AC-influenced Designated Forwarder Election for (PBB-)EVPN
draft-rabadan-bess-evpn-ac-df-00
The information below is for an old version of the document.
| Document | Type |
This is an older version of an Internet-Draft whose latest revision state is "Replaced".
|
|
|---|---|---|---|
| Authors | Jorge Rabadan , Kiran Nagaraj , Senthil Sathappan , Wim Henderickx | ||
| Last updated | 2014-10-24 | ||
| Replaced by | draft-ietf-bess-evpn-ac-df | ||
| RFC stream | (None) | ||
| Formats | |||
| Stream | Stream state | (No stream defined) | |
| Consensus boilerplate | Unknown | ||
| RFC Editor Note | (None) | ||
| IESG | IESG state | I-D Exists | |
| Telechat date | (None) | ||
| Responsible AD | (None) | ||
| Send notices to | (None) |
draft-rabadan-bess-evpn-ac-df-00
BESS Workgroup J. Rabadan
Internet Draft K. Nagaraj
S. Sathappan
Intended status: Standards Track W. Henderickx
Alcatel-Lucent
Expires: April 27, 2015 October 24, 2014
AC-influenced Designated Forwarder Election for (PBB-)EVPN
draft-rabadan-bess-evpn-ac-df-00
Abstract
The Designated Forwarder (DF) in (PBB-)EVPN networks is the PE
responsible for sending multicast, broadcast and unknown unicast
traffic to a multi-homed CE, on a given Ethernet Tag on a particular
Ethernet Segment (ES). The DF is selected based on the list of PEs
that advertise the Ethernet Segment Identifier (ESI) to the EVPN
network. While PE node or link failures trigger the DF re-election
for a given <ESI, EVI>, individual Attachment Circuit (AC) or MAC-VRF
failures do not trigger such DF re-election and the traffic may
therefore be permanently impacted, even though there is an
alternative path. This document improves the DF election algorithm so
that the AC status can influence the result of the election and this
type of "logical" failures can be protected too.
Status of this Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt
Rabadan et al. Expires April 27, 2015 [Page 1]
Internet-Draft AC-based DF election for EVPN October 24, 2014
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html
This Internet-Draft will expire on April 27, 2015.
Copyright Notice
Copyright (c) 2014 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.
Table of Contents
1. Problem Statement . . . . . . . . . . . . . . . . . . . . . . . 2
2. Solution description . . . . . . . . . . . . . . . . . . . . . 4
2.1. AC-influenced DF Election for EVPN . . . . . . . . . . . . 4
2.1.1. Current DF election procedure and AC failures . . . . . 5
2.1.2. The Attachment Circuit (AC) influenced DF election . . 6
2.2. AC-influenced DF Election for PBB-EVPN . . . . . . . . . . 7
2.2.1. Current DF election procedure and AC failures . . . . . 8
2.2.2. BGP encoding for advertising ACS per <ESI, ISID> . . . 9
2.2.3. The AC-influenced DF Election . . . . . . . . . . . . . 10
3. Solution benefits . . . . . . . . . . . . . . . . . . . . . . . 11
4. Conventions used in this document . . . . . . . . . . . . . . . 11
5. Security Considerations . . . . . . . . . . . . . . . . . . . . 11
6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 12
7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 12
7.1. Normative References . . . . . . . . . . . . . . . . . . . 12
7.2. Informative References . . . . . . . . . . . . . . . . . . 12
8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . 12
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 12
1. Problem Statement
[EVPN] defines the Designated Forwarder (DF) as the (PBB-)EVPN PE
responsible for:
Rabadan et al. Expires April 27, 2015 [Page 2]
Internet-Draft AC-based DF election for EVPN October 24, 2014
o Flooding Broadcast, Unknown unicast and Multicast traffic (BUM), on
a given Ethernet Tag on a particular Ethernet Segment (ES), to the
CE. This is valid for single-active and all-active EVPN
multi-homing.
o Sending unicast traffic on a given Ethernet Tag on a particular ES
to the CE. This is valid for single-active multi-homing.
The default DF election algorithm defined by [EVPN] is called
service-carving and, for a given ESI, is based on a (V mod N)= i
function that provides a local DF election of a PEi at <ESI, EVI>
level (for EVPN) and at <ESI, ISID> level for PBB-EVPN. V is the
Ethernet Tag associated to the EVI (the numerically lowest Ethernet
Tag value in case of multiple Ethernet Tags), whereas N is the number
of PEs for which ES routes have been successfully imported. In other
words, EVPN's service-carving takes into account only two variables
in the DF election for a given ESI: the existence of the PE's IP
address on the candidate list and the locally provisioned Ethernet
Tags (ISID identifiers in the case of PBB-EVPN).
If the DF for an <ESI, EVI/ISID> fails (due to physical link/node
failures) an ES route withdrawn will make the Non-DF (NDF) PEs re-
elect the DF for that <ESI, EVI/ISID> and the service will be
recovered.
However the current DF election procedure does not provide a
protection against "logical" failures or human errors that may occur
at service level on the DF, while the list of active PEs for a given
ESI does not change. These failures may have an impact not only on
the local PE where the issue happens, but also on the rest of the PEs
of the ESI. Some examples of such logical failures are listed below:
a) A given individual Attachment Circuit (AC) defined in an ESI is
accidentally shutdown or even not provisioned yet (hence the
Attachment Circuit Status - ACS - is DOWN), while the ESI is
operationally active (since the ES route is active).
b) A given MAC-VRF - with an ESI defined - is shutdown or not
provisioned yet, while the ESI is operationally active (since the
ES route is active). In this case, the ACS of all the AC defined
in that MAC-VRF is considered to be DOWN.
Neither (a) nor (b) will trigger the DF re-election on the remote PEs
for a given ESI since the ACS is not taken into account in the DF
election procedures. While the ACS is used as a DF election tie-
breaker and trigger in [VPLS-MH], there is no procedure defined in
[EVPN] to trigger the DF re-election based on the ACS change on the
DF.
Rabadan et al. Expires April 27, 2015 [Page 3]
Internet-Draft AC-based DF election for EVPN October 24, 2014
This document improves the [EVPN] service-carving procedure so that
the ACS may be taken into account as a variable in the DF election,
and therefore EVPN can provide protection against logical failures.
2. Solution description
The ACS for a given Ethernet Tag on an ESI is implicitly conveyed in
the corresponding EVPN A-D per EVI route for that given <ESI,
Ethernet Tag>. This section describes how to use the A-D per EVI
routes to improve the DF election algorithm in both cases, EVPN and
PBB-EVPN.
2.1. AC-influenced DF Election for EVPN
Figure 1 illustrates an example EVPN network that will be used to
describe the proposed solution for EVPN.
EVI-1 is defined in PE-1, PE-2, PE-3 and PE-4. CE12 is a multi-homed
CE connected to ESI12 in PE-1 and PE-2. Similarly CE23 is multi-homed
to PE-2 and PE-3 using ESI23. CE12-VID 1 (VLAN ID 1 on CE12) is
associated to AC1 and AC2 in EVI-1, whereas CE23-VID 1 is associated
to AC3 and AC4 in EVI-1. Note that there are other ACs defined on
these ESIs mapped to different EVIs.
Rabadan et al. Expires April 27, 2015 [Page 4]
Internet-Draft AC-based DF election for EVPN October 24, 2014
+---+
|CE4|
+---+
|
PE-4 |
+-----+-----+
+---------------| +-----+ |---------------+
| | |EVI-1| | |
| +-----------+ |
| |
| EVPN |
| |
| PE-1 PE-2 PE-3 |
| (NDF) (DF) (NDF)|
+-----------+ +-----------+ +-----------+
| |EVI-1| | | |EVI-1| | | |EVI-1| |
| +-----+ |-------| +-----+ |-------| +-----+ |
+-----------+ +-----------+ +-----------+
AC1\ ESI12 /AC2 AC3\ ESI23 /AC4
\ / \ /
\ / \ /
+----+ +----+
|CE12| |CE23|
+----+ +----+
Figure 1 EVPN network example
2.1.1. Current DF election procedure and AC failures
After running the service-carving DF election algorithm, PE-2 turns
out to be the DF for ESI12 and ESI23 in EVI-1. The following two
examples illustrate the issues with the existing defined procedure in
[EVPN]:
a) If AC2 is accidentally shutdown or even not configured, CE12
traffic will be impacted. In case of all-active multi-homing, only
the BUM traffic to CE12 will be impacted, whereas for single-active
multi-homing all the traffic to/from CE12 will be discarded. This is
due to the fact that a logical failure in PE-2 AC2 will not trigger
an ES route withdrawn for ESI12 (since there are still other ACs
active on ESI12) and therefore PE-1 will not re-run the DF election
procedures.
b) If EVI-1 is administratively shutdown or even not configured yet
on PE-2, CE12 and CE23 will both be impacted: BUM traffic to both CEs
will be discarded in case of all-active multi-homing and all traffic
will be discarded to/from the CEs in case of single-active
multi-homing. This is due to the fact that PE-1 and PE-3 will not
Rabadan et al. Expires April 27, 2015 [Page 5]
Internet-Draft AC-based DF election for EVPN October 24, 2014
re-run the DF election procedures and will keep assuming PE-2 is the
DF.
2.1.2. The Attachment Circuit (AC) influenced DF election
Modifying the service-carving DF election procedure in the following
way solves the issue:
1. When PE-1 and PE-2 discover ESI12, they advertise an ES route for
ESI12 with the associated ES-import extended community, starting a
timer at the same time. Likewise, PE-2 and PE-3 advertise an ES
route for ESI23 and start a timer.
2. Similarly PE-1 and PE-2 advertise an Ethernet A-D per ES route for
ESI12, and PE-2/PE-3 advertise an Ethernet A-D per ES route for
ESI23.
3. In addition, PE-1/PE-2/PE-3 advertise an Ethernet A-D per EVI
route for AC1, AC2, AC3 and AC4 as soon as the ACs are enabled.
4. When the timer expires, each PE builds an ordered "candidate" list
of the IP addresses of all the PE nodes connected to the Ethernet
Segment (including itself), in increasing numeric order. The
candidate list is based on the Originator Router's IP addresses of
the ES routes, excluding all the PEs for which no Ethernet A-D per
ES route has been received.
5. When electing the DF for a given EVI, a PE will not be considered
candidate until an Ethernet A-D per EVI route has been received
from that PE. In other words, the ACS on the ESI for a given PE
must be UP so that the PE is considered as candidate for a given
EVI. For example, PE-1 will not consider PE-2 as candidate for DF
election for <ESI12, EVI-1> until an Ethernet A-D per EVI route is
not received from PE-2 for <ESI12, EVI-1>.
6. Once the PEs with ACS = DOWN for a given EVI have been eliminated
from the candidate list, the (V mod N) = i function can be applied
for the remaining N candidates, as per [EVPN].
Note that this procedure does not modify the existing EVPN control
plane whatsoever. It only modifies the candidate list of PEs taken
into account for the DF election algorithm defined in [EVPN].
In addition to the procedure described above, the following events
SHALL modify the candidate PE list and trigger the DF re-election in
a PE for a given <ESI,EVI>:
a) Local ESI going DOWN due to a physical failure or reception of an
Rabadan et al. Expires April 27, 2015 [Page 6]
Internet-Draft AC-based DF election for EVPN October 24, 2014
ES route withdraw for that ESI.
b) Local ESI going UP due to its detection/configuration or reception
of a new ES route update for that ESI.
c) Local AC going DOWN/UP.
d) Reception of a new Ethernet A-D per EVI update/withdraw for the
<ESI, EVI>.
e) Reception of a new Ethernet A-D per ES update/withdraw for the
ESI.
2.2. AC-influenced DF Election for PBB-EVPN
Figure 2 illustrates an example PBB-EVPN network that will be used to
describe the proposed solution for PBB-EVPN.
In this case, the B-component EVI-B is defined in BEB-1, BEB-2, BEB-
3 and BEB-4 and runs EVPN. An I-component with ISID-N is also defined
in all the BEBs and attached to EVI-B as per [PBB-EVPN]. As in the
previous section, CE12 and CE 23 are multi-homed CEs connected to
ESI12 and ESI23 respectively, but in this case the ACs are associated
to the I-component on each BEB. Note that there are other ACs defined
on these ESIs mapped to different ISIDs.
Rabadan et al. Expires April 27, 2015 [Page 7]
Internet-Draft AC-based DF election for EVPN October 24, 2014
+---+
|CE4|
+---+
|
BEB-4 |
+-----+-----+
| ISID-N |
| | |
+---------------| +-----+ |---------------+
| | |EVI-B| | |
| +-----------+ |
| |
| PBB-EVPN |
| |
| BEB-1 BEB-2 BEB-3 |
| (NDF) (DF) (NDF) |
+-----------+ +-----------+ +-----------+
| |EVI-B| | | |EVI-B| | | |EVI-B| |
| +-----+ | | +-----+ | | +-----+ |
| | |-------| | |-------| | |
| ISID-N | | ISID-N | | ISID-N |
+-----------+ +-----------+ +-----------+
AC1\ ESI12 /AC2 AC3\ ESI23 /AC4
\ / \ /
\ / \ /
+----+ +----+
|CE12| |CE23|
+----+ +----+
Figure 2 PBB-EVPN network example
2.2.1. Current DF election procedure and AC failures
After running the service-carving DF election algorithm, BEB-2 turns
out to be the DF for ESI12 and ESI23 in ISID-N. The following two
examples illustrate the issues with the existing defined procedure in
[EVPN] when applied to PBB-EVPN:
a) If AC2 is accidentally shutdown or even not configured, CE12
traffic will be impacted - only the BUM traffic to CE12 if all-
active multi-homing or all the traffic in case of single-active
multi-homing. This is due to the fact that a logical failure in
BEB- 2 AC2 will not trigger an ES route withdrawn for ESI12 (since
there are still other ACs active on ESI12) and therefore BEB-1
will not re-run the DF election procedures.
b) If ISID-N is administratively shutdown or even not configured yet
on BEB-2, CE12 and CE23 will both be impacted - again only BUM or
Rabadan et al. Expires April 27, 2015 [Page 8]
Internet-Draft AC-based DF election for EVPN October 24, 2014
all the traffic depending on the multi-homing type. This is due to
the fact that BEB-1 and BEB-3 will not re-run the DF election
procedures and they will keep assuming BEB-2 is the DF.
Note that regardless the B-MAC address assignment the user chooses
for a PE (shared B-MAC for multiple ES or dedicated B-MAC per ES),
the withdrawal of a B-MAC cannot be used as an indication of an
individual AC going UP/DOWN. For instance, assuming a dedicated B-
MAC per ES is used, when AC2 goes down, BEB-2 will not withdraw the
B-MAC for ESI12 since there might be some other ACs in the same ESI
using the B-MAC.
2.2.2. BGP encoding for advertising ACS per <ESI, ISID>
This document proposes the use of an Ethernet A-D per ISID route to
advertise the ACS of a given ISID in a given ESI. Note that the
Ethernet A-D routes are not used in [PBB-EVPN] hence the procedure
described here is OPTIONAL and, if enabled, it MUST be supported in
all the BEBs part of the same ESI. The advertisement of ACS SHOULD be
administratively enabled at system level and MAY be administratively
enabled at ESI level.
When enabled, the BEB will advertise an Ethernet A-D per ISID route
as per [EVPN] with the following fields:
- RD = B-component EVI's RD
- ESI = non-reserved ESI where the ISID is defined
- Ethernet Tag ID = ISID
- MPLS Label = 0
- The ES-Import Route Target used by the ES route for the same ESI
The ES-Import Route Target will be used in the same way it is used
with ES routes. It will only be imported by the BEBs where the ESI is
defined and it is subject to RT Constraint procedures [RFC4684].
In the example of Figure 2, BEB-2 will advertise an Ethernet A-D per
ISID route for AC2 with the following fields:
- RD = EVI-B RD
- ESI = ESI12
- Ethernet Tag ID = ISID-N
And similarly it will send an Ethernet A-D per ISID route for AC3.
If AC2 is shutdown, BEB-2 will withdraw the A-D per ISID route for
AC2. That will be an indication for BEB-1 that AC2's status is DOWN.
If ISID-N is administratively shutdown, BEB-2 will withdraw the A-D
Rabadan et al. Expires April 27, 2015 [Page 9]
Internet-Draft AC-based DF election for EVPN October 24, 2014
per ISID routes for AC2 and AC3. This will be interpreted by BEB-1
and BEB-3 respectively as ACS DOWN indications.
2.2.3. The AC-influenced DF Election
When this procedure is enabled for a given ESI, the BEBs will
advertise Ethernet A-D per ISID routes and the DF election procedure
will be modified in the following way:
1. When BEB-1 and BEB-2 discover ESI12, they advertise an ES route
for ESI12 with the associated ES-import extended community,
starting a timer at the same time. Likewise, BEB-2 and BEB-3
advertise an ES route for ESI23 and start a timer.
2. In addition, BEB-1/BEB-2/BEB-3 advertise an Ethernet A-D per ISID
route for AC1, AC2, AC3 and AC4 as soon as the ACs are enabled.
3. When the timer expires, each PE builds an ordered "candidate" list
of the IP addresses of all the PE nodes connected to the Ethernet
Segment (including itself), in increasing numeric order. The
candidate list is based on the Originator Router's IP addresses of
the ES routes.
4. When electing the DF for a given ISID, a PE will not be considered
candidate until an Ethernet A-D per ISID route has been received
from that PE. In other words, the ACS on the ESI for a given PE
must be UP so that the PE is considered as candidate for a given
ISID. For example, BEB-1 will not consider BEB-2 as candidate for
DF election for <ESI12, ISID-N> until an Ethernet A-D per ISID
route is not received from BEB-2 for <ESI12, ISID-N>.
5. Once the PEs with ACS = DOWN for a given ISID have been eliminated
from the candidate list, the (V mod N) = i function can be applied
for the remaining N candidates, as per [EVPN].
This procedure modifies the PBB-EVPN control plane procedures, but
does not define new routes or attributes. It reuses the components
already existing in [EVPN] in order to define a DF election procedure
that is consistent for EVPN and PBB-EVPN and protects the network
against "logical" failures.
In addition to the procedure described above, the following events
SHALL modify the candidate PE list and trigger the DF re-election in
a PE for a given <ESI,ISID>:
a) Local ESI going DOWN due to a physical failure or reception of an
ES route withdraw for that ESI.
Rabadan et al. Expires April 27, 2015 [Page 10]
Internet-Draft AC-based DF election for EVPN October 24, 2014
b) Local ESI going UP due to its detection/configuration or reception
of a new ES route update for that ESI.
c) Local AC going DOWN/UP.
d) Reception of a new Ethernet A-D per ISID update/withdraw for the
<ESI, ISID>.
3. Solution benefits
The solution described in this document provides the following
benefits:
a) Improves the DF election procedures for EVPN so that failures due
to human errors, logical failures or even delay in provisioning of
Attachment Circuits can be protected by multi-homing.
b) The above benefit is added to EVPN without the need to modify or
add any BGP new attributes or NLRI changes.
c) Improves the DF election procedures for PBB-EVPN so that the same
failures as in (a) can be protected by multi-homing also in a
PBB-EVPN network.
d) The above benefit is provided without adding any BGP new
attributes or NLRI changes. It simply re-uses the existing
Ethernet A-D route in PBB-EVPN. PBB-EVPN implementations not
supporting this document SHOULD ignore Ethernet A-D routes.
4. 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 RFC-2119 [RFC2119].
In this document, these words will appear with that interpretation
only when in ALL CAPS. Lower case uses of these words are not to be
interpreted as carrying RFC-2119 significance.
In this document, the characters ">>" preceding an indented line(s)
indicates a compliance requirement statement using the key words
listed above. This convention aids reviewers in quickly identifying
or finding the explicit compliance requirements of this RFC.
5. Security Considerations
Will be added.
Rabadan et al. Expires April 27, 2015 [Page 11]
Internet-Draft AC-based DF election for EVPN October 24, 2014
6. IANA Considerations
Will be added.
7. References
7.1. Normative References
[RFC4684] Marques, P., Bonica, R., Fang, L., Martini, L., Raszuk,
R., Patel, K., and J. Guichard, "Constrained Route Distribution for
Border Gateway Protocol/MultiProtocol Label Switching (BGP/MPLS)
Internet Protocol (IP) Virtual Private Networks (VPNs)", RFC 4684,
November 2006, <http://www.rfc-editor.org/info/rfc4684>.
7.2. Informative References
[EVPN] Sajassi et al., "BGP MPLS Based Ethernet VPN", draft-ietf-
l2vpn-evpn-11.txt, work in progress, October, 2014
[VPLS-MH] Kothari, Henderickx et al., "BGP based Multi-homing in
Virtual Private LAN Service", draft-ietf-l2vpn-vpls-multihoming-
07.txt, work in progress, July, 2014
8. Acknowledgments
Will be added.
Authors' Addresses
Jorge Rabadan
Alcatel-Lucent
777 E. Middlefield Road
Mountain View, CA 94043 USA
Email: jorge.rabadan@alcatel-lucent.com
Kiran Nagaraj
Alcatel-Lucent
Email: kiran.nagaraj@alcatel-lucent.com
Senthil Sathappan
Alcatel-Lucent
Email: senthil.sathappan@alcatel-lucent.com
Rabadan et al. Expires April 27, 2015 [Page 12]
Internet-Draft AC-based DF election for EVPN October 24, 2014
Wim Henderickx
Alcatel-Lucent
Email: wim.henderickx@alcatel-lucent.com
Rabadan et al. Expires April 27, 2015 [Page 13]