TRILL Working Group H. Zhai
Internet-Draft ZTE
Intended status: Standards Track T. Senevirathne
Expires: February 27, 2013 Cisco Systems
R. Perlman
Intel Labs
D. Eastlake 3rd
M. Zhang
Huawei
F. Hu
ZTE
August 26, 2012
RBridge: Pseudo-Nickname
draft-hu-trill-pseudonode-nickname-03
Abstract
At the edg of TRILL campus, some RBridges provide end-station
services to their attached end stations; these RBridges are called
edge RBridges. To avoid potential frame duplication or loops in
TRILL campus, only one edge RBridge is permitted to provide such
services in a VLAN at all times to its attached end stations even
they are also attached to other edge RBridges. However, in some
application scenarios, for example in Link Aggregation Group (LAG),
more than one edge RBridge is required to provide such services to an
end station even in a VLAN to improve resiliency and maximize the
available network bandwidth, which causes the flip-flopping of the
egress RBridge nickname for such an end station in remote RBridges'
forwarding tables. In this document, the concept of Virtual RBridge,
along with its pseudo-nickname, is introduced to address the above
problem.
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). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at http://datatracker.ietf.org/drafts/current/.
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."
Zhai, et al. Expires February 27, 2013 [Page 1]
Internet-Draft Pseudo-Nickname August 2012
This Internet-Draft will expire on February 27, 2013.
Copyright Notice
Copyright (c) 2012 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.
Zhai, et al. Expires February 27, 2013 [Page 2]
Internet-Draft Pseudo-Nickname August 2012
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1. Terminology and Acronyms . . . . . . . . . . . . . . . . . 5
2. Problem Statement . . . . . . . . . . . . . . . . . . . . . . 5
2.1. Appointed Forwarders on Shared Links . . . . . . . . . . . 5
2.2. Multi-homing and Link Aggregation to TRILL Network . . . . 6
3. Concept of Virtual RBridge and Pseudo-nickname . . . . . . . . 7
3.1. VLAN-x Appointed Forwarder for member interfaces in RBv . 8
3.2. Announcing Pseudo-Nickname of RBv . . . . . . . . . . . . 8
4. Distribution Trees for Member RBridges in RBv . . . . . . . . 8
5. Frame Processing . . . . . . . . . . . . . . . . . . . . . . . 10
5.1. Data Frames . . . . . . . . . . . . . . . . . . . . . . . 10
5.1.1. Native Frames Ingressing . . . . . . . . . . . . . . . 10
5.1.2. TRILL Data Frames Egressing . . . . . . . . . . . . . 10
5.1.2.1. Unicast TRILL Data Frames . . . . . . . . . . . . 10
5.1.2.2. Multi-Destination TRILL Data Frames . . . . . . . 11
6. Member Link Failure in RBv . . . . . . . . . . . . . . . . . . 12
7. OAM Frames . . . . . . . . . . . . . . . . . . . . . . . . . . 12
8. Configuration Consistency . . . . . . . . . . . . . . . . . . 13
9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13
10. Security Considerations . . . . . . . . . . . . . . . . . . . 13
11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 13
12. Normative References . . . . . . . . . . . . . . . . . . . . . 13
Appendix A. Reasons for MAC Sharing among Member RBriges . . . . 14
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 15
Zhai, et al. Expires February 27, 2013 [Page 3]
Internet-Draft Pseudo-Nickname August 2012
1. Introduction
The IETF TRILL protocol [RFC6325] provides optimal pair-wise data
frame forwarding without configuration, safe forwarding even during
periods of temporary loops, and support for multi-pathing of both
unicast and multicast traffic. TRILL accomplishes this by using
[IS-IS] [RFC1195] link state routing and encapsulating traffic using
a header that includes a hop count. The design supports VLANs and
optimization of the distribution of multi-destination frames based on
VLANs and IP derived multicast groups. Devices that implement TRILL
are called RBridges.
In TRILL protocol, RBridges are identified by nicknames (16-bits).
Different RBridge has different nickname(s). At the edge of TRILL
network, some RBridges connect to legacy networks on one side and to
TRILL network on the other side. These RBridges are called edge
RBridges. For the connectivity between the two types of network,
edge RBridges provide end-station services to end stations located in
legacy networks. When receiving a native frame from such a local end
station S, the service edge RBridge RB1 encapsulates the frame in a
TRILL header, addressing the packet to RBridge RBx to which the
destination end station D is attached. The TRILL header contains an
"ingress RBridge nickname" field (RB1's nickname), an "egress RBridge
nickname" field (RBx's nickname), and a hop count. On receiving such
a frame, RBx removes the TRILL header and forwards it onto D in its
native form. Meanwhile, based on the de-capsulation of such a frame,
RBx learns the (ingress RBridge nickname, source MAC address, VLAN
ID) triplet. Edge RBridges maintain such triplets in their
forwarding table for the potential following transmission of native
frames.
Due to failures, reconfiguration and other network dynamics, service
edge RBridge for S may change over from RB1 to another edge RBridge.
In this event, remote traffic addressed to S will be still forwarded
to RB1 by remote RBridge RBx before perceiving this change, and then
the traffic gets dropped at RB1, causing traffic disruption.
Furthermore, to improve resiliency and maximize the available network
bandwidth, an end station typically is multi-homed to several edge
RBridges and treats all the uplink links as a Link Aggregation Group
(LAG) bundle. In this scenario, all those edge RBridges work in an
active-active load sharing model to provide end-station services for
an end station even in same VLAN. When remote RBridge RB2 receives
different frames, which are originated by such an end station S and
ingressed into TRILL campus by different such edge RBridge, flip-
flopping of ingress RBridge nickname for MAC of S will be observed by
RBx during de-capsulating such frames. This flip-flopping will cause
disorder of different frames in traffic, worsening the traffic
disruption.
Zhai, et al. Expires February 27, 2013 [Page 4]
Internet-Draft Pseudo-Nickname August 2012
In this document, concept of Virtual RBridge group, together with its
Pseudo-nickname, is introduced to address the above issues. For a
member RBridge in such a group, it uses the pseudo-nickname of this
group, instead of its own device nickname, as ingress RBridge
nickname when encapsulating a frame to its TRILL form with a TRILL
header. So, in a RBridge Group, even if there are more than one
RBridge providing end-station services for a end station or the
service RBridge changes over from one member RBridge to another in
same set of VLANs, the ingress RBridge nickname for the MAC of this
end station will still remain unchanged in remote RBridges'
forwarding tables.
This document is organized as following: Section 2 is problem
statement, which describes why virtual RBridge and its pseudo-
nickname are required. Section 3 gives the concept of virtual
RBridge. Section 4 describes the consideration for pseudo-nickname
used in ingressing multi-destination frames. Section 5 covers
processing of transit frame traffic when considering pseudo-nickname.
Familiarity with [RFC6325] is assumed in this document.
1.1. Terminology and Acronyms
This document uses the acronyms defined in [RFC6325] and the
following additional acronym:
AF - Appointed Forwarder
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].
When used in lower case, these words convey their typical use in
common language, and are not to be interpreted as described in
[RFC2119].
2. Problem Statement
2.1. Appointed Forwarders on Shared Links
If there are multiple RBridges on a shared link, together with end
stations, only one RBridge is permitted to provide end-station
services in a VLAN at all times for the end stations to avoid
possible frame duplication or loops in TRILL campus. The service
RBridge is called VLAN-x Appointed Forwarder (AF) on a shared link.
However, AF for any set of VLANs on a shared link may change over
Zhai, et al. Expires February 27, 2013 [Page 5]
Internet-Draft Pseudo-Nickname August 2012
from one RBridge to another, due to failure, configurations and other
network dynamics, etc. If such change occurs, local end stations may
not perceive it, so the end station cannot timely notify remote
RBridges to update the correspondence between ingress RBridge
nickname and the MAC of this end station in their forwarding tables.
As a result, remote RBridges may continue to forward traffic to the
previous AF and the traffic may be dropped at the previous egress
RBridge, causing traffic disruption.
2.2. Multi-homing and Link Aggregation to TRILL Network
In order to improve the reliability of connection to TRILL network,
multi-homing technique may be employed by a legacy device, such as a
switch or end host. For example, in Figure 1, switch SW1 multi-homes
to TRILL network by connecting to both RBridge RB1 and RB2 with
respective links. Then the end stations (e.g., S1), attached to SW1,
still get end-station services from TRILL network even if one
connection of SW1 to TRILL network, e.g., SW1-RB1, fails. That is to
say, the service RBridge for S1 changes over from RB1 to RB2 after
the connection of SW1-RB1 fails, which causes traffic disruption
temporarily (e.g., traffic from Sx to S1), similar to AF changing in
Section 2.1.
...........................................
: TRILL Network :
: +-----+ /\-/\-/\-/\-/\ :
+------| RB1 |-----/ \ :
| : +-----+ / \ :
+-----+ : | / Transit \ +-----+ :
S1 o--| SW1 | : | < RBridges >---| RBx |---o Sx
+-----+ : | \ Campus / +-----+ :
| : +-----+ \ / :
+------| RB2 |-----\ / :
: +-----+ \/\-/\-/\-/\-/ :
: :
...........................................
Figure 1
Multi-homing to TRILL Network
Furthermore, SW1 may treat the two links as a LAG (Link Aggregation
Group) bundle, so that the two links form active-active load sharing
model instead of previous active-stanby model. That is to say, in
Figure 1, two service RBridges (e.g., RB1 and RB2) must provide end-
station services simultaneously to S1 in that VLAN. And this will
result in flip-flopping of the ingress RBridge for the MAC of S1 in
remote RBridges' (e.g., RBx) forwarding tables. AS a result, this
flip-flopping will cause much more disorder packets and worsen the
Zhai, et al. Expires February 27, 2013 [Page 6]
Internet-Draft Pseudo-Nickname August 2012
traffic disruption.
Besides switches, end stations can also directly multi-home to TRILL
network and treat the multi-homing links as a LAG bundle. The issue
of traffic disruption also occurs in this scenario if such an end
station balances different traffic load in a same VLAN among the
member links.
In the following sections, concept of RBridge Group, together with
its nickname, is introduced to fix these issues.
3. Concept of Virtual RBridge and Pseudo-nickname
A Virtual RBridge (RBv) represents a group of different ports on
different edge RBridges, on which these RBridges provide end-station
service to a set of their attached end stations. After joining RBv,
such an RBridge port is called a member port of RBv, and such an
RBridge becomes a member RBridge of RBv. In an RBridge RBv is
identified by its virtual nickname in TRILL campus, and virtual
nickname is also referred to as pseudo-nickname in this
specification.
After joining an RBv, a member RBridge will announce its connection
to RBv by including the information of that RBv, e.g., the pseudo-
nickname of RBv, in its self-originated LSP. From such LSPs, remote
RBridges that are not a members of RBv can deduce that, one or more
shortest paths are available from itself to RBv.
When receiving a native frame on such a port, the member RBridge uses
the RBv's nickname, instead of its own nickname, as ingress nickname
in TRILL header if necessary to encapsulate the frame into TRILL data
form. By de-capsulating such a TRILL-encapsulated data frame, a
remote RBridge learns that S is reachable through RBv.
NOTE: An RBridge port can join at most one RBv at any time, but
different ports on the same RBridge can join the same RBv or
different RBvs. After joining an RBv, such a port becomes a member
port of the RBv, and the RBridge becomes a member RBridge of the RBv.
Furthermore, for a member RBridge, it MUST move out of RBv and clear
the RBv's information from its self-originated LSPs when it loses the
last member port from this group, due to port down, configuration,
etc.
Use of the Appointed Forwarder framework specified in [RFC6325], this
specification allows to utilize a single framework for both shared
LAN and point-point edge connectivity. Additionally this allows to
Zhai, et al. Expires February 27, 2013 [Page 7]
Internet-Draft Pseudo-Nickname August 2012
o Detect and protect against mis-configuration at the edge, e.g., on
the device SW1 the two interfaces are not configured as LAG or
o Accept TRILL and native frames on the RBridge interface connecting
S1 in above Figure 1.
o Avoid loops in the event S1 and S2 were connected by a native
Ethernet Link.
3.1. VLAN-x Appointed Forwarder for member interfaces in RBv
If member RBridges in RBv cannot see each others' Hellos on their
member ports (e.g., in the LAG scenario), then each RBridge becomes
Designated RBridge (DRB) for that port and appoints itself as AF for
all VLANs as it does not see any TRILL hellos on that port. However,
it MAY acts as appointed forwarders only for parts of VLANs on that
port, if it knows explicitly the sets of service VLANs on that port
via other means. For example, administrator can statically
configured the sets of service VLANs on that port, or a lower
protocol (e.g., LAG protocol) informs TRILL the sets of service VLANs
on that port, etc.
However, if they can see each others' Hellos on the member ports in
RBv (e.g., in the shared link scenario), the TRILL Hello protocol in
[RFC6325] is used for DRB election and for VLAN-x AFs appointment on
those ports. Then the DRB appoints different member ports as AFs for
different sets of VLANs.
Among the member RBridges of RBv, only the VLAN-x forwarder is
responsible to ingress native traffic (both unicast and non-unicast
traffic) in this VLAN into TRILL campus, but non-forwarder member
RBridge is also permitted to egress unicast TRILL data traffic out of
TRILL campus. For the multi-destination TRILL data frames, only the
VLAN-x forwarder can egress their out of TRILL campus.
3.2. Announcing Pseudo-Nickname of RBv
Each member RBridge advertises the RBv's pseudo-nickname using the
nickname sub-TLV [rfc6326bis], along with its regular nickname or
nicknames, in its LSPs. When a member RBridge leaves from RBv due to
losing its last member ports in RBv, it MUST clear RBv's pseudo-
nickname from its update LSPs.
4. Distribution Trees for Member RBridges in RBv
In TRILL, RBridges use distribution trees to forward multi-
destination frames. When a native frame either to destinations whose
Zhai, et al. Expires February 27, 2013 [Page 8]
Internet-Draft Pseudo-Nickname August 2012
location is unknown or to multicast/broadcast groups is necessary to
be ingressed into TRILL campus, the ingress RBridge encapsulates it
into multi-destination TRILL data frame and forwards it along a
chosen distribution tree. In the TRILL header of the TRILL frame,
the ingress nickname identifies the ingress RBridge and the egress
nickname represents the root of the chosen tree. After receiving a
multi-destination TRILL data frame, the RBridge performs Reverse Path
Forwarding (RPF) check, along with other checks, on the multi-
destination frame to further control potentially looping traffic.
RPF specifies that a multi-destination TRILL data frame ingressed by
an RBridge and forwarded along a distribution tree can only be
received by an RBridge on an expected port. If not on that port,
that frame MUST be dropped by that RBridge.
However, member RBridges employ RBv's pseudo-nickname other than
their own nicknames as ingress nickname when they ingress native
frames received on member ports, regardless unicast or non-unicast
frames, into TRILL campus. Therefore, when these frames reach a
remote RBridge, they will be treated, by that RBridge, as frames
ingressed by the same RBridge, i.e., RBv. If they are multi-
destination frames and the same distribution tree is chosen by
different member RBridges to forward these frames, they will travel
along the tree and reach a remote RBridge on different ports. Then
the RPF check is violated, and some of the frames reaching the
RBridge on unexpected ports are dropped by the RBridge.
To fix the above issue, a scalable and resilient approach is proposed
in [CMT], where different member RBridges are assigned different
distribution trees for forwarding the multi-destination TRILL data
frames that using RBv's pseudo-nickname as ingress nickname in their
TRILL header. And a new TLV, named Affinity sub-TLV, is also
introduced for a member RBridge to announce its assigned distribution
tree for RBv in its self-originated LSPs. After receiving such LSPs,
remote RBridges can calculate their RPF check information for RBv on
those specified trees.
In this specification, the approach proposed in [CMT] is employed for
RBv to assign different distribution trees to different member
RBridges and the Affinity sub-TLV for member RBridges to announce
their assigned trees in LSPs.
When a member RBridge joins into or leaves from a virtual RBridge
group RBv due to its last member ports up/down or its configuration
changing, etc., the distribution trees assigned to different member
RBridges may change. That change and its influence on frame
processing are beyond the scope of this document.
Zhai, et al. Expires February 27, 2013 [Page 9]
Internet-Draft Pseudo-Nickname August 2012
5. Frame Processing
Although, there are five types of Layer 2 frames in [RFC6325], e.g.,
native frame, TRILL data frame, TRILL control frames, etc., pseudo-
nickname of RBv is only used for native frame and TRILL data frame in
this specification.
5.1. Data Frames
5.1.1. Native Frames Ingressing
When RB1 receives a native frame on one of its valid member ports of
RBv, it uses the pseudo-nickname of RBv, instead of its own nickname,
as ingress nickname, if it is the appointed forwarder for the VLAN of
the frame on the port. If the frame is not received on a member
port, RB1 MUST NOT use RBv's pseudo-nickname as ingress nickname when
doing TRILL-encapsulation on the frame. Otherwise, the reverse
traffic may be forwarded to another member RBridge that does not
connect to the link containing the destination, which may cause the
traffic disruption.
If the above native frame is ingressed by RB1 as a multi-destination
TRILL data frame, e.g., its destination is unknown to RB1 or it is
non-unicast frame, RB1 can only choose one of its assigned
distribution trees for RBv to distribute the TRILL-encapsulated frame
[CMT]. If not so, the multi-destination TRILL data frame will fail
RPF check on another RBridge and be dropped.
Furthermore, for such a frame, its source MAC address information ( {
VLAN, Outer.MacSA, port } ) is learned by default if its source
address is unicast. Then the learned information is shared with
other member RBridges of RBv (See Appendix A for more details for the
information sharing).
5.1.2. TRILL Data Frames Egressing
This section describes egress processing of the received TRILL data
frames on a member RBridge(RBn, say) in virtual RBridge group of RBv.
Section 5.1.2.1 describes unicast TRILL data frames egress
processing. Section 5.1.2.2 covers multi-destination TRILL data
frames egress processing.
5.1.2.1. Unicast TRILL Data Frames
When receiving a unicast TRILL data frame, RBn checks the egress
nickname in the TRILL header of the frame. If the egress nickname is
one of RBn's own nicknames, the frame is processed as Section 4.6.2.4
in [RFC6325].
Zhai, et al. Expires February 27, 2013 [Page 10]
Internet-Draft Pseudo-Nickname August 2012
If the egress nickname is RBv's pseudo-nickname and RBn is a valid
member RBridge of RBv, the Inner.MacSA and Inner.VLAN ID are, by
default, learned associated with the ingress nickname, unless that
nickname is unknown or reserved or the Inner.MacSA is not unicast.
If the learned {Inner.MacSA, Inner.VLAN ID, ingress nickname} triplet
is a new one or updates the locally stored one, this triplet is
shared with other member RBridges of RBv (See Appendix A for more
details for the triplet sharing).
Then the frame being forwarded is de-capsulated to native form. The
Inner.MacDA and Inner.VLAN ID are looked up in RBn's local forwarding
address cache, and one of the three following cases occurs:
o If the destination end station identified by the Inner.MacDA and
Inner.VLAN ID is on a local link to RBv, this frame is sent onto
the link containing the destination.
o else if RBn can reach the destination through another RBridge RBk,
it re-encapsulate the native frame into a unicast TRILL data frame
and sends it to RBk. RBn uses RBk's own nickname, instead of
RBv's pseudo-nickname as egress nickname for the re-encapsulation,
and remains the ingress nickname unchanged.
o Else, RBn does not know how to reach the destination; it sends the
native frame out of all its member ports of RBv on which it is
appointed forwarders for the Inner.VLAN.
5.1.2.2. Multi-Destination TRILL Data Frames
If the RBn is an appointed forwarder for the Inner.VLAN ID of the
frame, the Inner.MacSA and Inner.VLAN ID are, by default, learned as
associated with the ingress nickname unless that nickname is unknown
or reserved or the Inner.MacSA is not unicast. If the learned
{Inner.MacSA, Inner.VLAN ID, ingress nickname} triplet is a new one
or updates the locally stored one, this triplet is shared among the
members RBridges in virtual RBridge group RBv (See Appendix A for
more details for the triplet sharing).
Then a copy of the frame is de-capsulated into its native form.
Before the native frame is sent out of ports on which RBn is
appointed forwarder for the frame's VLAN, the following extra check
is performed:
o Assigned Distribution Trees Check (ADTC): If the flag for this
check (ADTC_flag) is not zero on such a port, the distribution
tree T along which the TRILL data frame arrives at RBn is checked.
Only if T is one of RBn's assigned distribution trees in RBv, the
native frame can be send out of this port. If not, the frame
Zhai, et al. Expires February 27, 2013 [Page 11]
Internet-Draft Pseudo-Nickname August 2012
cannot be sent out of this port.
The value of ADTC_flag on a RBridge's end-station-servicing port
depends on whether the port is a member port of RBv and RBn can not
receive Hellos from other member RBridges on that port or not.
If a port is a member port of RBv and RBn is the appointed forwarder
for all VLANs on that port, the ADTC_flag MUST be set 1. For all
other cases ADTC_flag MUST be set to zero.
6. Member Link Failure in RBv
In Figure 1, if the link SW1-RB1 fails, RB1 loses its only local link
to S1. When that failure is detected, the MAC entries through the
failed link to S1 are removed from RB1's forwarding table
immediately, and other MAC entries to S1 shared by other member
RBridges of RBv (See Appendix A for more details), e.g., RB2, are
installed into RB1's forwarding table when RB1 still has at least one
valid member port in RBv. Then when the TRILL-encapsulated traffic
to S1 is delivered to RB1, it can be re-encapsulated by RB1 and
forwarded, based on the available MAC entries, to another member
RBridge which has direct link to S1 and egresses the traffic to S1.
On the other hand, if RB1 has lost all its member ports of RBv, it
MUST updates its self-orignated LSPs to announce its giving up of
membership of RBv and no longer utilizes pseudo-nickname of RBv to
ingress/egress traffic into/out of TRILL campus until one of its
member ports of RBv becomes valid.
NOTE: Although on an edge RBridge different ports that connect to
different LAGs or LANs can join the same RBv, for simplicity, it is
RECOMMENDED that on an edge RBridge different ports connecting to
different LAGs or LANs join different RBvs in practical deployment,
each RBv per LAG or per LAN. Then for such an edge RBridge, when all
its member ports connecting to a LAG or LAN failed, it can move out
of this RBv and no longer uses the RBv's pseudo-nickname to ingress/
egress data traffic into/out of TRILL campus.
7. OAM Frames
Special attention must be paid when generate the OAM frames. When an
OAM frame is generated with ingress nickname of RBv, originator
RBridge's nickname MUST be included in the OAM message to ensure
response is returned to the originating member of RBv group.
Zhai, et al. Expires February 27, 2013 [Page 12]
Internet-Draft Pseudo-Nickname August 2012
8. Configuration Consistency
It is important that VLAN membership of member ports of end switch
SW1 is consistent across all of the member ports it is attaching to
member RBridges of RBv in the point-piont scenario. Any
inconsistencies in VLAN membership may result in packet loss or
having to through an extra hop RBridge before the packet reaches its
destination end station.
As an example consider, in Figure 1, on RB1 link SW1-RB1 has VLAN1
and VLAN2 configured. Consider only VLAN1 is configured on RB2 on
SW1-RB2 link. Both RB1 and RB2 use the same ingress nickname RBv for
all frames originating from S1. Hence, RBx will learn MAC address
from S1 on VLAN2 as originating from RBv. As a result, on the return
path RBx may deliver VLAN2 traffic to RB2. RB2, does not have VLAN2
configured on SW1-RB2 link and hence may drop the frame or has to
forward the frame to RB1 to egress it to S1 if RB2 knows RB1 can
reach S1 in VLAN2.
9. IANA Considerations
TBD.
10. Security Considerations
TBD.
11. Acknowledgements
We would like to thank Mingjiang Chen for his contributions to this
document. Additionally, we would like to thank Erik Nordmark, Les
Ginsberg, Ayan Banerjee, Dinesh Dutt, Anoop Ghanwani, Janardhanan
Pathang, and Jon Hudson for their good questions and comments.
12. Normative References
[CMT] Senevirathne, T., Pathangi, J., and J. Hudson,
"Coordinated Multicast Trees (CMT)for TRILL",
draft-ietf-trill-cmt-00.txt Work in Progress, April 2012.
[RFC1195] Callon, R., "Use of OSI IS-IS for routing in TCP/IP and
dual environments", RFC 1195, December 1990.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Zhai, et al. Expires February 27, 2013 [Page 13]
Internet-Draft Pseudo-Nickname August 2012
Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC6325] Perlman, R., Eastlake, D., Dutt, D., Gai, S., and A.
Ghanwani, "Routing Bridges (RBridges): Base Protocol
Specification", RFC 6325, July 2011.
[rfc6326bis]
Eastlake 3rd, D., Banerjee, A., Ghanwani, A., and R.
Perlman, "Transparent Interconnection of Lots of Links
(TRILL) Use of IS-IS",
draft-eastlake-isis-rfc6326bis-07.txt Work in Progress,
March 2012.
Appendix A. Reasons for MAC Sharing among Member RBriges
With the introduction of virtual RBridge, MAC flip-flopping problem
in LAN or LAG is resolved. However, in order to forward traffic
effectively, member RBridges should share some of their learned MAC
addresses with each other. For example, see Figure 2 shown below.
...........................................
: TRILL Network :
^ : +-----+ /\-/\-/\-/\-/\ :
+-----| RB1 |-----/ \ :
/ : +-----+ / \ :
/ : | / Transit \ +-----+ :
S1 o RBv : | < RBridges >---| RBx |---o Sx
\ : | \ Campus / +-----+ :
\ : +-----+ \ / :
+-----| RB2 |-----\ / :
V : +-----+ \/\-/\-/\-/\-/ :
: :
...........................................
Figure 2 RBv in
LAG scenario
In VLAN-x, native frames from S1 to Sx will enter TRILL campus
through one member RBridge of the RBv, such as RB1 in Figure 2, so
RB1 learns the location of S1 in VLAN-x; but with regard to reverse
traffic, RBx may deliver it to RB2 if it thinks the shortest path to
RBv is through RB2. Then, if RB2 knows the location of S1 and the
link RB2-S1 is good, it egresses the traffic directly to S1.
However, if the link fails and RB2 has not learn the location of S1
in VLAN-x, RB2 cannot transmit the traffic properly to S1.
Zhai, et al. Expires February 27, 2013 [Page 14]
Internet-Draft Pseudo-Nickname August 2012
Thus, the MAC addresses of attached end stations on one member
RBridge SHOULD be shared with the rest member RBridges in an RBv.
With these informations shared, when RB2 receives reverse frames, it
can determine how to forward them to S1, for example, forward them to
RB1 if the link RB2-S1 fails.
On the other hand, RBx always delivers the reverse traffic to RB2 if
it thinks the shortest path to RBv is through RB2. Then RB2 egresses
the traffic and learns the location of Sx in the case its link to S1
is good. RB1 will not know where Sx is if neighbor it has other ways
to get the location of Sx nor RB2 shares this information with it.
As a result, it has to always treat the traffic from S1 to Sx as
unknown destination traffic and multicast it in TRILL. Always multi-
casting such traffic adds additional forwarding burden on TRILL
network.
Therefore, in addition to local attached end station MAC addresses,
the learned remote MAC addresses should also be shared among all
other member RBridges in an RBv. With such information sharing, RB1
can treat the traffic to Sx as known destination traffic and unicast
it to RBx.
Although we can extend ESADI (End Stations Address Distribution
Information) protocol or LAG protocol, etc., for such MAC sharing,
ways for the sharing are beyond the scope of this document.
Authors' Addresses
Hongjun Zhai
ZTE
68 Zijinghua Road, Yuhuatai District
Nanjing, Jiangsu 210012
China
Phone: +86 25 52877345
Email: zhai.hongjun@zte.com.cn
Tissa Senevirathne
Cisco Systems
375 East Tasman Drive
San Jose, CA 95134
USA
Phone: +1-408-853-2291
Email: tsenevir@cisco.com
Zhai, et al. Expires February 27, 2013 [Page 15]
Internet-Draft Pseudo-Nickname August 2012
Radia Perlman
Intel Labs
2200 Mission College Blvd
Santa Clara, CA 95054-1549
USA
Phone: +1-408-765-8080
Email: Radia@alum.mit.edu
Donald Eastlake 3rd
Huawei
155 Beaver Street
Milford, MA 01757
USA
Phone: +1-508-333-2270
Email: d3e3e3@gmail.com
Mingui Zhang
Huawei
Huawei Building, No.156 Beiqing Rd.
Beijing, Beijing 100095
China
Email: zhangmingui@huawei.com
Fangwei Hu
ZTE
889 Bibo Road, Pudong District
Shanghai, Shanghai 201203
China
Phone: +86 21 68896273
Email: hu.fangwei@zte.com.cn
Zhai, et al. Expires February 27, 2013 [Page 16]