulated (without IP/UDP headers), for PW. Fault
Detection and AC/PW Fault Status Signaling) for fault detection
only which are described in Section 6.1.3 of this document,
o defects within the PE that result in an inability to forward
traffic between the AC and the PW,
o defects of the AC or in the L2 network affecting the AC as per the
rules detailed in Section 5 for the "single emulated OAM loop"
mode and the "coupled OAM loops" modes.
When the extended Circuit Status definition of [RFC5641] is not
supported, a PE SHALL use the A bit in the Circuit Status AVP in the
SLI to report:
o defects of the AC or in the L2 network affecting the AC as per the
rules detailed in Section 5 for the "single emulated OAM loop"
mode and the "coupled OAM loops" modes.
When the extended Circuit Status definition of [RFC5641] is not
supported, a PE MAY use the CDN and StopCCN messages in a similar way
to an MPLS PW label withdrawal to report:
o defects detected through defect detection mechanisms in the
L2TPv3/IP PSN (using StopCCN),
o defects detected through VCCV (pseudowire level) (using CDN),
o defects within the PE that result in an inability to forward
traffic between ACs and PW (using CDN).
For ATM L2TPv3/IP pseudowires, in addition to the Circuit Status AVP,
a PE MAY use the ATM Alarm Status AVP [RFC4454] to indicate the
reason for the ATM circuit status and the specific alarm type, if
any. This AVP is sent in the SLI message to indicate additional
information about the ATM circuit status.
L2TP control connections use Hello messages as a keep-alive facility.
It is important to note that if PSN failure is detected by keep-alive
timeout, the control connection is cleared. L2TP Hello messages are
sent in-band so as to follow the data plane with respect to the
source and destination addresses, IP protocol number, and UDP port
(when UDP is used).
6.1.3. BFD Diagnostic Codes
BFD [RFC5880] defines a set of diagnostic codes that partially
overlap the set of defects that can be communicated through LDP
Status TLV or L2TP Circuit Status AVP. This section describes the
behavior of the PEs with respect to using one or both of these
methods for detecting and propagating defect state.
In the case of an MPLS PW established via LDP signaling, the PEs
negotiate VCCV capabilities during the label mapping messages
exchange used to establish the two directions of the PW. This is
achieved by including a capability TLV in the PW Forward Error
Correction (FEC) interface parameters TLV. In the L2TPv3/IP case,
the PEs negotiate the use of VCCV during the pseudowire session
initialization using the VCCV AVP [RFC5085].
The CV Type Indicators field in the OAM capability TLV or VCCV AVP
defines a bitmask used to indicate the specific OAM capabilities that
the PE can use over the PW being established.
A CV type of 0x04 or 0x10 [RFC5885] indicates that BFD is used for PW
fault detection only. These CV types MAY be used any time the PW is
established using LDP or L2TP control planes. In this mode, only the
following diagnostic (Diag) codes specified in [RFC5880] will be
used:
0 - No diagnostic
1 - Control detection time expired
3 - Neighbor signaled session down
7 - Administratively Down
A PE using VCCV-BFD MUST use diagnostic code 0 to indicate to its
peer PE that it is correctly receiving BFD control messages. It MUST
use diagnostic code 1 to indicate to its peer that it has stopped
receiving BFD control messages and will thus declare the PW to be
down in the receive direction. It MUST use diagnostic code 3 to
confirm to its peer that the BFD session is going down after
receiving diagnostic code 1 from this peer. In this case, it will
declare the PW to be down in the transmit direction. A PE MUST use
diagnostic code 7 to bring down the BFD session when the PW is
brought down administratively. All other defects, such as AC/PW
defects and PE internal failures that prevent it from forwarding
traffic, MUST be communicated through the LDP Status TLV in the case
of MPLS or MPLS/IP PSN, or through the appropriate L2TP codes in the
Circuit Status AVP in the case of L2TPv3/IP PSN.
A CV type of 0x08 or 0x20 in the OAM capabilities TLV indicates that
BFD is used for both PW fault detection and Fault Notification. In
addition to the above diagnostic codes, a PE uses the following codes
to signal AC defects and other defects impacting forwarding over the
PW service:
6 - Concatenated Path Down
8 - Reverse Concatenated Path Down
As specified in [RFC5085], the PEs negotiate the use of VCCV during
PW setup. When a PW transported over an MPLS-PSN is established
using LDP, the PEs negotiate the use of the VCCV capabilities using
the optional VCCV Capability Advertisement Sub-TLV parameter in the
Interface Parameter Sub-TLV field of the LDP PW ID FEC or using an
Interface Parameters TLV of the LDP Generalized PW ID FEC. In the
case of L2TPv3/IP PSNs, the PEs negotiate the use of VCCV during the
pseudowire session initialization using VCCV AVP.
Note that a defect that causes the generation of the "PW not
forwarding code" (diagnostic code 6 or 8) does not necessarily result
in the BFD session going down. However, if the BFD session times
out, then diagnostic code 1 MUST be used since it signals a state
change of the BFD session itself. In general, when a BFD session
changes state, the PEs MUST use state change diagnostic codes 0, 1,
3, and 7 in accordance with [RFC5880], and they MUST override any of
the AC/PW status diagnostic codes (codes 6 or 8) that may have been
signaled prior to the BFD session changing state.
The forward and reverse defect indications used in this document map
to the following BFD codes:
Forward defect indication corresponds to the logical OR of:
* Concatenated Path Down (BFD diagnostic code 06)
* Pseudowire Not Forwarding (PW status code 0x00000001).
Reverse defect indication corresponds to:
* Reverse Concatenated Path Down (BFD diagnostic code 08).
These diagnostic codes are used to signal forward and reverse defect
states, respectively, when the PEs negotiated the use of BFD as the
mechanism for AC and PW fault detection and status signaling
notification. As stated in Section 6.1, these CV types SHOULD NOT be
used when the PW is established with the LDP or L2TP control plane.
6.2. PW Defect State Entry/Exit
6.2.1. PW Receive Defect State Entry/Exit Criteria
PE1, as downstream PE, will enter the PW receive defect state if one
or more of the following occurs:
o It receives a forward defect indication (FDI) from PE2 indicating
either a receive defect on the remote AC or that PE2 detected or
was notified of downstream PW fault.
o It detects loss of connectivity on the PSN tunnel upstream of PE1,
which affects the traffic it receives from PE2.
o It detects a loss of PW connectivity through VCCV-BFD or VCCV-
PING, which affects the traffic it receives from PE2.
Note that if the PW control session (LDP session, the L2TP session,
or the L2TP control connection) between the PEs fails, the PW is torn
down and needs to be re-established. However, the consequent actions
towards the ACs are the same as if the PW entered the receive defect
state.
PE1 will exit the PW receive defect state when the following
conditions are met. Note that this may result in a transition to the
PW operational state or the PW transmit defect state.
o All previously detected defects have disappeared, and
o PE2 cleared the FDI, if applicable.
6.2.2. PW Transmit Defect State Entry/Exit Criteria
PE1, as upstream PE, will enter the PW transmit defect state if the
following conditions occur:
o It receives a Reverse Defect Indication (RDI) from PE2 indicating
either a transmit fault on the remote AC or that PE2 detected or
was notified of a upstream PW fault, and
o it is not already in the PW receive defect state.
PE1 will exit the transmit defect state if it receives an OAM message
from PE2 clearing the RDI, or it has entered the PW receive defect
state.
For a PW over L2TPv3/IP using the basic Circuit Status AVP [RFC3931],
the PW transmit defect state is not valid and a PE can only enter the
PW receive defect state.
7. Procedures for ATM PW Service
The following procedures apply to Asynchronous Transfer Mode (ATM)
pseudowires [RFC4717]. ATM terminology is explained in Appendix A.2
of this document.
7.1. AC Receive Defect State Entry/Exit Criteria
When operating in the coupled OAM loops mode, PE1 enters the AC
receive defect state when any of the following conditions are met:
a. It detects or is notified of a physical layer fault on the ATM
interface.
b. It receives an end-to-end Flow 4 OAM (F4) Alarm Indication Signal
(AIS) OAM flow on a Virtual Path (VP) AC or an end-to-end Flow 5
(F5) AIS OAM flow on a Virtual Circuit (VC) as per ITU-T
Recommendation I.610 [I.610], indicating that the ATM VPC or VCC
is down in the adjacent Layer 2 ATM network.
c. It receives a segment F4 AIS OAM flow on a VP AC, or a segment F5
AIS OAM flow on a VC AC, provided that the operator has
provisioned segment OAM and the PE is not a segment endpoint.
d. It detects loss of connectivity on the ATM VPC/VCC while
terminating segment or end-to-end ATM continuity check (ATM CC)
cells with the local ATM network and CE.
When operating in the coupled OAM loops mode, PE1 exits the AC
receive defect state when all previously detected defects have
disappeared.
When operating in the single emulated OAM loop mode, PE1 enters the
AC receive defect state if any of the following conditions are met:
a. It detects or is notified of a physical layer fault on the ATM
interface.
b. It detects loss of connectivity on the ATM VPC/VCC while
terminating segment ATM continuity check (ATM CC) cells with the
local ATM network and CE.
When operating in the single emulated OAM loop mode, PE1 exits the AC
receive defect state when all previously detected defects have
disappeared.
The exact conditions under which a PE enters and exits the AIS state,
or declares that connectivity is restored via ATM CC, are defined in
Section 9.2 of [I.610].
7.2. AC Transmit Defect State Entry/Exit Criteria
When operating in the coupled OAM loops mode, PE1 enters the AC
transmit defect state if any of the following conditions are met:
a. It terminates an end-to-end F4 RDI OAM flow, in the case of a VPC,
or an end-to-end F5 RDI OAM flow, in the case of a VCC, indicating
that the ATM VPC or VCC is down in the adjacent L2 ATM.
b. It receives a segment F4 RDI OAM flow on a VP AC, or a segment F5
RDI OAM flow on a VC AC, provided that the operator has
provisioned segment OAM and the PE is not a segment endpoint.
PE1 exits the AC transmit defect state if the AC state transitions to
working or to the AC receive defect state. The exact conditions for
exiting the RDI state are described in Section 9.2 of [I.610].
Note that the AC transmit defect state is not valid when operating in
the single emulated OAM loop mode, as PE1 transparently forwards the
received RDI cells as user cells over the ATM PW to the remote CE.
7.3. Consequent Actions
In the remainder of this section, the text refers to AIS, RDI, and CC
without specifying whether there is an F4 (VP-level) flow or an F5
(VC-level) flow, or whether it is an end-to-end or a segment flow.
Precise ATM OAM procedures for each type of flow are specified in
Section 9.2 of [I.610].
7.3.1. PW Receive Defect State Entry/Exit
On entry to the PW receive defect state:
a. PE1 MUST commence AIS insertion into the corresponding AC.
b. PE1 MUST cease generation of CC cells on the corresponding AC, if
applicable.
c. If the PW defect was detected by PE1 without receiving FDI from
PE2, PE1 MUST assume PE2 has no knowledge of the defect and MUST
notify PE2 by sending RDI.
On exit from the PW receive defect state:
a. PE1 MUST cease AIS insertion into the corresponding AC.
b. PE1 MUST resume any CC cell generation on the corresponding AC, if
applicable.
c. PE1 MUST clear the RDI to PE2, if applicable.
7.3.2. PW Transmit Defect State Entry/Exit
On entry to the PW Transmit Defect State:
a. PE1 MUST commence RDI insertion into the corresponding AC.
b. If the PW failure was detected by PE1 without receiving RDI from
PE2, PE1 MUST assume PE2 has no knowledge of the defect and MUST
notify PE2 by sending FDI.
On exit from the PW Transmit Defect State:
a. PE1 MUST cease RDI insertion into the corresponding AC.
b. PE1 MUST clear the FDI to PE2, if applicable.
7.3.3. PW Defect State in ATM Port Mode PW Service
In case of transparent cell transport PW service, i.e., "port mode",
where the PE does not keep track of the status of individual ATM VPCs
or VCCs, a PE cannot relay PW defect state over these VCCs and VPCs.
If ATM CC is run on the VCCs and VPCs end-to-end (CE1 to CE2), or on
a segment originating and terminating in the ATM network and spanning
the PSN network, it will time out and cause the CE or ATM switch to
enter the ATM AIS state.
7.3.4. AC Receive Defect State Entry/Exit
On entry to the AC receive defect state and when operating in the
coupled OAM loops mode:
a. PE1 MUST send FDI to PE2.
b. PE1 MUST commence insertion of ATM RDI cells into the AC towards
CE1.
When operating in the single emulated OAM loop mode, PE1 must be able
to support two options, subject to the operator's preference. The
default option is the following:
On entry to the AC receive defect state:
a. PE1 MUST transparently relay ATM AIS cells, or, in the case of a
local AC defect, commence insertion of ATM AIS cells into the
corresponding PW towards CE2.
b. If the defect interferes with NS OAM message generation, PE1 MUST
send FDI to PE2.
c. PE1 MUST cease the generation of CC cells on the corresponding PW,
if applicable.
In certain operational models, for example, in the case that the ATM
access network is owned by a different provider than the PW, an
operator may want to distinguish between defects detected in the ATM
access network and defects detected on the AC directly adjacent to
the PE. Therefore, the following option MUST also be supported:
a. PE1 MUST transparently relay ATM AIS cells over the corresponding
PW towards CE2.
b. Upon detection of a defect on the ATM interface on the PE or in
the PE itself, PE1 MUST send FDI to PE2.
c. PE1 MUST cease generation of CC cells on the corresponding PW, if
applicable.
On exit from the AC receive defect state and when operating in the
coupled OAM loops mode:
a. PE1 MUST clear the FDI to PE2.
b. PE1 MUST cease insertion of ATM RDI cells into the AC.
On exit from the AC receive defect state and when operating in the
single emulated OAM loop mode:
a. PE1 MUST cease insertion of ATM AIS cells into the corresponding
PW.
b. PE1 MUST clear the FDI to PE2, if applicable.
c. PE1 MUST resume any CC cell generation on the corresponding PW, if
applicable.
7.3.5. AC Transmit Defect State Entry/Exit
On entry to the AC transmit defect state and when operating in the
coupled OAM loops mode:
* PE1 MUST send RDI to PE2.
On exit from the AC transmit defect state and when operating in the
coupled OAM loops mode:
* PE1 MUST clear the RDI to PE2.
8. Procedures for Frame Relay PW Service
The following procedures apply to Frame Relay (FR) pseudowires
[RFC4619]. Frame Relay (FR) terminology is explained in Appendix A.1
of this document.
8.1. AC Receive Defect State Entry/Exit Criteria
PE1 enters the AC receive defect state if one or more of the
following conditions are met:
a. A Permanent Virtual Circuit (PVC) is not deleted from the FR
network and the FR network explicitly indicates in a full status
report (and optionally by the asynchronous status message) that
this PVC is inactive [Q.933]. In this case, this status maps
across the PE to the corresponding PW only.
b. The Link Integrity Verification (LIV) indicates that the link from
the PE to the Frame Relay network is down [Q.933]. In this case,
the link down indication maps across the PE to all corresponding
PWs.
c. A physical layer alarm is detected on the FR interface. In this
case, this status maps across the PE to all corresponding PWs.
PE1 exits the AC receive defect state when all previously detected
defects have disappeared.
8.2. AC Transmit Defect State Entry/Exit Criteria
The AC transmit defect state is not valid for a FR AC.
8.3. Consequent Actions
8.3.1. PW Receive Defect State Entry/Exit
The A (Active) bit indicates whether the FR PVC is ACTIVE (1) or
INACTIVE (0) as explained in [RFC4591].
On entry to the PW receive defect state:
a. PE1 MUST clear the Active bit for the corresponding FR AC in a
full status report, and optionally in an asynchronous status
message, as per [Q.933], Annex A.
b. If the PW failure was detected by PE1 without receiving FDI from
PE2, PE1 MUST assume PE2 has no knowledge of the defect and MUST
notify PE2 by sending RDI.
On exit from the PW receive defect state:
a. PE1 MUST set the Active bit for the corresponding FR AC in a full
status report, and optionally in an asynchronous status message,
as per [Q.933], Annex A. PE1 does not apply this procedure on a
transition from the PW receive defect state to the PW transmit
defect state.
b. PE1 MUST clear the RDI to PE2, if applicable.
8.3.2. PW Transmit Defect State Entry/Exit
On entry to the PW transmit defect state:
a. PE1 MUST clear the Active bit for the corresponding FR AC in a
full status report, and optionally in an asynchronous status
message, as per [Q.933], Annex A.
b. If the PW failure was detected by PE1 without RDI from PE2, PE1
MUST assume PE2 has no knowledge of the defect and MUST notify PE2
by sending FDI.
On exit from the PW transmit defect state:
a. PE1 MUST set the Active bit for the corresponding FR AC in a full
status report, and optionally in an asynchronous status message,
as per [Q.933], Annex A. PE1 does not apply this procedure on a
transition from the PW transmit defect state to the PW receive
defect state.
b. PE1 MUST clear the FDI to PE2, if applicable.
8.3.3. PW Defect State in the FR Port Mode PW Service
In case of port mode PW service, STATUS ENQUIRY and STATUS messages
are transported transparently over the PW. A PW Failure will
therefore result in timeouts of the Q.933 link and PVC management
protocol at the Frame Relay devices at one or both sites of the
emulated interface.
8.3.4. AC Receive Defect State Entry/Exit
On entry to the AC receive defect state:
* PE1 MUST send FDI to PE2.
On exit from the AC receive defect state:
* PE1 MUST clear the FDI to PE2.
8.3.5. AC Transmit Defect State Entry/Exit
The AC transmit defect state is not valid for an FR AC.
9. Procedures for TDM PW Service
The following procedures apply to SAToP [RFC4553], CESoPSN [RFC5086]
and TDMoIP [RFC5087]. These technologies utilize the single emulated
OAM loop mode. RFC 5087 distinguishes between trail-extended and
trail-terminated scenarios; the former is essentially the single
emulated loop model. The latter applies to cases where the NS
networks are run by different operators and defect notifications are
not propagated across the PW.
Since TDM is inherently real-time in nature, many OAM indications
must be generated or forwarded with minimal delay. This requirement
rules out the use of messaging protocols, such as PW status messages.
Thus, for TDM PWs, alternate mechanisms are employed.
The fact that TDM PW packets are sent at a known constant rate can be
exploited as an OAM mechanism. Thus, a PE enters the PW receive
defect state whenever a preconfigured number of TDM PW packets do not
arrive in a timely fashion. It exits this state when packets once
again arrive at their proper rate.
Native TDM carries OAM indications in overhead fields that travel
along with the data. TDM PWs emulate this behavior by sending urgent
OAM messages in the PWE control word.
The TDM PWE3 control word contains a set of flags used to indicate PW
and AC defect conditions. The L bit is an AC forward defect
indication used by the upstream PE to signal NS network defects to
the downstream PE. The M field may be used to modify the meaning of
receive defects. The R bit is a PW reverse defect indication used by
the PE to signal PSN failures to the remote PE. Upon reception of
packets with the R bit set, a PE enters the PW transmit defect state.
L bits and R bits are further described in [RFC5087].
9.1. AC Receive Defect State Entry/Exit Criteria
PE1 enters the AC receive defect state if any of the following
conditions are met:
a. It detects a physical layer fault on the TDM interface (Loss of
Signal, Loss of Alignment, etc., as described in [G.775]).
b. It is notified of a previous physical layer fault by detecting
AIS.
The exact conditions under which a PE enters and exits the AIS state
are defined in [G.775]. Note that Loss of Signal and AIS detection
can be performed by PEs for both structure-agnostic and structure-
aware TDM PW types. Note that PEs implementing structure-agnostic
PWs cannot detect Loss of Alignment.
9.2. AC Transmit Defect State Entry/Exit Criteria
PE1 enters the AC transmit defect state when it detects RDI according
to the criteria in [G.775]. Note that PEs implementing structure-
agnostic PWs cannot detect RDI.
9.3. Consequent Actions
9.3.1. PW Receive Defect State Entry/Exit
On entry to the PW receive defect state:
a. PE1 MUST commence AIS insertion into the corresponding TDM AC.
b. PE1 MUST set the R bit in all PW packets sent back to PE2.
On exit from the PW receive defect state:
a. PE1 MUST cease AIS insertion into the corresponding TDM AC.
b. PE1 MUST clear the R bit in all PW packets sent back to PE2.
Note that AIS generation can, in general, be performed by both
structure-aware and structure-agnostic PEs.
9.3.2. PW Transmit Defect State Entry/Exit
On entry to the PW Transmit Defect State:
* A structure-aware PE1 MUST commence RDI insertion into the
corresponding AC.
On exit from the PW Transmit Defect State:
* A structure-aware PE1 MUST cease RDI insertion into the
corresponding AC.
Note that structure-agnostic PEs are not capable of injecting RDI
into an AC.
9.3.3. AC Receive Defect State Entry/Exit
On entry to the AC receive defect state and when operating in the
single emulated OAM loop mode:
a. PE1 SHOULD overwrite the TDM data with AIS in the PW packets sent
towards PE2.
b. PE1 MUST set the L bit in these packets.
c. PE1 MAY omit the payload in order to conserve bandwidth.
d. A structure-aware PE1 SHOULD send RDI back towards CE1.
e. A structure-aware PE1 that detects a potentially correctable AC
defect MAY use the M field to indicate this.
On exit from the AC receive defect state and when operating in the
single emulated OAM loop mode:
a. PE1 MUST cease overwriting PW content with AIS and return to
forwarding valid TDM data in PW packets sent towards PE2.
b. PE1 MUST clear the L bit in PW packets sent towards PE2.
c. A structure-aware PE1 MUST cease sending RDI towards CE1.
10. Procedures for CEP PW Service
The following procedures apply to SONET/SDH Circuit Emulation
[RFC4842]. They are based on the single emulated OAM loop mode.
Since SONET and SDH are inherently real-time in nature, many OAM
indications must be generated or forwarded with minimal delay. This
requirement rules out the use of messaging protocols, such as PW
status messages. Thus, for CEP PWs alternate mechanisms are
employed.
The CEP PWE3 control word contains a set of flags used to indicate PW
and AC defect conditions. The L bit is a forward defect indication
used by the upstream PE to signal to the downstream PE a defect in
its local attachment circuit. The R bit is a PW reverse defect
indication used by the PE to signal PSN failures to the remote PE.
The combination of N and P bits is used by the local PE to signal
loss of pointer to the remote PE.
The fact that CEP PW packets are sent at a known constant rate can be
exploited as an OAM mechanism. Thus, a PE enters the PW receive
defect state when it loses packet synchronization. It exits this
state when it regains packet synchronization. See [RFC4842] for
further details.
10.1. Defect States
10.1.1. PW Receive Defect State Entry/Exit
In addition to the conditions specified in Section 6.2.1, PE1 will
enter the PW receive defect state when one of the following becomes
true:
o It receives packets with the L bit set.
o It receives packets with both the N and P bits set.
o It loses packet synchronization.
10.1.2. PW Transmit Defect State Entry/Exit
In addition to the conditions specified in Section 6.2.2, PE1 will
enter the PW transmit defect state if it receives packets with the R
bit set.
10.1.3. AC Receive Defect State Entry/Exit
PE1 enters the AC receive defect state when any of the following
conditions are met:
a. It detects a physical layer fault on the TDM interface (Loss of
Signal, Loss of Alignment, etc.).
b. It is notified of a previous physical layer fault by detecting of
AIS.
The exact conditions under which a PE enters and exits the AIS state
are defined in [G.707] and [G.783].
10.1.4. AC Transmit Defect State Entry/Exit
The AC transmit defect state is not valid for CEP PWs. RDI signals
are forwarded transparently.
10.2. Consequent Actions
10.2.1. PW Receive Defect State Entry/Exit
On entry to the PW receive defect state:
a. PE1 MUST commence AIS-P/V insertion into the corresponding AC.
See [RFC4842].
b. PE1 MUST set the R bit in all PW packets sent back to PE2.
On exit from the PW receive defect state:
a. PE1 MUST cease AIS-P/V insertion into the corresponding AC.
b. PE1 MUST clear the R bit in all PW packets sent back to PE2.
See [RFC4842] for further details.
10.2.2. PW Transmit Defect State Entry/Exit
On entry to the PW Transmit Defect State:
a. A structure-aware PE1 MUST commence RDI insertion into the
corresponding AC.
On exit from the PW Transmit Defect State:
a. A structure-aware PE1 MUST cease RDI insertion into the
corresponding AC.
10.2.3. AC Receive Defect State Entry/Exit
On entry to the AC receive defect state:
a. PE1 MUST set the L bit in these packets.
b. If Dynamic Bandwidth Allocation (DBA) has been enabled, PE1 MAY
omit the payload in order to conserve bandwidth.
c. If Dynamic Bandwidth Allocation (DBA) is not enabled, PE1 SHOULD
insert AIS-V/P in the SDH/SONET client layer in the PW packets
sent towards PE2.
On exit from the AC receive defect state:
a. PE1 MUST cease overwriting PW content with AIS-P/V and return to
forwarding valid data in PW packets sent towards PE2.
b. PE1 MUST clear the L bit in PW packets sent towards PE2.
See [RFC4842] for further details.
11. Security Considerations
The mapping messages described in this document do not change the
security functions inherent in the actual messages. All generic
security considerations applicable to PW traffic specified in Section
10 of [RFC3985] are applicable to NS OAM messages transferred inside
the PW.
Security considerations in Section 10 of RFC 5085 for VCCV apply to
the OAM messages thus transferred. Security considerations
applicable to the PWE3 control protocol of RFC 4447 Section 8.2 apply
to OAM indications transferred using the LDP status message.
Since the mechanisms of this document enable propagation of OAM
messages and fault conditions between native service networks and
PSNs, continuity of the end-to-end service depends on a trust
relationship between the operators of these networks. Security
considerations for such scenarios are discussed in Section 7 of
[RFC5254].
12. Contributors and Acknowledgments
Mustapha Aissaoui, Peter Busschbach, Luca Martini, Monique Morrow,
Thomas Nadeau, and Yaakov (J) Stein, were each, in turn, editors of
one or more revisions of this document. All of the above were
contributing authors, as was Dave Allan, david.i.allan@ericsson.com.
The following contributed significant ideas or text:
Matthew Bocci, matthew.bocci@alcatel-lucent.co.uk
Simon Delord, Simon.A.DeLord@team.telstra.com
Yuichi Ikejiri, y.ikejiri@ntt.com
Kenji Kumaki, kekumaki@kddi.com
Satoru Matsushima, satoru.matsushima@tm.softbank.co.jp
Teruyuki Oya, teruyuki.oya@tm.softbank.co.jp
Carlos Pignataro, cpignata@cisco.com
Vasile Radoaca, vasile.radoaca@alcatel-lucent.com
Himanshu Shah, hshah@ciena.com
David Watkinson, david.watkinson@alcatel-lucent.com
The editors would like to acknowledge the contributions of Bertrand
Duvivier, Adrian Farrel, Tiberiu Grigoriu, Ron Insler, Michel
Khouderchah, Vanson Lim, Amir Maleki, Neil McGill, Chris Metz, Hari
Rakotoranto, Eric Rosen, Mark Townsley, and Ben Washam.
13. References
13.1. Normative References
[RFC2119] Bradner, S., "Key words for use in RFCs to
Indicate Requirement Levels", BCP 14, RFC 2119,
March 1997.
[RFC4379] Kompella, K. and G. Swallow, "Detecting Multi-
Protocol Label Switched (MPLS) Data Plane
Failures", RFC 4379, February 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 YJ. Stein, "Structure-Agnostic
Time Division Multiplexing (TDM) over Packet
(SAToP)", RFC 4553, June 2006.
[RFC4591] Townsley, M., Wilkie, G., Booth, S., Bryant, S.,
and J. Lau, "Frame Relay over Layer 2 Tunneling
Protocol Version 3 (L2TPv3)", RFC 4591,
August 2006.
[RFC4619] Martini, L., Kawa, C., and A. Malis,
"Encapsulation Methods for Transport of Frame
Relay over Multiprotocol Label Switching (MPLS)
Networks", RFC 4619, September 2006.
[RFC4717] Martini, L., Jayakumar, J., Bocci, M., El-Aawar,
N., Brayley, J., and G. Koleyni, "Encapsulation
Methods for Transport of Asynchronous Transfer
Mode (ATM) over MPLS Networks", RFC 4717,
December 2006.
[RFC4842] Malis, A., Pate, P., Cohen, R., and D. Zelig,
"Synchronous Optical Network/Synchronous Digital
Hierarchy (SONET/SDH) Circuit Emulation over
Packet (CEP)", RFC 4842, April 2007.
[RFC5036] Andersson, L., Minei, I., and B. Thomas, "LDP
Specification", RFC 5036, October 2007.
[RFC5085] Nadeau, T. and C. Pignataro, "Pseudowire Virtual
Circuit Connectivity Verification (VCCV): A
Control Channel for Pseudowires", RFC 5085,
December 2007.
[RFC5641] McGill, N. and C. Pignataro, "Layer 2 Tunneling
Protocol Version 3 (L2TPv3) Extended Circuit
Status Values", RFC 5641, August 2009.
[RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding
Detection (BFD)", RFC 5880, June 2010.
[RFC5885] Nadeau, T. and C. Pignataro, "Bidirectional
Forwarding Detection (BFD) for the Pseudowire
Virtual Circuit Connectivity Verification
(VCCV)", RFC 5885, June 2010.
[G.707] "Network node interface for the synchronous
digital hierarchy", ITU-T Recommendation G.707,
December 2003.
[G.775] "Loss of Signal (LOS), Alarm Indication Signal
(AIS) and Remote Defect Indication (RDI) defect
detection and clearance criteria for PDH
signals", ITU-T Recommendation G.775,
October 1998.
[G.783] "Characteristics of synchronous digital hierarchy
(SDH) equipment functional blocks", ITU-
T Recommendation G.783, March 2006.
[I.610] "B-ISDN operation and maintenance principles and
functions", ITU-T Recommendation I.610,
February 1999.
[Q.933] "ISDN Digital Subscriber Signalling System No. 1
(DSS1) Signalling specifications for frame mode
switched and permanent virtual connection control
and status monitoring", ITU- T Recommendation
Q.993, February 2003.
13.2. Informative References
[RFC0792] Postel, J., "Internet Control Message Protocol",
STD 5, RFC 792, September 1981.
[RFC3031] Rosen, E., Viswanathan, A., and R. Callon,
"Multiprotocol Label Switching Architecture",
RFC 3031, January 2001.
[RFC3209] Awduche, D., Berger, L., Gan, D., Li, T.,
Srinivasan, V., and G. Swallow, "RSVP-TE:
Extensions to RSVP for LSP Tunnels", RFC 3209,
December 2001.
[RFC3916] Xiao, X., McPherson, D., and P. Pate,
"Requirements for Pseudo-Wire Emulation Edge-to-
Edge (PWE3)", RFC 3916, September 2004.
[RFC3931] Lau, J., Townsley, M., and I. Goyret, "Layer Two
Tunneling Protocol - Version 3 (L2TPv3)",
RFC 3931, March 2005.
[RFC3985] Bryant, S. and P. Pate, "Pseudo Wire Emulation
Edge-to-Edge (PWE3) Architecture", RFC 3985,
March 2005.
[RFC4023] Worster, T., Rekhter, Y., and E. Rosen,
"Encapsulating MPLS in IP or Generic Routing
Encapsulation (GRE)", RFC 4023, March 2005.
[RFC4377] Nadeau, T., Morrow, M., Swallow, G., Allan, D.,
and S. Matsushima, "Operations and Management
(OAM) Requirements for Multi-Protocol Label
Switched (MPLS) Networks", RFC 4377,
February 2006.
[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.
[RFC4446] Martini, L., "IANA Allocations for Pseudowire
Edge to Edge Emulation (PWE3)", BCP 116,
RFC 4446, April 2006.
[RFC4454] Singh, S., Townsley, M., and C. Pignataro,
"Asynchronous Transfer Mode (ATM) over Layer 2
Tunneling Protocol Version 3 (L2TPv3)", RFC 4454,
May 2006.
[RFC5086] Vainshtein, A., Sasson, I., Metz, E., Frost, T.,
and P. Pate, "Structure-Aware Time Division
Multiplexed (TDM) Circuit Emulation Service over
Packet Switched Network (CESoPSN)", RFC 5086,
December 2007.
[RFC5087] Stein, Y(J)., Shashoua, R., Insler, R., and M.
Anavi, "Time Division Multiplexing over IP
(TDMoIP)", RFC 5087, December 2007.
[RFC5254] Bitar, N., Bocci, M., and L. Martini,
"Requirements for Multi-Segment Pseudowire
Emulation Edge-to-Edge (PWE3)", RFC 5254,
October 2008.
[RFC6073] Martini, L., Metz, C., Nadeau, T., Bocci, M., and
M. Aissaoui, "Segmented Pseudowire", RFC 6073,
January 2011.
[Eth-OAM-Inter] Mohan, D., Bitar, N., DeLord, S., Niger, P.,
Sajassi, A., and R. Qiu, "MPLS and Ethernet OAM
Interworking", Work in Progress, March 2011.
[Static-PW-Status] Martini, L., Swallow, G., Heron, G., and M.
Bocci, "Pseudowire Status for Static
Pseudowires", Work in Progress, June 2011.
[I.620] "Frame relay operation and maintenance principles
and functions", ITU-T Recommendation I.620,
October 1996.
Appendix A. Native Service Management (Informative)
A.1. Frame Relay Management
The management of Frame Relay Bearer Service (FRBS) connections can
be accomplished through two distinct methodologies:
a. Based on [Q.933], Annex A, Link Integrity Verification procedure,
where STATUS and STATUS ENQUIRY signaling messages are sent using
DLCI=0 over a given User-Network Interface (UNI) and Network-
Network Interface (NNI) physical link.
b. Based on FRBS Local Management Interface (LMI), and similar to ATM
Integrated LMI (ILMI) where LMI is common in private Frame Relay
networks.
In addition, ITU-T I.620 [I.620] addressed Frame Relay loopback.
This Recommendation was withdrawn in 2004, and its deployment was
limited.
It is possible to use either, or both, of the above options to manage
Frame Relay interfaces. This document will refer exclusively to
Q.933 messages.
The status of any provisioned Frame Relay PVC may be updated through:
a. Frame Relay STATUS messages in response to Frame Relay STATUS
ENQUIRY messages; these are mandatory.
b. Optional unsolicited STATUS updates independent of STATUS ENQUIRY
(typically, under the control of management system, these updates
can be sent periodically (continuous monitoring) or only upon
detection of specific defects based on configuration).
In Frame Relay, a Data Link Connection (DLC) is either up or down.
There is no distinction between different directions. To achieve
commonality with other technologies, down is represented as a receive
defect.
Frame Relay connection management is not implemented over the PW
using either of the techniques native to FR; therefore, PW mechanisms
are used to synchronize the view each PE has of the remote Native
Service/Attachment Circuit (NS/AC). A PE will treat a remote NS/AC
failure in the same way it would treat a PW or PSN failure, that is,
using AC facing FR connection management to notify the CE that FR is
down.
A.2. ATM Management
ATM management and OAM mechanisms are much more evolved than those of
Frame Relay. There are five broad management-related categories,
including fault management (FT), Performance management (PM),
configuration management (CM), Accounting management (AC), and
Security management (SM). [I.610] describes the functions for the
operation and maintenance of the physical layer and the ATM layer,
that is, management at the bit and cell levels. Because of its
scope, this document will concentrate on ATM fault management
functions. Fault management functions include the following:
a. Alarm Indication Signal (AIS).
b. Remote Defect Indication (RDI).
c. Continuity Check (CC).
d. Loopback (LB).
Some of the basic ATM fault management functions are described as
follows: Alarm Indication Signal (AIS) sends a message in the same
direction as that of the signal, to the effect that an error has been
detected.
The Remote Defect Indication (RDI) sends a message to the
transmitting terminal that an error has been detected. Alarms
related to the physical layer are indicated using path AIS/RDI.
Virtual path AIS/RDI and virtual channel AIS/RDI are also generated
for the ATM layer.
OAM cells (F4 and F5 cells) are used to instrument virtual paths and
virtual channels, respectively, with regard to their performance and
availability. OAM cells in the F4 and F5 flows are used for
monitoring a segment of the network and end-to-end monitoring. OAM
cells in F4 flows have the same VPI as that of the connection being
monitored. OAM cells in F5 flows have the same VPI and VCI as that
of the connection being monitored. The AIS and RDI messages of the
F4 and F5 flows are sent to the other network nodes via the VPC or
the VCC to which the message refers. The type of error and its
location can be indicated in the OAM cells. Continuity check is
another fault management function. To check whether a VCC that has
been idle for a period of time is still functioning, the network
elements can send continuity-check cells along that VCC.
Appendix B. PW Defects and Detection Tools
B.1. PW Defects
Possible defects that impact PWs are the following:
a. Physical layer defect in the PSN interface.
b. PSN tunnel failure that results in a loss of connectivity between
ingress and egress PE.
c. Control session failures between ingress and egress PE.
In case of an MPLS PSN and an MPLS/IP PSN there are additional
defects:
a. PW labeling error, which is due to a defect in the ingress PE, or
to an over-writing of the PW label value somewhere along the LSP
path.
b. LSP tunnel label swapping errors or LSP tunnel label merging
errors in the MPLS network. This could result in the termination
of a PW at the wrong egress PE.
c. Unintended self-replication; e.g., due to loops or denial-of-
service attacks.
B.2. Packet Loss
Persistent congestion in the PSN or in a PE could impact the proper
operation of the emulated service.
A PE can detect packet loss resulting from congestion through several
methods. If a PE uses the sequence number field in the PWE3 Control
Word for a specific pseudowire [RFC3985] and [RFC4385], it has the
ability to detect packet loss. Translation of congestion detection
to PW defect states is beyond the scope of this document.
There are congestion alarms that are raised in the node and to the
management system when congestion occurs. The decision to declare
the PW down and to select another path is usually at the discretion
of the network operator.
B.3. PW Defect Detection Tools
To detect the defects listed above, Service Providers have a variety
of options available.
Physical Layer defect detection and notification mechanisms include
SONET/SDH Loss of Signal (LOS), Loss of Alignment (LOA), and AIS/RDI.
PSN defect detection mechanisms vary according to the PSN type.
For PWs over L2TPv3/IP PSNs, with L2TP as encapsulation protocol, the
defect detection mechanisms described in [RFC3931] apply. These
include, for example, the keep-alive mechanism performed with Hello
messages for detection of loss of connectivity between a pair of
LCCEs (i.e., dead PE peer and path detection). Furthermore, the
tools Ping and Traceroute, based on ICMP Echo Messages [RFC0792]
apply and can be used to detect defects on the IP PSN. Additionally,
VCCV-Ping [RFC5085] and VCCV-BFD [RFC5885] can also be used to detect
defects at the individual pseudowire level.
For PWs over MPLS or MPLS/IP PSNs, several tools can be used:
a. LSP-Ping and LSP-Traceroute [RFC4379] for LSP tunnel connectivity
verification.
b. LSP-Ping with Bi-directional Forwarding Detection [RFC5885] for
LSP tunnel continuity checking.
c. Furthermore, if Resource Reservation Protocol - Traffic
Engineering (RSVP-TE) is used to set up the PSN Tunnels between
ingress and egress PE, the hello protocol can be used to detect
loss of connectivity [RFC3209], but only at the control plane.
B.4. PW Specific Defect Detection Mechanisms
[RFC4377] describes how LSP-Ping and BFD can be used over individual
PWs for connectivity verification and continuity checking,
respectively.
Furthermore, the detection of a fault could occur at different points
in the network and there are several ways the observing PE determines
a fault exists:
a. Egress PE detection of failure (e.g., BFD).
b. Ingress PE detection of failure (e.g., LSP-PING).
c. Ingress PE notification of failure (e.g., RSVP Path-err).
Authors' Addresses
Mustapha Aissaoui
Alcatel-Lucent
600 March Rd
Kanata, ON K2K 2E6
Canada
EMail: mustapha.aissaoui@alcatel-lucent.com
Peter Busschbach
Alcatel-Lucent
67 Whippany Rd
Whippany, NJ 07981
USA
EMail: busschbach@alcatel-lucent.com
Luca Martini
Cisco Systems, Inc.
9155 East Nichols Avenue, Suite 400
Englewood, CO 80112
USA
EMail: lmartini@cisco.com
Monique Morrow
Cisco Systems, Inc.
Richtistrase 7
CH-8304 Wallisellen
Switzerland
EMail: mmorrow@cisco.com
Thomas Nadeau
CA Technologies
273 Corporate Dr.
Portsmouth, NH 03801
USA
EMail: Thomas.Nadeau@ca.com
Yaakov (Jonathan) Stein
RAD Data Communications
24 Raoul Wallenberg St., Bldg C
Tel Aviv 69719
Israel
EMail: yaakov_s@rad.com