draft-vasseur-mpls-backup-computation-00.txt June 2002
Jean-Philippe Vasseur
Anna Charny
Francois Le Faucheur
Cisco Systems, Inc.
Javier Achirica
Telefonica Data Espagna
IETF Internet Draft
Expires: December, 2002
June, 2002
draft-vasseur-mpls-backup-computation-00.txt
MPLS Traffic Engineering Fast reroute: backup tunnel path computation
for bandwidth protection
Status of this Memo
This document is an Internet-Draft and is in full conformance with all
provisions of Section 10 of RFC2026. 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.
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
Vasseur, Charny, Le Faucheur and Achirica 1
draft-vasseur-mpls-backup-computation-00.txt June 2002
Content
1. Terminology ------------------------------------------------------ 4
2. Introduction ----------------------------------------------------- 5
3. Background and Motivation ---------------------------------------- 5
4. Various backup tunnel path computation models -------------------- 6
5. Limitations of the independent CSPF-based computation model ------ 6
5.1 Bandwidth sharing between backup tunnels ------------------------ 7
5.2 Potential inability to find a placement of a set of backup tunnels
satisfying constraints ---------------------------------------------- 8
6. Facility based computation model ----------------------------------8
6.1 Centralized backup path computation scenario -------------------- 8
6.1.1 Server responsible for both the primary and backup tunnels path
computation --------------------------------------------------------- 9
6.1.2 Server responsible for backup tunnels path computation only (not
primary TE LSPs) --------------------------------------------------- 11
6.2 Distributed backup tunnel path computation scenario ------------ 12
6.2.1 Node Protection ---------------------------------------------- 13
6.2.2 Link protection ---------------------------------------------- 14
6.2.3 SRLG protection ---------------------------------------------- 15
6.2.4 Protection order --------------------------------------------- 15
6.3 Signaled parameters -------------------------------------------- 15
6.3.1 Element to protect ------------------------------------------- 15
6.3.2 Bandwidth to protect ----------------------------------------- 15
6.3.3 Affinities --------------------------------------------------- 16
6.3.4 Maximum number of backup tunnels ----------------------------- 16
6.3.5 Minimum bandwidth on any element of a set of backup tunnels -- 16
6.3.6 Class Type (CT) to protect ----------------------------------- 16
6.3.7 Set of already in place backup tunnels ----------------------- 16
7. Validity of the independent failure assumption ------------------ 16
8. Operations with DS-TE and multiple Class-Types ------------------ 18
8.1 Single backup pool --------------------------------------------- 18
8.2 Multiple backup pool ------------------------------------------- 21
9. Interaction with Scheduling ------------------------------------- 23
10. Routing and signaling extensions ------------------------------- 25
10.1 Routing (IGP-TE) extensions ----------------------------------- 25
10.2 Signaling (RSVP-TE) extensions -------------------------------- 26
10.2.1 PCC -> PCS signaling : specification of a set of constraints 26
10.2.2 PCS->PCC signaling: sending of the computed set of backup
tunnels ------------------------------------------------------------ 29
11 Backup Tunnel - - Make before break ------------------------------- 30
12 Stateless versus statefull PCS ---------------------------------- 30
13 Packing algorithm ----------------------------------------------- 30
14 Interoperability in a mixed environment ------------------------- 31
15 Security consideration ------------------------------------------ 31
Vasseur, Charny, Le Faucheur and Achirica 2
draft-vasseur-mpls-backup-computation-00.txt June 2002
References
Appendix A: Limitations/inefficiency of the independent CSPF-based
computation model
Appendix B: Bandwidth to protect
Appendix C: Backup tunnel path computation triggering and path changes
Appendix D ''Push'' versus ''Pull'' mode
Vasseur, Charny, Le Faucheur and Achirica 3
draft-vasseur-mpls-backup-computation-00.txt June 2002
Abstract
This draft proposes an efficient model called ''Facility based
computation model'' for computing bypass tunnels paths in the context of
the MPLS TE Fast Reroute, while allowing bandwidth sharing between
backup tunnel protecting independent resources. Both a centralized and
a distributed path computation scenarios are described. The required
signaling extensions are also addressed in the draft.
1. Terminology
LSR - Label Switch Router
LSP - An MPLS Label Switched Path
PCS - Path Computation Server (may be any kind of LSR (ABR, ...)
or a centralized path computation server
PCC - Path Computation Client (any head-end LSR) requesting a path
computation of the Path Computation Server.
Local Repair - Techniques used to repair LSP tunnels quickly
when a node or link along the LSPs path fails.
Protected LSP - An LSP is said to be protected at a given hop if
it has one or multiple associated backup tunnels
originating at that hop.
Detour LSP - An MPLS LSP used to re-route traffic around a failure
in one-to-one backup.
Bypass Tunnel - An LSP that is used to protect a set of LSPs
passing over a common facility.
Backup Tunnel - The LSP that is used to backup up one of the many
LSPs in many-to-one backup.
PLR - Point of Local Repair. The head-end of a backup tunnel or
a detour LSP.
MP - Merge Point. The LSR where detour or backup tunnels meet
the protected LSP. In case of one-to-one backup, this is where
multiple detours converge. A MP may also be a PLR.
NHOP Bypass Tunnel - Next-Hop Bypass Tunnel. A backup tunnel
which bypasses a single link of the protected LSP.
Vasseur, Charny, Le Faucheur and Achirica 4
draft-vasseur-mpls-backup-computation-00.txt June 2002
NNHOP Bypass Tunnel - Next-Next-Hop Bypass Tunnel. A backup
tunnel which bypasses a single node of the protected LSP.
Reroutable LSP - Any LSP for with the "Local protection desired"
bit is set in the Flag field of the
SESSION_ATTRIBUTE object of its Path messages.
CSPF - Constraint-based Shortest Path First.
2. Introduction
The focus of this document is ''Bandwidth protection'' in the context of
local repair capability of MPLS Fast Reroute. We concentrate on the
issues related to the computation of local backup (also called bypass)
tunnels satisfying capacity constraints in the context of facility
backup. We do not propose another method for MPLS traffic Engineering
Fast Reroute. This draft makes the assumption that the fast reroute
technique named Facility backup and described in [FAST-REROUTE]is used
to provide fast recovery in case of link/node failure.
The exact algorithms for placement of the backup tunnels with bandwidth
guarantees are outside the scope of this draft. Rather, we concentrate
on the mechanisms enabling the backup tunnel path computation to be
performed by a server which holds sufficient information in order to
achieve efficient sharing of bandwidth between backup tunnels
protecting independent failures. The mechanisms are described in the
context of both a centralized and a distributed computation. We
specifically address the signaling involved for such computation
between the PLR and the server (also called PCC-PCS signaling).
3. Background and Motivation
As defined in [FAST-REROUTE], a TE LSP can explicitly request:
- local protection (''Local protection desired'' bit set in
the SESSION-ATTRIBUTE object carried in the Path message),
- bandwidth protection (''Bandwidth protection desired'' bit
set in the SESSION-ATTRIBUTE object carried in the Path
message),
or
- local protection and bandwidth protection carrying a FAST-
REROUTE object in the Path message,
- other parameters,
Bandwidth protection will typically be requested for TE LSPs carrying
very sensitive traffic (Voice trunking, ...).
When a link or a node failure occurs, the PLR (Point of Local Repair)
fast reroutes the protected LSPs onto their backup tunnel. The PLR will
also send a Path Error notifying the head-end LSRs that the protected
Vasseur, Charny, Le Faucheur and Achirica 5
draft-vasseur-mpls-backup-computation-00.txt June 2002
LSPs have been locally repaired so that head-ends should trigger a re-
optimization, and potentially reroute the TE LSP in a non disruption
fashion (make before break) following a more optimal path, provided
such a path exists.
The bandwidth of the backup tunnels that the protected LSPs will be
rerouted onto will dictate the level of bandwidth protection and so the
QOS during failure until the TE LSPs are being re-optimized (if such a
re-optimization can be performed, depending on the available network
resources).
Various constraints can be taken into account for the backup tunnels:
(1) must be diversely routed from the protected element
(link/node/SRLG diverse),
(2) must be setup in such a way that they get enough
bandwidth so that the protected LSPs requesting bandwidth
protection should receive the same level of QOS when
rerouted. Note that the notion of bandwidth protection is on
a per LSP basis.
(1) must always be satisfied and makes FRR an efficient protection
mechanism to reroute protected TE LSP in 10s of milliseconds in case of
link or node failure.
(2) allows FRR to provide an equivalent level of QOS during failure to
the TE LSPs that have requested bandwidth protection.
4. Various backup tunnel path computation models
Various backup tunnel path computation models have been proposed:
independent CSPF-based computation, [KINI], [BP-PLACEMENT], ... A new
model, named ''facility based computation model'' is proposed in this
draft.
5. Limitations of the independent CSPF-based computation model
The simplest mechanism to get bandwidth protection available today is
to rely on existing IGP advertisement and for the head-end of the
backup tunnel:
- to determine the bandwidth requirements of the desired backup
tunnel(s),
- to determine the path in the network where the appropriate
amount of bandwidth is available using standard CSPF-based
computation,
- to signal the bandwidth requirements of the individual backup
tunnels explicitly.
While this approach is quite attractive for its simplicity, it presents
a substantial set of challenges:
- bandwidth sharing between backup tunnels,
Vasseur, Charny, Le Faucheur and Achirica 6
draft-vasseur-mpls-backup-computation-00.txt June 2002
- potential inability to find a placement of the backup tunnels
satisfying the bandwidth constraints.
5.1. Bandwidth sharing between backup tunnels
Since local repair is expected to be used for only a short period of
time after failure, followed by re-optimization of the affected primary
LSPs, it is reasonable to expect that the probability of multiple
failures in this short period of time is small. As a result, being able
to share bandwidth on the link by backup tunnels protecting different
failures typically results in large savings in the bandwidth required
for protection. This is what we refer many times in this document as
''efficient bandwidth sharing'' or as achieving ''bandwidth sharing''. Note
also that the single failure assumption needed for such bandwidth
sharing is a pre-requisite to any protection approach which uses pre-
computed backup tunnels - - clearly even two completely link and node
disjoint paths can both fail if more than one failure can occur. It is
worth underlining that the single failure of a SRLG may result in the
actual failure of multiple links.
Once head-ends will received the Path Error (''Tunnel locally repaired'')
reoptimization should be triggered followed by an LSP reroute making
use of the ''Make Before Break'' technique to avoid traffic disruption,
if such a more optimal path obeying the constraints within the new
network topology can be found. If such a path cannot be found, the TE
LSP will not be reoptimized and will still be fast rerouted by the
immediately upstream PLR attached to the failed element. If a second
failure occurs before the TE LSP should be reoptimized, this results in
a multiple independent failures situation where bandwidth protection
cannot be ensured. Now this case should not be considered as a
showstopper as:
- in networks where bandwidth is a reasonably available
resource, this situation is unlikely to happen as the TE LSP
reoptimization will succeed,
- in networks where bandwidth is a very scarce resource,
bandwidth protection without backup bandwidth sharing is out of
question anyway.
As a result, bandwidth sharing among backup tunnels protecting
independent failures is highly desirable.
However, achieving such sharing using explicit bandwidth reservations
for the backup tunnels requires extensive signaling and routing
extensions:
- routing extensions propagating the set of backup LSPs as
well as their bandwidth and the element(s) they protect [BP-
PLACEMENT]. In [lakshman]it was proposed to substantially
reduce the amount of state that needs to be propagated in the
routing updates at the price of sacrificing the amount of
achievable sharing.
- signaling extensions to perform specific call admission
Vasseur, Charny, Le Faucheur and Achirica 7
draft-vasseur-mpls-backup-computation-00.txt June 2002
control for the backup LSPs.
5.2. Potential inability to find a placement of a set of backup
tunnels satisfying constraints
Another well-known issue with independent CSPF-based computation with
explicitly signaled bandwidth requirements is its potential inability
to find a placement of the backup tunnels satisfying the bandwidth
constraints, even if such a placement exists. This issue is not
specific to the placement of the backup tunnels - - rather it is due to
the sub-optimality of a greedy on-demand nature of the CSPF approach.
See appendix A for a detailed example.
6. Facility based computation model
In this draft we propose another model for the backup tunnel path
computation referred as the ''Facility based computation model''.
The facility based computation model can be implemented in two
different ways: centralized (with two sub cases) or distributed. In all
of these scenarios the facility based computation enables efficient
sharing of bandwidth among backup tunnels protecting independent
failures. In addition, all of these scenarios also allow overcoming
some of the limitations of the greedy independent CSPF-based placement
of the backup tunnels increasing the chances of finding a backup
tunnels placement satisfying the constraints if such a solution exists.
While some of these approaches can benefit from an IGP-TE extension
maintaining and advertising an addition bandwidth pool, all of these
approaches can be usefully deployed in a limited fashion in the
existing networks without any additional routing extensions at all. As
shown bellow, the required signaling extensions could be based on
[PATH-COMP] with a few additional objects (described in section 11.).
6.1. Centralized backup path computation scenario
In the centralized scenario, the backup tunnel path computation is
being performed on a central PCS (which can be a workstation or another
LSR). The PCS will be responsible for the computation of the backup
tunnels for some or all the LSRs in the network. Typically, there could
be one PCS per area in the context of a multi-area network. The PCS(s)
address may be manually configured on every LSR or automatically
discovered using IGP extensions (see [ISIS-PCSD] and [OSPF-PCSD]).
To compute the backup tunnels protecting a given element, the server
needs to know:
- the network topology,
- the desired amount of primary traffic that needs to be
protected (this could be either the actual bandwidth reserved
Vasseur, Charny, Le Faucheur and Achirica 8
draft-vasseur-mpls-backup-computation-00.txt June 2002
by primary LSPs or the bandwidth pool that could be reserved by
the primary LSPs - - see Appendix A for a detailed discussion),
- the amount of bandwidth available for the placement of the
backup tunnels (also referred to as backup bandwidth).
The first two items are available directly from IGP_TE database. The
third one depends on the exact model and is discussed separately in
each case.
However, whether or not this information is sufficient, depends on
whether the server is also responsible for the computation of primary
tunnels or not. This is discussed below.
6.1.1. Server responsible for both the primary and backup tunnels path
computation
In this scenario, the PCS can easily take advantage of knowing all the
primary tunnels to define tunnel bandwidth requirements based on actual
primary LSPs.
There is substantial flexibility in choosing what bandwidth can be used
for the backup tunnel placement. One approach might be to use for the
backup tunnels protecting traffic of a particular pool the same
bandwidth pool as the corresponding primary LSPs.
At some point the user will have to specify the policy to the server.
For example, protect traffic of a pool X with a backup tunnel in the
same pool but also the proportion of pool X that can be used for backup
and primary. For pool X, the user could specify: ''up to y% of pool X
can be used for backup''.
Since in this scenario the server is responsible for the placement of
both the primary traffic and the backup tunnels, at any given time in
the computation of the backup tunnels it has complete information about
the topology and the current placement of all backup and primary
tunnels. Therefore, the server can compute the backup tunnels
protecting one element at a time, and when placing its backup tunnels
simply ignore the bandwidth of any backup tunnels already placed if
those protect a different element, thus ensuring implicitly the desired
bandwidth sharing. In this case, there is no need to specify a notion
of backup bandwidth pool.
Signaling Backup tunnels with zero bandwidth
Having computed the backup tunnels, the server needs to inform the head
ends of the backup tunnels about the placement of the backup tunnels,
their bandwidth requirements, and the elements they protect.
Depending on whether the server is an LSR or not, this could be done
either via a network management interface, or signaled using RSVP
extensions similar to those described in draft [PATH-COMP] (with a new
Vasseur, Charny, Le Faucheur and Achirica 9
draft-vasseur-mpls-backup-computation-00.txt June 2002
RSVP object needed to achieve this communication described in section
11).
Once an LSR has received the information about the backup tunnels for
one or more elements it is the head-end for, it needs to establish
those tunnels along the specified paths. At first glance, given the
need to ensure bandwidth protection, it seems natural to signal the
bandwidth requirements of the backup tunnel explicitly. However, as
discussed in [BP-PLACEMENT], such approach requires that the local
admission control is changed to be aware of the bandwidth sharing, and
additional signaling extensions need to be implemented to enable an LSR
to tell a primary LSP from a backup lSP so that admission control can
be performed differently in the two cases.
However, since the placement of both the primary and the backup tunnels
in this case is done by the server which maintains the bandwidth
requirements of all these primary and backup lSPs, it is sufficient to
signal zero-bandwidth tunnels, thus avoiding the need for any
additional signaling extensions or changes to admission control. Even
though the bandwidth will not be explicitly signaled, the required
bandwidth will nevertheless be available along the path upon failure by
virtue of the computation of this placement by the server which is
fully aware of the global topology and places of all lSPs in such a way
that their bandwidth requirements are satisfied.
Note also that although the bandwidth requirements are not explicitly
signaled, the head-end may store this information locally, since it may
be needed in determination of which primary LSPs to assign to which
backup tunnels in the case where more than one backup tunnel exists
(see section 14).
PCC-PCS signaling
If the path computation server uses a network management interface to
obtain the topology information and communicate the paths of the
computed backup tunnels to their head ends, this approach requires no
signaling extensions at all. However, in the case the path computation
server is an LSR itself, additional signaling mechanisms are required
to communicate to the server a request to compute backup tunnels for a
particular element, and for the server to communicate the backup
tunnels to their head-ends. These extensions, described in detail in
sections 11 are built on those proposed in [PATH-COMP]. Of course, the
same extensions could be also used even if the PCS is a network
management station.
Note that the benefit of having an LSR be the PCS as opposed to an off-
line tool is the LSR's real-time visibility to any topology changes in
the network. In particular, the LSR-based approach can be expected to
recompute the backup tunnels affected by a failure much faster than a
network-management based solution, thus making a single failure
assumption more reliable. In addition, as will be discussed later in
section 6.2, the ability of an LSR to compute backup tunnels for other
Vasseur, Charny, Le Faucheur and Achirica 10
draft-vasseur-mpls-backup-computation-00.txt June 2002
elements is especially useful in the context of a more distributed
backup tunnel computation.
6.1.2. Server responsible for backup tunnels path computation only
(not primary TE LSPs)
The main benefit of the previous scenario (PCS computing both the
primary and backup LSPs) was residing in the fact that the PCS could
make use, for the backup tunnels, of any available bandwidth not
reserved for primary TE LSPs. As a consequence, this was not requiring
a separate backup pool. On the other hand, if the PCS is just
responsible for the backup tunnels paths (i.e the primary tunnels are
established on-line or by any other mechanism external to the backup
path computation server), and the backup tunnels are signaled with
zero bandwidth to enable efficient bandwidth sharing, the backup
tunnels cannot draw bandwidth from the same pool as the primary traffic
they protect. This is because primary TE LSPs could make use of all
bandwidth including the bandwidth the PCS used for backup tunnel path
computation, which would result in bandwidth protection violation.
Achieving efficient bandwidth sharing in this case requires the
definition of a separate backup pool that could only be used for backup
tunnels.
Note that the notion of backup bandwidth pool is similar to that
described in [BP-PLACEMENT].
In this case, the bandwidth requirements for backup tunnels is based on
the full bandwidth pool available for primary, independently of how
much of this bandwidth pools is currently used by the primary LSPs;
The backup bandwidth pool approach can be used in two ways:
- being advertised in IGP
- without being advertised in IGP
Backup Pool advertised in IGP
In this approach, an additional bandwidth pool is established, and is
flooded in the routing updates. See section 11 for more details
If the backup path computation server uses the value of the backup
bandwidth pool for its computation, no bandwidth overbooking will ever
occur, since the primary tunnels now use the bandwidth from a different
pool. The additional state that needs to be flooded in routing updates
to implement the backup bandwidth pool does not impact the IGP
scalability as the bandwidth protection pool being announced by IGP-TE
is a static value i.e does not dynamically change as backup TE LSP are
set up, which preserves IGP scalability. As the bandwidth protection
pool is being defined on a per link basis, this allows for different
policies depending on the link characteristics.
Vasseur, Charny, Le Faucheur and Achirica 11
draft-vasseur-mpls-backup-computation-00.txt June 2002
Backup Pool not being advertised in IGP
It should be pointed out that in the context of the facility based
computation model, such routing extension is desirable but not
necessary to deploy this approach in the existing network in a limited,
but useful fashion.
Since the computation of the backup tunnels in this approach is
performed by a centralized server, the server can use the notion of the
backup bandwidth pool implicitly. Just as in the case of a server
computing the placement of both primary and backup LSPs, such policy
may be simply configured on the server for every link. The policy must
ensure that the backup pool never overlaps with the pool requiring
bandwidth protection.
Thus, substantial benefits may be achieved in this approach without
actually deploying any additional IGP-TE extensions at all. The only
drawback is that the policy will have to be the same for the whole
network or may be specified on a per link basis which requires some
extra configuration work on the PCS. Just as in the previous approach
(section 6.1.1) signaling extensions can be used between a PCC and a
PCS whether the PCS is an LSR or a network management station.
6.2. Distributed backup tunnel path computation scenario
While there are several clear advantages of a centralized (off-line)
model, there are also well-known disadvantages of it (such as the
single point of failure, the necessity to provide reliable
communication channels to the server, etc.) While most of these issues
can be addressed by the proper architectural design of the network, a
dynamic distributed solution is clearly desirable.
This section presents the use of the ''facility-based computation''
solution in a distributed model.
6.2.1. Node Protection
Consider first the problem of node protection. The key idea is to shift
the computation of the backup tunnels from the head-ends of those
backup tunnels to the node that is being protected. Essentially, each
node protects itself by computing the placement of all the backup
tunnels that are required to protect the bandwidth of traffic
traversing this node in the case of its failure. Once the backup
tunnels are computed, they need to be communicated to their head-ends
(in this case the neighbors of the protected node) for installation.
The backup tunnel head-ends play the role of PLR. Essentially, each
Vasseur, Charny, Le Faucheur and Achirica 12
draft-vasseur-mpls-backup-computation-00.txt June 2002
node becomes a PCS for all of its neighbors, computing all NNHOP backup
tunnels between each pair of its neighbors which are necessary for its
own protection. The fact that the backup tunnels to protect a node X
are being computed by a single PCS (node X) is essential and much more
efficient than the non-coordinated independent CSPF-based computation.
The key pieces that make this model work are those already described in
the context of the centralized server:
1) making use of explicitly defined backup bandwidth pool
2) taking advantage of a single failure assumption to do bandwidth
sharing
3) installing backup tunnels with zero bandwidth.
These three things together allow the computation of the placement of
backup tunnels for a given node to be completely independent of the
placement of backup tunnels for any other node. Essentially, each node
has the entire backup bandwidth pool available for itself. The problem
it needs to solve is how to place a set of NNHOP backup tunnels (one or
more for each pair of its direct neighbors) in a network with available
capacity on each link equal to the backup bandwidth pool. This problem
can be solved by any algorithm for finding a feasible placement of a
set of flows with given demands in a network with links of given
capacity.
While the details of such algorithm are beyond the scope of this draft,
it is clear that since the node now has control over all backup tunnels
protecting itself, it is more likely that it can find such a placement,
and potentially find a more optimal placement, than is possible if the
head-ends of the backup tunnels compute the placement of these tunnels
independently of each other.
Just as in the case of a centralized server, installing the backup
tunnels with zero bandwidth ensures that no changes to admission
control are necessary to allow sharing of the backup pool by backup
tunnels protecting different nodes, thus enabling bandwidth sharing
between independent failures. Yet, by virtue of the computation, the
backup tunnels protecting a given node will also have enough bandwidth
in the case of that node's failure.
Also, the backup pools can be implicitly derived from the routing
information already available. This could be done by configuring max
global reservable pool to being less than the link speed by the desired
value of the backup pool. Every node computing its backup tunnels then
can by default use link speed minus the max global reservable pool as
the value of the backup pool to use in its computation of the backup
tunnels placement. However, there is substantial benefit in defining
the backup pool explicitly and advertise its value as part of the
topology in the routing updates. This clearly requires an IGP-TE
extension as described in section 11 . The benefit of doing so is that
it provides much more flexibility in the design of the network. But
again IGP-TE extensions is a benefit not a requirement for this
Vasseur, Charny, Le Faucheur and Achirica 13
draft-vasseur-mpls-backup-computation-00.txt June 2002
solution to work. Signaling extensions required for communication
between the node serving as path computation server and the head-ends
of the backup tunnels are the same as for an off-line server and are
defined in sections 11.
6.2.2. Link Protection
In order to protect a link with FRR in both directions, two backup
tunnels protecting each direction of this link are installed by the
corresponding head-end of that link. To make sure that traffic
requesting bandwidth protection traversing this link is protected in
case of a link failure (if both directions fail simultaneously), it is
necessary to account for the interaction of the backup tunnels
protecting different directions of this link. That is, one needs to
make sure that if a backup tunnel T1 protecting bandwidth B1 on a
directed link A->B and the tunnel T2 protecting bandwidth B2 on a
directed link B->A traverse the same directed link L, then link L has
spare capacity of at least B1+B2.
If the two ends of the link compute their backup tunnels independently,
the way to ensure this condition would be to explicitly signal the
bandwidth of the backup tunnels. However, as discussed earlier, this
approach makes the sharing of bandwidth between the backup tunnels
protecting different elements impractical and would require IGP and
admission control extensions. To achieve this goal in a distributed
setting we propose that one of the two end-nodes of the link takes
responsibility for computing the backup tunnels for both directions
using the backup pools explicitly or implicitly defined. We propose
that by default the node with the smaller IGP id serves as the server
(PCS) for the other end of the link. Therefore, by default a node with
id X serves as a PCS for NNHOP backup tunnels protecting itself and
NHOP backup tunnels protecting any adjacent bi-directional link for
which the other end has an IGP id larger than X.
6.2.3. SRLG protection
This version of the draft does not support a case where a link is part
of more than one SRLG.
We propose to use exactly the same approach as for the bi-directional
link. That is, if an SRLG consists of a set of bi-directional links,
the node with the smallest IGP id of all the endpoints of these links
serves by default as a path computation server.
6.2.4. Protection order
Vasseur, Charny, Le Faucheur and Achirica 14
draft-vasseur-mpls-backup-computation-00.txt June 2002
If an LSR serves as a PCS for itself, some of its adjacent links and
SRLGs to which this link belongs, for any link that is part of SRLG it
needs to compute the backup paths for the entire SRLG and use it for
the backup of all links in this SRLG it is responsible for. The node
can compute the NNHOP tunnels it is responsible for in any order with
respect to the SRLG and link protection.
6.3. Signaled parameters
The PCC (an LSR) will send a backup tunnel path computation request to
the PCS using the RSVP TE extensions defined in [PATH-COMP] and the
newly BACKUP-TUNNEL object defined in this draft.
The PCC's request will be characterized by the specification of several
parameters that are discussed bellow.
6.3.1. Element to protect
The PCC specifies the element to protect: Link, Node or SRLG.
Typically, a link protection request will result in a set of NHOP
backup tunnels as a node protection request will result in a set of
NNHOP backup tunnels.
6.3.2. Bandwidth to protect
There are two different approaches for the bandwidth to protect
constraint:
- the backup tunnel bandwidth may be based on the amount of reservable
bandwidth on a particular network resource,
- the backup tunnel bandwidth may be based on the sum of bandwidths
actually reserved by established TE LSPs on a particular resource.
Each approach is having pros and cons that are being extensively
discussed in Appendix B.
6.3.3. Affinities
Affinities constraint may be also specified by the requesting node.
Affinities for the backup tunnel may be configured on the PLR by the
network administrator or derived from the FAST-REROUTE object of the
protected TE LSP, if used. In this latter case, this would require some
rules to derive the affinities of the backup tunnel from the affinities
of the protected TE LSPs making use of this backup tunnel.
6.3.4. Maximum number of backup tunnels
Vasseur, Charny, Le Faucheur and Achirica 15
draft-vasseur-mpls-backup-computation-00.txt June 2002
It may happen that no single backup tunnel can fulfill the constraints
requirements. In such a situation, a set of backup tunnels could be
computed such that the sum of the bandwidths of every element in the
set is at least equal to the required bandwidth. It may be desirable to
bound the number of elements in this set.
6.3.5. Minimum bandwidth on any element of a set of backup tunnels
When a solution can be found with a set of backup tunnels it may also
be desirable to provide some constraint on the minimal bandwidth value
for any backup tunnel in the set. As an example, if a 100M backup
tunnel is required, a set of 1000 tunnels each having 100K is likely to
be unacceptable. Also, it is worth reminding that a single protected TE
LSP will make use of a single backup tunnel at a given time.
6.3.6. Class Type to protect
Specifies the Class-Type(s) to protect. See section 8 on operations
with DS-TE.
6.3.7. Set of already in place backup tunnels
In certain circumstances, it may also be useful for the PCC to provide
to the PCS the set of already in place backup tunnels with their
corresponding constraints for the PCS to try to minimize the
incremental changes, especially when the PCS can handle the ''minimal
perturbation problem''. This will be further discussed in section 11.
7. Validity of the Independent failure assumption
The facility based computation model is heavily dependent on the
independent failure assumption. That is, it is assumed that the
probability of multiple independent element failures in the interval of
time required for the network to re-optimize primary tunnels affected
by a given failure and re-compute the backup tunnels for other elements
is low.
In a distributed model both of these tasks are likely to be
accomplished within a very short time so the assumption typically can
be justified. The loss of bandwidth protection in the rare cases that
the assumption is violated is offset by the benefit of sharing the
bandwidth between backup tunnels protecting different elements.
However, not all elements are independent. One example of elements that
are not independent is a set of links in the same SRLG. Therefore, as
discussed above, SRLG is treated as a single element and is protected
as a single entity.
Vasseur, Charny, Le Faucheur and Achirica 16
draft-vasseur-mpls-backup-computation-00.txt June 2002
Another example of failures that are not independent is a failure of a
node and links adjacent to it. It is possible (and is frequently the
case) that a failure of a node results also in the failure of the
link(s). However, in the approach described in the draft link and node
protection is done independently. This is necessary to ensure that
NNHOP tunnels for a node can be computed completely independently of
the NHOP tunnels for adjacent links, thus enabling the distributed
computation. The justification for this is that when a node fails,
traffic that does not terminate at this node is protected because it is
rerouted over the NNHOP tunnels, and traffic that does terminate at the
failed node does not need to be protected against the failure of
adjacent links since it is dropped anyway.
Thus, the underlying assumption is that if a node fails, the NHOP
tunnels protecting the link are not used, while if a link fails but the
router does not, the NHOP tunnels are used. So they can in fact be
computed independently. However, this reasoning only works if it is in
fact possible to identify the failure correctly and use the appropriate
set of tunnels depending on the failure.
There are several cases to be considered:
- a downstream router fails but the link does not
- the link fails but the downstream router does not
- the link fails because the downstream router failed
The first case is typically identifiable by means of RSVP hello or some
fast IGP hellos mechanism. However, using the currently deployed
mechanisms a node adjacent to the failed link cannot tell within the
time appropriate for Fast Reroute whether the node on the other side of
that link is operational or not. Hence, to protect both traffic that
terminates at the failed node in case the failure was a link failure,
and at the same time to protect traffic transit through the failed node
in case it was a node failure, the LSR adjacent to the failed link is
forced to use both the NHOP and the NNHOP tunnels at the same time.
This may lead to a violation of bandwidth guarantees, since the NHOP
and NNHOP tunnels were computed independently using the same backup
bandwidth pool, and so they may share a link with enough bandwidth for
only one but not the other.
A similar issue occurs in the case of bi-directional link failure.
Since the two nodes on each side of the link will see the failure of an
adjacent link, unless they can detect that it was a link and not a node
failure, they will be forced to activate the NHOP tunnel protecting the
link, and the NNHOP tunnel protecting the node on the other side.
Essentially, the system will operate as if two failures have occurred
simultaneously when in reality only a single (directed) link failed.
This clearly can result in a violation of a bandwidth guarantee.
To address this issue, one needs a mechanism to differentiate a link
from a node failure.
Note
Vasseur, Charny, Le Faucheur and Achirica 17
draft-vasseur-mpls-backup-computation-00.txt June 2002
It is worth mentioning there are also some cases where:
- just NNHOP bypass tunnels are required (links are already
protected at layer 1 or 2)
- just NHOP bypass tunnels are required (nodes are considered as
''enough'' reliable) and just links are protected against failures.
8. Operations with DS-TE and multiple Class-Types
This section assumes the reader is familiar with Diff-Serv-aware MPLS
Traffic Engineering as specified in [DSTE-REQTS] and [DSTE-PROTO] and
with its associated concepts such as Class-Types (CTs), Bandwidth
Constraints (BCs) and the Russian Dolls bandwidth constraint model.
The bandwidth protection approach described in this document supports
DS-TE and operations with multiple Class-Types.
It is worth mentioning that both the primary and backup bandwidth pools
sizes have to be carefully determined by the network administrator as
their values dictate the congestion level in case of failure, as
discussed bellow. In the absence of failure, up to the max reservable
bandwidth pool (i.e the primary bandwidth pool) of traffic will be
forwarded onto a link. In case of failure, potentially up to "Primary
bandwidth pool" + "backup bandwidth pool" of traffic will be active on
a link. Various scenarios as to what the backup bandwidth should be
reserved for, are discussed in the following sections. The
determination of their values compared to the link speed is a critical
factor.
8.1. Single backup pool
Several bandwidth protection scenarios only require the use of a single
backup pool.
First, when a single Class-Type is used (i.e. network which do not use
Diff-Serv or use Diff-Serv but only enforce a single bandwidth
constraint to all the TE tunnels), bandwidth protection can be achieved
via a single bandwidth pool.
Second, when multiple Class-Types are used, a single backup pool can be
used to provide bandwidth protection to LSPs from a single Class-Type
CTc, which is the active CT with the highest index c, (in other words
the active CT with the smallest Bandwidth Constraint), while LSPs from
the other Class-Types do not get bandwidth protection.
Here is an example of such scenario. Let's consider the following
network where:
- DS-TE and the Russian Dolls bandwidth constraint model are
used
- two Class-Types (CTs) are used:
o CT1 is used for Voice Traffic
Vasseur, Charny, Le Faucheur and Achirica 18
draft-vasseur-mpls-backup-computation-00.txt June 2002
o CT0 is used for Data traffic
From a bandwidth protection perspective, let's assume that:
- Voice traffic (i.e. CT1 LSPs) needs Bandwidth Protection
during failure
- Data traffic (i.e. CT0 LSPs) does not need Bandwidth
Protection during failure.
Let's further assume that the network administrator has elected to use
the notion of backup pool and thus specify bandwidth requirements for
backup tunnels based on the full bandwidth pool of primary tunnels
(i.e. BC1) as configured towards the protected facility (as opposed to
the amount of bandwidth currently used by the primary LSPs; see
Appendix B for a detailed discussion).
Then, for every link the network administrator will configure:
- BC0, the Bandwidth Constraint on the aggregate across all
primary LSPs (CT0+CT1)
- BC1, the Bandwidth Constraint for primary CT1 LSPs
- BCbu, the Bandwidth Constraint for the Backup CT1 LSPs
The bandwidth requirement of each backup LSP is configured based on the
value of BC1 configured towards the facility it protects. In other
words, the backup LSPs are only sized to protect voice traffic
transiting via the protected facility.
Purely for illustration purposes, the diagram below builds on the one
presented in section 9 of [DSTE-PROTO] to represent these bandwidth
constraints in a pictorial manner.
I------------------------------------------------------I ----------------I
I--------------I I I
I CT1 I I I
I Primary I I I
I--------------I I CT1 Backup I
I CT1 + CT0 I I
I------------------------------------------------------I ----------------I
I-----BC1------>
I---------------------------------------------BC0------> I----BCbu------->
Note that while this scenario assumes Data traffic does not need
Bandwidth protection during failure, Data traffic can be either not
protected at all by Fast Reroute or be protected by Fast Reroute but
without bandwidth protection during failure. In the former case, CT0
LSPs transporting Data traffic would not be rerouted into backup LSPs
on failure. In the latter case, CT0 LSPs would be rerouted onto backup
LSPs upon failure; the backup tunnels could either be a different set
of backup tunnel from the backup tunnels for voice, or could be the
same backup tunnels as for Voice assuming appropriate Diffserv marking
Vasseur, Charny, Le Faucheur and Achirica 19
draft-vasseur-mpls-backup-computation-00.txt June 2002
and scheduling differentiation are configured properly, as discussed
below.
From a scheduling perspective, a possible approach is for Voice traffic
to be treated as the exact same Ordered Aggregate (i.e. use the same EF
PHB) whether it is transported on primary LSPs or on backup LSPs. In
that case, on a given link, BC1 and BCbu must clearly be configured in
such a way that the Voice QoS objectives are met when there is
simultaneously, on that link, up to BC1 worth of traffic on primary CT1
LSPs and up to BCbu worth of Voice Traffic on backup LSPs. A more
detailed discussion on scheduling is provided in the following section.
The size of the backup pool BCbu is configured on all links such that
the CT1 LSP QoS objectives are met when there is simultaneously, on
that link, up to BC1 worth of primary LSPs and up to BCbu worth of
backup CT1 traffic.
Notes
- If the objective for CT1 traffic is only to protect CT1 bandwidth
then the network administrator must just make sure that: BC1+BCbuLink Speed, CT0 traffic may experiment congestion during
failure but CT1 traffic is still bandwidth-protected.
Other scenarios can be addressed with a single bandwidth pool. This
includes the case where all Class-Types need bandwidth protection but
it is acceptable to relax delay guarantee to these classes during the
failure and only offer bandwidth protection. Operations is very similar
to the previous scenario described (e.g. size backup tunnel based on
BC0), the only difference is that QoS objectives other than bandwidth
guarantee of other CTs than CT0 are not are not necessarily guaranteed
to be preserved during failure. These CTs only get bandwidth
assurances.
8.2. Multiple backup pools
When DS-TE is used and multiple Class-Types are supported, the
operations described above can be easily extended to multiple bandwidth
pools in the case where backup LSPs are sized based on the actual
amount of established LSPs (See appendix B for discussion on the pros
and cons of this approach): one backup pool can be used to separately
constrain the bandwidth used by backup LSPs of each Class-Type.
In that case, each CT can be given bandwidth protection during failure
with guarantee that each CT will also meet all its respective QoS
objectives during the failure and without any bandwidth wastage.
Vasseur, Charny, Le Faucheur and Achirica 20
draft-vasseur-mpls-backup-computation-00.txt June 2002
Here is an example of such scenario. Let's consider the following
network where:
- DS-TE and the Russian Dolls bandwidth constraint model are
used
- two Class-Types (CTs) are used:
o CT1 is used for Voice Traffic
o CT0 is used for Data traffic
From a bandwidth protection perspective, let's assume that:
- Voice traffic (i.e. CT1 LSPs) needs Bandwidth Protection
during failure
- Data traffic (i.e. CT0 LSPs) also needs Bandwidth Protection
during failure.
Let's further assume that the network administrator has elected to
specify bandwidth requirements for backup tunnels based on the the
actual amount of established primary LSPs (as opposed to the the full
bandwidth pool of primary tunnels as configured towards the protected
facility; see Appendix B for a detailed discussion).
Then, for every link the network administrator will configure:
- BC0, the Bandwidth Constraint on the aggregate across all
primary LSPs (CT0+CT1)
- BC1, the Bandwidth Constraint for primary CT1 LSPs
- BCbu0, the Bandwidth Constraint on the aggregate across all
backup LSPs (CT0+CT1)
- BCbu1, the Bandwidth Constraint on the CT1 backup LSPs
The bandwidth requirement of each CT0 backup LSP is configured based on
the actual amount of established CT0 primary LSPs it protects. The
bandwidth requirement of each CT1 backup LSP is configured based on the
actual amount of established CT1 primary LSPs it protects.
Purely for illustration purposes, the diagram below represents these
bandwidth constraints in a pictorial manner.
I----------------------------------------------I--------------------I
I--------------I I----------I I
I CT1 I I CT1 I I
I Primary I I Backup I I
I--------------I I----------I I
I CT1 + CT0 Primary I CT1+CT0 Backup I
I----------------------------------------------I--------------------I
I-----BC1------> I--BCbu1-->
I-------------------------------------BC0------>I-------BCbu0------->
The size of the backup pool BCbu0 is configured on all links such that
the CT0 LSP QoS objectives are met when there is simultaneously, on
Vasseur, Charny, Le Faucheur and Achirica 21
draft-vasseur-mpls-backup-computation-00.txt June 2002
that link, up to BC0 worth of CT0 primary LSPs and up to BCbu0 worth of
backup CT0 traffic.
The size of the backup pool BCbu1 is configured on all links such that
the CT1 LSP QoS objectives are met when there is simultaneously, on
that link, up to BC1 worth of CT1 primary LSPs and up to BCbu1 worth of
backup CT1 traffic.
In the case where backup LSPs are sized based on the amount of
reservable bandwidth (See appendix B for discussion on the pros and
cons of this approach), it is also possible to extend operations to
multiple bandwidth pools in the same way, but this may result in
bandwidth wastage. This is because, BC1 will be effectively reserved
both from BC1bu and from BC0bu.
Here is an example of such scenario. Let's consider the following
network where:
- DS-TE and the Russian Dolls bandwidth constraint model are
used
- two Class-Types (CTs) are used:
o CT1 is used for Voice Traffic
o CT0 is used for Data traffic
From a bandwidth protection perspective, let's assume that:
- Voice traffic (i.e. CT1 LSPs) needs Bandwidth Protection
during failure
- Data traffic (i.e. CT0 LSPs) also needs Bandwidth Protection
during failure.
Let's further assume that the network administrator has elected to
specify bandwidth requirements for backup tunnels based on the full
bandwidth pool of primary tunnels as configured towards the protected
facility (as opposed to the amount of bandwidth currently used by the
primary LSPs; see Appendix B for a detailed discussion).
Then, for every link the network administrator will configure:
- BC0, the Bandwidth Constraint on the aggregate across all
primary LSPs (CT0+CT1)
- BC1, the Bandwidth Constraint for primary CT1 LSPs
- BCbu0, the Bandwidth Constraint on the aggregate across all
backup LSPs (CT0+CT1)
- BCbu1, the Bandwidth Constraint on the CT1 backup LSPs
The bandwidth requirement of each CT1 backup LSP is configured based on
the value of BC1 configured towards the facility it protects. The
bandwidth requirement of each CT0 backup LSP is configured based on the
value of BC0 configured towards the facility it protects. Thus,
effectively the CT1 backup LSP and CT0 backup LSP will have an
aggregate bandwidth requirement of BC0+BC1 which represents a bandwidth
wastage since we know that the aggregate primary bandwidth across CT0
and CT1 is actually limited to BC0 (since BC0 is a bandwidth constraint
on CT0+CT1).
Vasseur, Charny, Le Faucheur and Achirica 22
draft-vasseur-mpls-backup-computation-00.txt June 2002
Operations with multiple backup pools will be discussed in more details
in subsequent versions of this draft.
9. Interaction with scheduling
The bandwidth protection approach described in this document does not
require any enhancement or modification to MPLS scheduling mechanisms
beyond those defined in [MPLS-DIFF]. In particular, scheduling can
remain entirely unaware of Fast Reroute and bandwidth protection; in
particular this approach does not require that scheduling behave
differently depending on whether a packet is transported on a primary
LSP or a backup LSP, nor does it require per-LSP scheduling.
This approach simply requires that the existing MPLS scheduling
mechanisms (e.g. Diff-Serv PHBs) are configured in a manner which is
compatible with the goal of bandwidth protection, because while the
bandwidth protection allocates bandwidth appropriately in the control
plane, it is scheduling which is responsible for the actual enforcement
in the data path of the corresponding service rates to packets in a way
which will achieve the targeted bandwidth protection.
The details of which configuration is appropriate depends on multiple
parameters such as the details of the Diff-Serv policy, the bandwidth
protection policy and the number of DS-TE Class-Types supported. Thus,
it is outside the scope of this draft.
For illustration purposes, we can expand on the scheduling aspects in
the example discussed in the previous section. A possible scheduling
approach based on MPLS Diff-Serv is the following:
- let's assume Voice uses EF PHB and Data uses AF11 ,AF12, AF21
and AF22 PHBs
- E-LSPs with preconfigured EXP<-->PHB mapping can be used
with:
o EXP=eee maps to EF
o EXP=aa0 maps to AF11
o EXP=aa1 maps to AF12
o EXP=bb0 maps to AF21
o EXP=bb1 maps to AF22
- separate E-LSPs are established for Voice and for Data
- Voice E-LSPs are established in CT1
- Data E-LSPs are established in CT0
- Separate E-LSPs are established for backup constrained by
Bcbu (but with signaled bandwidth set to zero as discussed in
section 6).
- BC1 and BCbu are configured on every link so that the EF PHB
can guarantee appropriate QoS to voice when there is BC1+BCbu
worth of voice traffic
- The uniform Diff-Serv tunneling mode defined in section 2.6
of [MPLS-DIFF] is used on the backup tunnels. In particular,
Vasseur, Charny, Le Faucheur and Achirica 23
draft-vasseur-mpls-backup-computation-00.txt June 2002
when a packet is steered into a backup tunnel by the PLR
(i.e. when the backup tunnel label entry is pushed onto the
packet) the EXP field of the packet is copied into the EXP
field of the backup tunnel label entry.
Then, upon a failure:
- voice packets have their EXP=eee regardless of whether they
are transported on a primary tunnel or backup tunnel. Thus
they will be scheduled by the EF PHB. Since our bandwidth
protection approach ensures that there is less CT1 LSPs than
BC1 and less backup LSPs than BCbu, and since we have
configured BC1 and BCbu so that EF can cope with that
aggregate load, QoS is indeed guaranteed to voice during
failure.
- Data packets have their EXP=aax or EXP=bbx regardless of
whether they are transported on a primary tunnel or a backup
tunnel. Thus, it is clear that they do not steal bandwidth
from the EF PHB.
In the example described in the previous section, we mentioned several
possible protection policies for Data. Let's assume that Data is
protected by Fast Reroute but without Bandwidth protection and let's
assume that the same backup tunnels are used as for voice. Then it must
be noted that even if Data is injecting traffic into the backup LSPs
(whose bandwidth constraint do NOT factor such load since they only
factor the voice traffic), this does NOT compromise the voice bandwidth
protection in anyway since:
- the admission control performed over backup LSPs factored the
voice load over the EF PHB
- the data packets transported on the backup LSP have their
EXP=aax or EXP=bbx and thus are scheduled in the AF PHBs
without affecting the EF PHB.
On the other hand, Data packets may experience QoS degradation during
failure. This is because a given link, in addition to data packets on
primary CT0 LSPs for which admission control has been performed, may
also receive data packets on backup LSPs for which effectively no
admission control has been performed (since this load was not factored
in the sizing of the backup LSPs). This is in line with the assumption
that Data traffic did not need bandwidth protection during failure.
In the particular case where the PLR could not establish a backup
tunnel with the full requested amount of bandwidth (due to some lack of
bandwidth in the backup pool) and instead established a backup tunnel
with a smaller bandwidth, when rerouting LSPs onto this backup tunnel,
the PLR may ensure that the amount of rerouted primary LSPs complies
with the actual bandwidth of the backup tunnel. Otherwise, this would
simply violate bandwidth protection (for traffic on this backup tunnel
as well as for all traffic on any LSP using the same PHBs) because more
traffic than reserved for would end up in the backup tunnel. In that
Vasseur, Charny, Le Faucheur and Achirica 24
draft-vasseur-mpls-backup-computation-00.txt June 2002
case, the primary LSPs which did not fit into the backup LSP would have
their traffic dropped.
10. Routing and signaling extensions
10.1. Routing (IGP-TE) extensions
In this section, we define an IGP-TE routing extensions to signal the
bandwidth protection pool. This extension is identical to the extension
defined in [BP-PLACEMENT] and is defined for ISIS-TE and OSPF-TE.
As explained earlier, this extension is purely optional and can be
considered as useful but not mandatory.
One new sub TLVs (in Link TLVs of TE LSA for OSPF, and in IS
reachability TLVs for ISIS) is defined:
Max reservable protection bandwidth sub-TLV: this sub-TLV contains
the maximum protection bandwidth that can be reserved on this link in
this direction (from the node originating the LSA to its
neighbors). The maximum protection bandwidth is encoded in 32 bits in
IEEE floating-point format. The units are bytes per second.
OSPF and ISIS types are TBD.
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| TBD | 4 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Max res prot bandwidth |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Again, the bandwidth protection pool being announced by IGP-TE is a
static value i.e does not dynamically change as backup TE LSP are set
up, which preserves IGP scalability.
As the bandwidth protection pool is being defined on a per link basis,
this allows for different policies depending on the link
characteristics.
10.2. Signaling (RSVP-TE) extensions
10.2.1. PCC -> PCS signaling : specification of a set of
constraints
The PCC (an LSR) will provide to the PCS a set of constraints to
satisfy for the backup tunnel path computation. The PCC-PCS signaling
protocol used in this draft is based on [PATH-COMP]. A new object
Vasseur, Charny, Le Faucheur and Achirica 25
draft-vasseur-mpls-backup-computation-00.txt June 2002
called BACKUP-TUNNEL, related to backup tunnel is defined in this
section.
As defined in [PATH-COMP], the path computation request has the
following format:
::= [ ]
[ |
] ... ]
[ ]
[ ]
[ ]
[]
[]
[]
[ ]
[ ... ]
::=
[ ]
[ ]
There are several constraints that should be taken into account when
computing the backup tunnel paths that have already been described in
section 6.3:
- element to protect,
- bandwidth,
- affinities,
- Max number of backup tunnels, (per link or per pair of links
through a node)
- Minimum bandwidth on a single backup tunnel,
- CT to protect,
- Existing backup tunnels,
- other optional parameters, e.g. maximum allowed propagation
delay increase of the backup tunnel over the segment of the
primary path protected by the tunnel.
Some are optional (see bellow).
The PCC can make use of a single path computation request even if
multiple backup tunnel path computations are requested. In that case,
the PCC must include a separate BACKUP-TUNNEL object per request.
BACKUP-TUNNEL Class-Num is [TBD by IANA] - C-Type is [TBD by
IANA]
0 1 2 3
Vasseur, Charny, Le Faucheur and Achirica 26
draft-vasseur-mpls-backup-computation-00.txt June 2002
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Flag | Length | ETP | CT |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Resource-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
// //
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Bandwidth |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Include-any |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Exclude-any |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Include-all |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| MAX-NB-BACKUP-TUNNEL |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| MIN-BW-BACKUP-TUNNEL |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Flags: 8 bits
0x01: specifies that the requesting PCC provides a set (possibly
reduced to a single element) of existing backup tunnels. For each
existing backup tunnel the corresponding ERO will be included
within the Path computation request.
0x02: specifies to the PCS that in case of negative reply (the PCC
cannot find a set of backup tunnels that fulfill the set of
requirements), the PCS should provide in the path computation
reply the best possible set of backup tunnels i.e the set of
backup tunnels that will protect the maximum possible amount of
bandwidth for the protected element.
Length
The Length contains the total length of the subobject in bytes.
The Length MUST be at least 4, and MUST be a multiple of 4.
ETP (Element to protect): 8 bits
0x00: Link
0x01: Node
0x02: SRLG
CT: Class-type to protect
Resource ID: identifies the resource to protect
- for a link, the PCC must specify the link IP address,
Vasseur, Charny, Le Faucheur and Achirica 27
draft-vasseur-mpls-backup-computation-00.txt June 2002
- for a node, the PCC must specify one the interface IP addresses of
the node or its router ID,
- for a SRLG, the PCC must specify the SRLG number
The BACKUP-TUNNEL object may contain more than one RESOURCE-ID
field, provided all the resources to protect (identified by their
respective RESOURCE-ID) share the same bandwidth protection
constraints.
Bandwidth: (32-bit IEEE floating point integer) in bytes-per-
second.
Affinities (optional)
This parameter is optional and must be set to 0x00000000 if not
used.
Exclude-any
A 32-bit vector representing a set of attribute filters
associated with a backup path any of which renders a link
unacceptable.
Include-any
A 32-bit vector representing a set of attribute filters
Associated with a backup path any of which renders a link
acceptable (with respect to this test). A null set (all bits set
to zero)automatically passes.
Include-all
A 32-bit vector representing a set of attribute filters
Associated with a backup path all of which must be present for a
link to be acceptable (with respect to this test). A null set
(all bits set to zero) automatically passes.
MAX-NB-BACKUP-TUNNEL: Maximum number of backup tunnels
This parameter is optional and must be set to 0x00000000 if not
used.
MIN-BW-BACKUP-TUNNEL: Minimum bandwidth of any element of the backup
tunnel set.
This parameter is optional and must be set to 0x00000000 if not
used.
Vasseur, Charny, Le Faucheur and Achirica 28
draft-vasseur-mpls-backup-computation-00.txt June 2002
10.2.2. PCS -> PCC signaling - - sending the computed set of
backup tunnels
After having processed a PCC request, the PCS will send a path
computation reply to the PCC.
The likelihood of finding a solution that will obey the set of
constraints will of course be conditioned by:
- the network resources (and particularly the backup
bandwidth/link bandwidth ratio)
- the set of constraints.
There are two possible results:
- the request can be satisfied (positive reply)
- the new request cannot be (fully) satisfied (negative reply).
As defined in PATH-COMP, the PCS' path computation reply message will
have the following form:
::= [ ]
[ | ]...]
[ ]
[ ]
[ []] ...
[ ]
[ ... ]
For each backup tunnel, the Path Computation Reply will contain:
- a BACKUP-TUNNEL object specifying the characteristics of the
computed backup tunnel (identification of the resource it
protects (ETP, resource-ID, ...) and backup tunnels attributes
(bandwidth, affinities). The MAX-NB-BACKUP-TUNNEL and MIN-BW-
BACKUP-TUNNEL fields will be set to 0x00000000.
- the Path of the computed backup tunnel (EXPLICIT_ROUTE).
A set of backup tunnels may be reduced to a single element if the PCS
can find a single backup tunnel that fulfills the requirements.
11. Backup tunnel - Make before break
In case of backup tunnel path change, the new backup tunnel may be set
up using make before break. This may or not be possible depending on
the change in the set of backup tunnels.
12. Stateless versus Statefull PCS
There are basically two options for the PCS:
Vasseur, Charny, Le Faucheur and Achirica 29
draft-vasseur-mpls-backup-computation-00.txt June 2002
- can be statefull: the PCS registers the various backup tunnels
computation requests and results. It will also monitor the network
states (backup tunnels in place, ...)
- can be stateless: the PCS does not maintain any state. This approach
is the recommended approach.
13. Packing algorithm
Once the set of backup tunnels is in place, the PLR should, for each
protected TE LSP successfully signaled, select a corresponding backup
tunnel. As per defined in [FAST-REROUTE], the bandwidth protection
requirement for the protected LSP can be specified using the FAST-
REROUTE object or by setting the ''Bandwidth protection desired'' bit
in the SESSION-ATTRIBUTE of the Path message . Based on the signaled
backup bandwidth requirement for the protected LSP, the PLR should
appropriately select the backup tunnel to use for the protected TE LSP,
making sure the requested backup bandwidth requirement is met.
14. Interoperability in a mixed environment
There could potentially be some interoperability issues when conformant
and non conformant nodes to this draft are mixed within the same
network. The following interoperability issues categories could be
identified:
* Ability of LSRs to communicate with the server: if the PCS is an LSR,
other LSRs need to communicate with the server using the signaling
extensions proposed in this draft,
* Interaction of different bandwidth protection FRR techniques.
- networks not supporting backup bandwidth pools,
- interaction with backup tunnels using explicit bandwidth reservation,
- interaction with 0-bandwidth best effort TE LSPs.
Interoperability issues will be covered in detailed in a further
revision of this draft.
15. Security Considerations
The practice described in this draft does not raise specific security
issues beyond those of existing TE.
Vasseur, Charny, Le Faucheur and Achirica 30
draft-vasseur-mpls-backup-computation-00.txt June 2002
References
[TE-REQ] Awduche et al, Requirements for Traffic Engineering over MPLS,
RFC2702, September 1999.
[OSPF-TE] Katz, Yeung, Traffic Engineering Extensions to OSPF, draft-
katz-yeung-ospf-traffic-05.txt, June 2001.
[ISIS-TE] Smit, Li, IS-IS extensions for Traffic Engineering, draft-
ietf-isis-traffic-03.txt, June 2001.
[RSVP-TE] Awduche et al, "RSVP-TE: Extensions to RSVP for LSP Tunnels",
RFC3209, December 2001.
[CR-LDP] Jamoussi et al., "Constraint-Based LSP Setup using LDP",
draft-ietf-mpls-cr-ldp-05.txt, February 2001
[METRICS] Fedyk et al, ''Multiple Metrics for Traffic Engineering with
IS-IS and OSPF'', draft-fedyk-isis-ospf-te-metrics-01.txt, November
2000.
[DS-TE] Le Faucheur et al, ''Requirements for support of Diff-Serv-aware
MPLS Traffic Engineering'', draft-ietf-tewg-diff-te-reqts-01.txt, June
2001.
[PATH-COMP] Vasseur et al, ''RSVP Path computation request and reply
messages'', draft-vasseur-mpls-computation-rsvp-02.txt, November 2001.
[FAST-REROUTE] Pan, P. et al., "Fast Reroute Techniques in
RSVP-TE", Internet Draft, draft-ietf-mpls-rsvp-lsp-fastreroute-00.txt
, January 2002
[BP-PLACEMENT] Leroux, Calvignac, ''A method for an Optimized Online
Placement of MPLS Bypass Tunnels'', draft-leroux-mpls-bypass-placement-
00.txt, February 2002.
[KINI] Kini et al, ''Shared Backup Label Switched Path Restoration'',
draft-kini-restoration-shared-backup-01.txt, May 2001.
[ISIS-PCSD] Vasseur and Shand, ''IS-IS Path Computation Server
discovery TLV'', draft-vasseur-mpls-isis-pcsd-discovery-00.txt, work in
progress.
[OSPF-PCSD] Vasseur, Psenak, ''OSPF Path Computation Server discovery'',
draft-vasseur-mpls-ospf-pcsd-discovery-00.txt, work in progress.
Vasseur, Charny, Le Faucheur and Achirica 31
draft-vasseur-mpls-backup-computation-00.txt June 2002
Authors' Address:
Jean Philippe Vasseur
Cisco Systems, Inc.
11, rue Camille Desmoulins
92782 Issy les Moulineaux Cedex 9
France
Email: jpv@cisco.com
Anna Charny
Cisco Systems, Inc.
300 Apollo Drive
Chelmsford, MA 01824
USA
Email: acharny@cisco.com
Francois Le Faucheur
Cisco Systems, Inc.
Village d'Entreprise Green Side - Batiment T3
400, Avenue de Roumanille
06410 Biot-Sophia Antipolis
France
Phone: +33 4 97 23 26 19
Email: flefauch@cisco.com
Javier Achirica
Telefnica Data Espaħa
Beatriz de Bobadilla, 14
28040 Madrid
Spain
javier.achirica@telefonica-data.com
Vasseur, Charny, Le Faucheur and Achirica 32
draft-vasseur-mpls-backup-computation-00.txt June 2002
Appendix A: Limitations/inefficiency of the independent CSPF-based
computation model
Let's give a simple illustration of the case where PLRs are
using an independent based CSPF approach and fail to find a
feasible placement of the backup tunnels.
R6---------R7
|\ |
| \ |
| \ |
R1----R2---R3----R4----R5
| |
| |
| |
R8---------R9
The goal is to find the backup tunnels protecting node R3.
Let's assume that the amount of bandwidth than needs to be
protected on links adjacent to R3 is given by:
R6-R3=5M
R2-R3=10M
Assume further that bandwidth on other links available for
placement of the backup tunnels is as follows:
R6-R7=10M
R6-R2=20M
R2-R8=5M
other links=100M
Bandwidth on a link in each direction is assumed the same (e.g.
link R8-R2 is also 5M).
In a distributed and non coordinated setting, the order in which
the direct neighbors of R3 compute and place their backup tunnels
protecting against the failure of R3 can be arbitrary.
Suppose R6 tries to compute a NNHOP backup tunnel to R4 with
bandwidth 5M and selects the shortest path to R4 with available
bandwidth and bypassing R3. That is R6-R7-R4. When R2 tries to
compute a NNHOP backup tunnel to R4 with bandwidth 10M, it
discovers that there in no feasible path it can take. In
contrast, and independent server using a more sophisticated
algorithm could discover this condition and find that the
solution:
NNHOP backup tunnel from R6 to R4: R6-R2-R8-R9-R4 (BW=5M),
NNHOP backup tunnel from R2 to R4: R2-R6-R7-R4 (BW=10M),
Vasseur, Charny, Le Faucheur and Achirica 33
draft-vasseur-mpls-backup-computation-00.txt June 2002
NNHOP backup tunnel from R4 to R2: R4-R7-R6-R2 (BW=5M),
NNHOP backup tunnel from R4 to R6: R4-R9-R8-R2-R6 (BW=10M),
NNHOP backup tunnel from R6 to R2: R6-R2 (BW=5M),
NNHOP backup tunnel from R2 to R6: R2-R6 (BW=5M)
satisfies the constraints. Since the general problem of finding a
feasible placement of given bandwidth demands in a general-
topology network is well-known to be NP-complete, it could be
argued that a centralized server cannot be expected to implement
an algorithm that is always guaranteed to find a solution in a
reasonable time in all cases anyway. While it is certainly true,
it is quite clear that a server-based implementation can run a
heuristic algorithm that is much more likely to find a solution
than simple greedy CSPF-based approach. Moreover, the centralized
model is much more amenable to supporting various optimality
criteria not available with the simple CSPF-based approach.
Vasseur, Charny, Le Faucheur and Achirica 34
draft-vasseur-mpls-backup-computation-00.txt June 2002
Appendix B: Bandwidth to protect
There are two different approaches for the bandwidth constraint of the
backup tunnels.
The backup tunnel bandwidth may be based on:
- the amount of reservable bandwidth on a particular network
resource,
- the sum of bandwidths actually reserved by established TE
LSPs on a particular resource.
Solution 1: primary reservable pool
In this case, the backup tunnel bandwidth requirement is based on the
primary reservable pool we need to protect.
Example:
R6---R7----R8
|\ | / |
| -- | -- |
| \|/ |
R1----R2---R3----R4----R5
| / \ |
| -- -- |
|/ \ |
R9---------R10
Objective: find a set of backup tunnels from R2 to R4 to protect R2
from a node failure of R3.
In this case, the backup tunnel bandwidth requirement is being driven
by the smaller of amount of max reservable bandwidth (the bandwidth
pools) defined on the links R2-R3 and R3-R4 (potentially multiplied by
some factor), independently on the current state of bandwidth
reservation on these links. In case of nested pools of bandwidth, the
outmost pool could be taken into account (that would cover all pools
nested inside) or just one of the subpools.
With this solution 1, in the example above, when R2 requests the server
to compute for it the backup tunnels protecting its traffic traversing
R3 against R3's failure, it should request the computation of 6
different NNHOP backup tunnels with headend in R2 and tailend at each
other direct neighbor of R3. The bandwidth of each of these backup
tunnels is determined by the minimum of the max reservable bandwidth of
the pool for which protection is desired on the link R2-R3 and the link
connecting R3 to the corresponding neighbor. For example, if max
reservable bandwidth is 10 Mbps on link R2-R3, and 8 Mbps on link R3-
R4, then the backup tunnel from R2 to R4 must have the bandwidth of
8Mbps available to it.
Vasseur, Charny, Le Faucheur and Achirica 35
draft-vasseur-mpls-backup-computation-00.txt June 2002
The obvious benefit of this approach is of course that the backup path
computation is not impacted by the dynamic network state (the TE LSPs
currently in place) which is a serious advantage in term of stability.
A new backup path computation should just be triggered in case of
network topology change (link/node down, change in the reservable
amount of bandwidth on a given link, ...). The drawback is that the
bandwidth requirement may be substantially higher than needed if the
actual amount of capacity is much larger than the actual amount of
reserved capacity of the TE LSPs in place; the higher is the bandwidth
requirement for the backup tunnel, the lower is the likelihood to find
a solution.
Aggregate bandwidth constraints for backup tunnels
When protecting a bi-directional link, an SRLG or a node, multiple
backup tunnels are typically required. For example, a bi-directional
link protection requires at least one backup tunnel for each of the two
directions of the link. For SRLG, at least one backup tunnel is
required for each link in the SRLG. For a node, at least one backup
tunnel is required for every pair of direct neighbors of this node.
At first glance, it may seem that if tunnels T1,T2,...TK with bandwidth
requirements b1,b2,..Bk protecting against a failure of some element F
traverse some link L, then link L must have at least b1+b2+...+bk
bandwidth available for backup placement. It is indeed always true for
link and SRLG protection. For node protection it is more complicated.
In the case when the actual amount of primary bandwidth is protected,
the above statement is also true. However, for the case when the backup
pool is protected, this statement is unnecessarily conservative.
To see this, consider the above example, and assume that the primary
pools (max reservable bandwidth for a particular subpool) on all links
adjacent to R3 are 10 Mbps, except for the link R3-R4, which has the
primary pool of 8 Mbps in each direction. Note now that backup tunnels
T1 (R6-R4) and T2(R2-R4) each need 8 Mbps. However, the total amount of
primary traffic traversing paths R6-R3-R4 and R2-R3-R4 is bounded by
the primary pool of link R3-R4, and so the aggregate bandwidth
requirements of both backups tunnels is only 8Mbps, and not 16Mbps. A
path computation server implementing solution 1 SHOULD take such
aggregate constraints into consideration when computing backup tunnels
placement.
Solution 2: total amount of bandwidth actually reserved on a given link
Another option is to make the backup tunnel bandwidth requirement a
function of the actual amount of reserved bandwidth. In the diagram
above, R2 would request a set of backup tunnels so that the backup
bandwidth is equal to the sum of the bandwidths of the currently
established TE LSPs crossing the R2-R3 link. This value may be
Vasseur, Charny, Le Faucheur and Achirica 36
draft-vasseur-mpls-backup-computation-00.txt June 2002
multiplied by some factor to allocate some spare room for new coming TE
LSPs.
With this solution, R2 would send a request to the PCS for the actual
amount of reserved bandwidth between it and each of the direct
neighbors of R3 to which it has primary traffic. For example, if there
is no primary TE LSP established between R2 and R6, there is no need to
request a backup tunnel connecting R2 to R6. Furthermore, if the total
bandwidth of all TE LSPs between R2 and R4 traversing R3 is 2 Mbps,
then the bandwidth requirement of the backup tunnel R2-R4 can be 2 Mbps
instead of 8Mbps in solution 1.
Note however, that the backup tunnels are signaled with zero bandwidth
and therefore do not reserve any bandwidth. Therefore, as long as the
set of backup tunnels protecting the entire pool exist (and can be
found by the algorithm computing their placement), the bandwidth
savings of solution 2 over solution 1 is irrelevant. However in the
cases when the backup bandwidth is so scarce that the backup tunnels
protecting the entire bandwidth pools cannot be found, solution 2
clearly provides a benefit.
The main drawback of solution 2 is the need for a potentially large
number of backup tunnel recomputations each time TE LSPs are set
up/torn down which creates additional load on the device computing the
placement, and results in additional signaling overhead. Furthermore,
recomputing and resignaling the new set of backup tunnels may take some
(albeit relatively short) time, leaving all primary TE LSPs traversing
the affected elements temporarily unprotected.
The risk of instability may be reduced by the use of some UP/DOWN
thresholds. In this case, each time a new TE LSP is set up, if a UP
threshold is crossed a new backup tunnel path computation is triggered.
Optionally, a DOWN threshold scheme may be used to better optimize the
backup bandwidth usage. In this case, when a TE LSP is torn down, if a
DOWN threshold is crossed, a backup tunnel path computation is
triggered. For obvious reasons, it is expected to have different UP and
DOWN thresholds.
Mix of solutions 1 and 2: another approach is also to combine the two
solutions described above.
Suppose the objective of full bandwidth protection cannot be met by the
PCS: in case of negative reply from the PCS that cannot find a solution
to the requested constraints, some algorithms may be implemented to
find the best possible solution (the closest to the initial request).
Three options exist:
- option 1: the intelligence is on the PCC. The PCC will send several
requests to the PCS until it gets a positive reply.
- option 2: the intelligence is on the PCS. The PCS in case of negative
reply tries to find the ''best'' possible solution and suggests those new
Vasseur, Charny, Le Faucheur and Achirica 37
draft-vasseur-mpls-backup-computation-00.txt June 2002
values to the PCC. Then the PCC will decide whether it can accept the
new values. If yes, it will resend a new request to the PCS with the
suggested value to get the result. Option 2 requires less signaling
overhead than option 1.
- option 3: the PCS directly answers with the best possible solution.
1) in solution 1 all bandwidth information is available at the PCS,
so there is actually no need to signal the bandwidth at all
2) in solution 2 or a mix, the server may or may not have primary
bandwidth info (e.g. is an LSR ''protects itself'', it already knows
all the actual primary bandwidth requirements, but if a PCS
protects some other element, in this case primary bandwidth needs
to be communicated to it.
Option 3 requires less signaling overhead than option 2.
Vasseur, Charny, Le Faucheur and Achirica 38
draft-vasseur-mpls-backup-computation-00.txt June 2002
Appendix C: Backup tunnel path computation triggering and path changes
This appendix deals with:
- backup tunnel path computation triggers,
- backup tunnel path changes,
Backup tunnel path computation triggers will of course depends on
whether solution 1 or 2 has been adopted (see Appendix B).
With solution 1: primary reservable pool
Backup tunnel path computation may be triggered when the network
resource to protect first comes up or when the first protected LSP is
signaled.
This is a matter of local policy.
Then the backup tunnel path computation is triggered:
- when the network topology has changed. Following a network
failure (link/node), the PLR may decide, after some
configurable time has elapsed, to trigger a new path
computation. This includes the situation where a new neighbor
of an already protected node comes up. This is a topology
change.
- when the reservable bandwidth of the protected section
changes,
- when the amount of bandwidth protection pool changes,
- when a backup tunnel path reoptimization is triggered: a PCC
may desire to trigger a backup tunnel path computation at any
time (using for instance a timer driven approach) in order to
see whether a more optimal set of backup tunnels could be
found.
With solution 2: sum of the bandwidth actually reserved on a given link
Backup tunnel path computation is triggered:
- when the network topology has changed. Following a network
failure (link/node), the PLR may decide, after some
configurable time has elapsed, to trigger a new path
computation. This includes the situation where a new neighbor
of an already protected node comes up. This is a topology
change.
- when the reservable bandwidth of the protected section
changes,
- when the amount of bandwidth protection pool changes,
- when the actual amount of reserved bandwidth changes (e.g
when a TE LSP is setup or torn down, or when a UP/DOWN
threshold is crossed)
- when a backup tunnel path reoptimization is triggered: a PCC
may desire to trigger a backup tunnel path computation at any
time (using for instance a timer driven approach) in order to
Vasseur, Charny, Le Faucheur and Achirica 39
draft-vasseur-mpls-backup-computation-00.txt June 2002
see whether a more optimal set of backup tunnels could be
found.
Backup tunnel path changes
Various conditions may generate some changes of existing backup tunnels
paths:
(1) when a backup tunnel path computation has been triggered
and as a result a new set of backup tunnels has been computed
that differs from the already in place setup (because the
backup tunnel constraints have changed or a more optimal backup
tunnel path exists),
(2) when as a result of a new backup path computation that has
been triggered by another node, the PCS has computed a new set
of backup tunnels for the node.
(1) is obvious.
Example of (2)
R6---R7----R8
|\ | / |
| -- | -- |
| \|/ |
R1----R2---R3----R4----R5
| / \ |
| -- -- |
|/ \ |
R9---------R10
As an example, suppose:
- Max backup bandwidth pool size along the R6-R7-R8-R4 path is
10M
- Max backup bandwidth pool size along the R2-R9-R10-R4 path is
15M
- On R6, the backup tunnel T1 to protect R6-R3-R4:
Min(R6-R3,R3-R4)=20M
Backup tunnel T1: path=R6-R7-R8-R4, bandwidth=10M
- On R2, the backup tunnel T2 to protect R2-R3-R4:
Min(R2-R3,R3-R4)=10M
Backup tunnel T2: path=R2-R9-R10-R4, bandwidth=5M
For some reason, R6 triggers a new backup tunnel path
computation, requesting for more bandwidth (15M).
To satisfy this new constraint, the PCS will find the following
solutions:
T1: R6-R2-R9-R10-R4
Vasseur, Charny, Le Faucheur and Achirica 40
draft-vasseur-mpls-backup-computation-00.txt June 2002
T2: R2-R6-R7-R8-R4
Which implies to reroute T2, although the backup requirements
of R2 have not changed.
This example shows that a change in a set of backup tunnels for a node
may have some consequences on the set of backup tunnels for some other
nodes.
Vasseur, Charny, Le Faucheur and Achirica 41
draft-vasseur-mpls-backup-computation-00.txt June 2002
Appendix D ''Push'' versus ''Pull'' mode
As discussed in Appendix C, a backup tunnel request from a node X may
result in some changes of the set of backup tunnels for other nodes.
Two scenarios may be implemented:
''Push'' mode: in this scenario, upon the receipt of a backup tunnel path
computation request, the PCS will trigger a simultaneous computation of
backup tunnels for all its neighbors and, in turns, returns the sets of
backup tunnels to all its neighbors (this includes not only the
requesting node but also all the PCS' neighbors).
The corresponding finite state machine would be:
(1) When a new backup tunnel path computation is triggered (see
appendix C), the PCC sends a request to the PCS specifying a set of
constraints (see section 6.3).
(2) When receiving a backup tunnel path computation request, the PCS
will:
(2.1) Optionally first request the set of backup tunnels already in
place to all its neighbors. See note 2 bellow.
(2.2) Perform the backup tunnel path computation simultaneously for all
its neighbors.
Two different situations may happen:
(2.2.1) the new request cannot be (fully) satisfied. In this
case, as defined in [PATH-COMP], the PCS will send a negative
reply including a NO-PATH-AVAILABLE object. Optionally, this
object may indicate the constraint that could not be fulfilled
and also optionally a suggested value for this constraint for
which a solution could have been found. The PCS may use an
algorithm to find the closest solution to initial request.
Optionally, as previously discussed, the PCS may return the
closest possible solution that could be found.
(2.2.2) the new request can be satisfied.
(2.3) send the new sets of backup tunnel to each neighbor
(2.4) each PCS' neighbor will then compare the new set of backup
tunnel(s) to the already in place set of backup tunnels. In case of no
change, then stop. If the new set of backup tunnel differs from the set
of backup tunnels already in place, the node will tear down the
existing backup tunnels and sets up the new set of backup tunnels
optionally with a make before break (if possible).
Note 1: if a PCC request cannot be fully satisfied by the PCS, as
discussed above, some algorithm may be used to find the closest
possible solution to the request. In this case, the PCS will provide
the set of backup tunnels and the amount of protected bandwidth. This
means the node will be partially protected (i.e the amount of protected
bandwidth is less than the amount of setup TE LSPs/reservable
bandwidth).
Vasseur, Charny, Le Faucheur and Achirica 42
draft-vasseur-mpls-backup-computation-00.txt June 2002
Note 2: this may be a very beneficial optimization if the PCS is
capable of minimizing the incremental change (problem known as Minimal
perturbation problem). A statefull PCS will have the knowledge of the
existing backup tunnels. A stateless PCS will have, upon the receipt of
the backup tunnel path computation request, to poll its neighbors to
get the sets of existing backup tunnels as well as the other parameters
(this would imply some additional signaling extension to [PATH-COMP]).
''Pull'' mode: in this mode, the PCS is not allowed to send to a node a
new set of backup tunnels unless explicitly requested by the node. On
the other hand, upon the receipt of a backup tunnel path computation
request from node X, the PCS can still trigger a simultaneous
computation for all its neighbors, provides the output to the
requesting node and registered the sets of backup tunnels of other
neighbors for a future use, provided the PCS is statefull.
Vasseur, Charny, Le Faucheur and Achirica 43