Use of Interior Gateway Protocol (IGP) Metric as a second MPLS Traffic Engineering (TE) Metric
How routers can use the ordinary IGP link cost as a second yardstick for placing MPLS traffic engineering tunnels.
Le Faucheur, F., Uppili, R., Vedrenne, A., Merckx, P., and T. Telkamp, "Use of Interior Gateway Protocol (IGP) Metric as a second MPLS Traffic Engineering (TE) Metric", BCP 87, RFC 3785, DOI 10.17487/RFC3785, May 2004, <https://www.rfc-editor.org/info/rfc3785>.
@misc{rfc3785,
series = {Request for Comments},
number = 3785,
howpublished = {RFC 3785},
publisher = {RFC Editor},
doi = {10.17487/RFC3785},
url = {https://www.rfc-editor.org/info/rfc3785},
author = {F. Le Faucheur and R. Uppili and A. Vedrenne and P. Merckx and T. Telkamp},
title = {{Use of Interior Gateway Protocol (IGP) Metric as a second MPLS Traffic Engineering (TE) Metric}},
pagetotal = 8,
year = 2004,
month = may,
} RFC text
Contents
Network Working Group F. Le Faucheur
Request for Comments: 3785 R. Uppili
BCP: 87 Cisco Systems, Inc.
Category: Best Current Practice A. Vedrenne
P. Merckx
Equant
T. Telkamp
Global Crossing
May 2004
Use of Interior Gateway Protocol (IGP) Metric
as a second MPLS Traffic Engineering (TE) Metric
Status of this Memo
This document specifies an Internet Best Current Practices for the
Internet Community, and requests discussion and suggestions for
improvements. Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2004). All Rights Reserved.
Abstract
This document describes a common practice on how the existing metric
of Interior Gateway Protocols (IGP) can be used as an alternative
metric to the Traffic Engineering (TE) metric for Constraint Based
Routing of MultiProtocol Label Switching (MPLS) Traffic Engineering
tunnels. This effectively results in the ability to perform
Constraint Based Routing with optimization of one metric (e.g., link
bandwidth) for some Traffic Engineering tunnels (e.g., Data Trunks)
while optimizing another metric (e.g., propagation delay) for some
other tunnels with different requirements (e.g., Voice Trunks). No
protocol extensions or modifications are required. This text
documents current router implementations and deployment practices.
1. Introduction
Interior Gateway Protocol (IGP) routing protocols (OSPF and IS-IS) as
well as MultiProtocol Label Switching (MPLS) signaling protocols
(RSVP-TE and CR-LDP) have been extended (as specified in [ISIS-TE],
[OSPF-TE], [RSVP-TE] and [CR-LDP]) in order to support the Traffic
Engineering (TE) functionality as defined in [TE-REQ].
These IGP routing protocol extensions currently include advertisement
of a single additional MPLS TE metric to be used for Constraint Based
Routing of TE tunnels.
However, the objective of traffic engineering is to optimize the use
and the performance of the network. So it seems relevant that TE
tunnel placement may be optimized according to different optimization
criteria. For example, some Service Providers want to perform
traffic engineering of different classes of service separately so
that each class of Service is transported on a different TE tunnel.
One example motivation for doing so is to apply different fast
restoration policies to the different classes of service. Another
example motivation is to take advantage of separate Constraint Based
Routing in order to meet the different Quality of Service (QoS)
objectives of each Class of Service. Depending on QoS objectives one
may require either (a) enforcement by Constraint Based Routing of
different bandwidth constraints for the different classes of service
as defined in [DS-TE], or (b) optimizing on a different metric during
Constraint Based Routing or (c) both. This document discusses how
optimizing on a different metric can be achieved during Constraint
Based Routing.
The most common scenario for a different metric calls for
optimization of a metric reflecting delay (mainly propagation delay)
when Constraint Based Routing TE Label Switched Paths (LSPs) that
will be transporting voice, while optimizing a more usual metric
(e.g., reflecting link bandwidth) when Constraint Based Routing TE
LSPs that will be transporting data.
Additional IGP protocol extensions could be defined so that multiple
TE metrics could be advertised in the IGP (as proposed for example in
[METRICS]) and would thus be available to Constraint Based Routing in
order to optimize on a different metric. However this document
describes how optimizing on a different metric can be achieved today
by existing implementations and deployments, without any additional
IGP extensions beyond [ISIS-TE] and [OSPF-TE], by effectively using
the IGP metric as a "second" TE metric.
2. Common Practice
In current MPLS TE deployments, network administrators often want
Constraint Based Routing of TE LSPs carrying data traffic to be based
on the same metric as the metric used for Shortest Path Routing.
Where this is the case, this practice allows the Constraint Based
Routing algorithm running on the Head-End LSR to use the IGP metric
advertised in the IGP to compute paths for data TE LSPs instead of
the advertised TE metric. The TE metric can then be used to convey
another metric (e.g., a delay-based metric) which can be used by the
Constraint Based Routing algorithm on the Head-End LSR to compute
path for the TE LSPs with different requirements (e.g., Voice TE
LSP).
In some networks, network administrators configure the IGP metric to
a value factoring the link propagation delay. In that case, this
practice allows the Constraint Based Routing algorithm running on the
Head-End LSR to use the IGP metric advertised in the IGP to compute
paths for delay-sensitive TE LSPs (e.g., Voice TE LSPs) instead of
the advertised TE metric. The TE metric can then be used to convey
another metric (e.g., bandwidth based metric) which can be used by
the Constraint Based Routing algorithm to compute paths for the data
TE LSPs.
More generally, the TE metric can be used to carry any arbitrary
metric that may be useful for Constraint Based Routing of the set of
LSPs which need optimization on another metric than the IGP metric.
2.1. Head-End LSR Implementation Practice
A Head-End LSR implements the current practice by:
(i) Allowing configuration, for each TE LSP to be routed, of
whether the IGP metric or the TE metric is to be used by the
Constraint Based Routing algorithm.
(ii) Enabling the Constraint Based Routing algorithm to make use of
either the TE metric or the IGP metric, depending on the above
configuration for the considered TE-LSP
2.2. Network Deployment Practice
A Service Provider deploys this practice by:
(i) Configuring, on every relevant link, the TE metric to reflect
whatever metric is appropriate (e.g., delay-based metric) for
Constraint Based Routing of some LSPs as an alternative metric
to the IGP metric
(ii) Configuring, for every TE LSP, whether this LSP is to be
constraint based routed according to the TE metric or IGP
metric
2.3. Constraints
The practice described in this document has the following
constraints:
(i) it only allows TE tunnels to be routed on either of two metrics
(i.e., it cannot allow TE tunnels to be routed on one of three,
or more, metrics). Extensions (for example such as those
proposed in [METRICS]) could be defined in the future if
necessary to relax this constraints, but this is outside the
scope of this document.
(ii) it can only be used where the IGP metric is appropriate as one
of the two metrics to be used for constraint based routing
(i.e., it cannot allow TE tunnels to be routed on either of two
metrics while allowing IGP SPF to be based on a third metric).
Extensions (for example such as those proposed in [METRICS])
could be defined in the future if necessary to relax this
constraints, but this is outside the scope of this document.
(iii) it can only be used on links which support an IGP adjacency so
that an IGP metric is indeed advertised for the link. For
example, this practice can not be used on Forwarding
Adjacencies (see [LSP-HIER]).
Note that, as with [METRICS], this practice does not recommend that
the TE metric and the IGP metric be used simultaneously during path
computation for a given LSP. This is known to be an NP-complete
problem.
2.4. Interoperability
Where path computation is entirely performed by the Head-End (e.g.,
intra-area operations with path computation on Head-end), this
practice does not raise any interoperability issue among LSRs since
the use of one metric or the other is a matter purely local to the
Head-End LSR.
Where path computation involves another component than the Head-End
(e.g., with inter-area operations where path computation is shared
between the Head-End and Area Boundary Routers or a Path Computation
Server), this practice requires that which metric to optimize on, be
signaled along with the other constraints (bandwidth, affinity) for
the LSP. See [PATH-COMP] for an example proposal on how to signal
which metric to optimize, to another component involved in path
computation when RSVP-TE is used as the protocol to signal path
computation information.
3. Migration Considerations
Service Providers need to consider how to migrate from the current
implementation to the new one supporting this practice.
Although the head-end routers act independently from each other, some
migration scenarios may require that all head-end routers be upgraded
to the new implementation to avoid any disruption on existing TE-LSPs
before two metrics can effectively be used by TE. The reason is that
routers with current implementation are expected to always use the TE
metric for Constraint Based Routing of all tunnels; so when the TE
metric is reconfigured to reflect the "second metric" (say to a
delay-based metric) on links in the network, then all TE-LSPs would
get routed based on the "second metric" metric, while the intent may
be that only the TE-LSPs explicitly configured so should be routed
based on the "second metric".
A possible migration scenario would look like this:
1) upgrade software on all head-end routers in the network to support
this practice.
2) change the TE-LSPs configuration on the head-end routers to use
the IGP metric (e.g., bandwidth-based) for Constraint Based
Routing rather than the TE metric.
3) configure TE metric on the links to reflect the "second metric"
(e.g., delay-based).
4) modify the LSP configuration of the subset of TE-LSPs which need
to be Constraint Based routed using the "second metric" (e.g.,
delay-based), and/or create new TE-LSPs with such a configuration.
It is desirable that step 2 is non-disruptive (i.e., the routing of a
LSP will not be affected in any way, and the data transmission will
not be interrupted) by the change of LSP configuration to use "IGP
metric" as long as the actual value of the "IGP metric" and "TE
metric" are equal on every link at the time of LSP reconfiguration
(as would be the case at step 2 in migration scenario above which
assumed that TE metric was initially equal to IGP metric).
4. Security Considerations
The practice described in this document does not raise specific
security issues beyond those of existing TE. Those are discussed in
the respective security sections of [TE-REQ], [RSVP-TE] and [CR-LDP].
5. Acknowledgment
This document has benefited from discussion with Jean-Philippe
Vasseur.
6. References
6.1. Normative References
[TE-REQ] Awduche, D., Malcolm, J., Agogbua, J., O'Dell, M. and J.
McManus, Requirements for Traffic Engineering over MPLS,
RFC 2702, September 1999.
[OSPF-TE] Katz, D., Kompella, K. and D. Yeung, "Traffic Engineering
(TE) Extensions to OSPF Version 2", RFC 3630, September
2003.
[ISIS-TE] Smit, H. and T. Li, "Intermediate System to Intermediate
System (IS-IS) Extensions for Traffic Engineering (TE),
RFC 3784, May 2004.
[RSVP-TE] Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V.
and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP
Tunnels", RFC 3209, December 2001.
[CR-LDP] Jamoussi, B., Andersson, L., Callon, R., Dantu, R., Wu,
L., Doolan, P., Worster, T., Feldman, N., Fredette, A.,
Girish, M., Gray, E., Heinanen, J., Kilty, T. and A.
Malis, "Constraint-Based LSP Setup using LDP", RFC 3212,
January 2002.
6.1. Informative References
[METRICS] Fedyk, et al., "Multiple Metrics for Traffic Engineering
with IS-IS and OSPF", Work in Progress, November 2000.
[DIFF-TE] Le Faucheur, F. and W. Lai, "Requirements for Support of
Differentiated Services-aware MPLS Traffic Engineering",
RFC 3564, July 2003.
[PATH-COMP] Vasseur, et al., "RSVP Path computation request and reply
messages", Work in Progress, June 2002.
[LSP-HIER] Kompella, et al., "LSP Hierarchy with Generalized MPLS
TE", Work in Progress, September 2002.
7. Authors' Addresses
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
Ramesh Uppili
Cisco Systems,
2000 Innovation Drive
Kanata,
ONTARIO,
Canada - K2K 3E8
Phone: 01-613-254 4578
Email: ruppili@cisco.com
Alain Vedrenne
Equant
Heraklion, 1041 route des Dolines, BP347
06906 Sophia Antipolis Cedex
FRANCE
Phone: +33 4 92 96 57 22
EMail: alain.vedrenne@equant.com
Pierre Merckx
Equant
1041 route des Dolines - BP 347
06906 SOPHIA ANTIPOLIS Cedex
FRANCE
Phone: +33 (0)492 96 6454
EMail: pierre.merckx@equant.com
Thomas Telkamp
Global Crossing, Ltd.
Croeselaan 148
NL-3521CG Utrecht
The Netherlands
Phone: +31 30 238 1250
EMail: telkamp@gblx.net
8. Full Copyright Statement
Copyright (C) The Internet Society (2004). This document is subject
to the rights, licenses and restrictions contained in BCP 78, and
except as set forth therein, the authors retain all their rights.
This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights or other rights that might be claimed to
pertain to the implementation or use of the technology described in
this document or the extent to which any license under such rights
might or might not be available; nor does it represent that it has
made any independent effort to identify any such rights. Information
on the procedures with respect to rights in RFC documents can be
found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat and any
assurances of licenses to be made available, or the result of an
attempt made to obtain a general license or permission for the use of
such proprietary rights by implementers or users of this
specification can be obtained from the IETF on-line IPR repository at
http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at ietf-
ipr@ietf.org.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
AI ELI5
On this page
§1Why a Second Metric Is Wanted
OSPF and IS-IS were extended so that routers can advertise one additional TE metric per link, used by Constraint Based Routing when it computes paths for MPLS traffic engineering tunnels (Label Switched Paths, or LSPs).
Providers sometimes want different classes of service carried on different TE tunnels, for example to apply different fast restoration policies or to meet different QoS objectives. The most common case in the RFC is optimizing a delay-related metric (mainly propagation delay) for tunnels carrying voice, while optimizing a more usual metric, such as one reflecting link bandwidth, for tunnels carrying data.
Further IGP extensions could advertise several TE metrics. This RFC instead describes what existing implementations and deployments already do, with no extension beyond the OSPF and IS-IS TE extensions.
Phone calls care about how fast the road is, and big downloads care about how wide the road is. One ruler cannot measure both, so you want two.
§2The Common Practice
The idea is simple. Data tunnels are often meant to follow the same metric as normal shortest path routing. In that case the head-end router computes their paths with the IGP metric, and the TE metric is freed up to carry some other value, such as a delay-based metric, for the tunnels that need it.
The reverse also works. If administrators set the IGP metric to factor in link propagation delay, delay-sensitive tunnels such as voice use the IGP metric, and the TE metric can carry a bandwidth-based value for data tunnels. More generally, the TE metric can carry any arbitrary metric useful for the tunnels that need something other than the IGP metric.
The road signs already show one cost. The memo says to put a second cost on a different sign and let each trip pick which sign to read.
| Tunnel type | Metric used for path computation | What the other metric carries |
|---|---|---|
| Data LSPs | IGP metric (same as shortest path routing) | TE metric: e.g., a delay-based value for voice LSPs |
| Voice LSPs | IGP metric, if it factors in propagation delay | TE metric: e.g., a bandwidth-based value for data LSPs |
§2.1What Routers and Operators Do
A head-end router implements the practice in two steps: it lets the operator configure, for each TE LSP, whether the IGP metric or the TE metric is used, and its path computation can then use either one according to that setting.
A service provider deploys it by configuring the TE metric on every relevant link to reflect the alternative metric (for example a delay-based one), and by configuring each TE LSP to be routed on either the TE metric or the IGP metric.
Each router gets a switch for every tunnel that says which ruler to use. The network owner sets the second ruler's numbers on every road.
§2.3Limits of the Approach
The RFC lists three constraints:
- Tunnels can be routed on one of two metrics only, not three or more.
- The IGP metric must be appropriate as one of the two metrics. It cannot be used if IGP shortest path routing needs to be based on a third metric.
- It only works on links that support an IGP adjacency, so that an IGP metric is actually advertised. For example, it cannot be used on Forwarding Adjacencies.
The practice also does not use the TE metric and the IGP metric at the same time for one LSP, since that is known to be an NP-complete problem.
You get exactly two rulers, one must be the regular one, and it only works on roads that are really on the map. Using both rulers at once for one trip is far too hard to work out.
§2.4Interoperability
When the head-end router does all the path computation, as in intra-area operation, there is no interoperability issue, because choosing one metric or the other is purely local to that router.
When another component also takes part, for example Area Boundary Routers or a Path Computation Server in inter-area operation, the choice of metric must be signaled along with the other constraints (bandwidth, affinity) for the LSP. The RFC points to an example proposal for doing this with RSVP-TE.
If one router plans the whole trip, nobody else needs to know which ruler it used. If helpers plan parts of the trip, they must be told which ruler to read.
§3Migrating to Two Metrics
Routers with the older behaviour always use the TE metric for all tunnels. If the TE metric is reconfigured to a delay-based value before every head-end is upgraded, all tunnels would be routed on it, not only the intended ones. Some migrations therefore need all head-ends upgraded first.
The RFC gives this sequence:
- Upgrade software on all head-end routers to support the practice.
- Change the TE LSP configuration to use the IGP metric instead of the TE metric.
- Configure the TE metric on the links to reflect the second metric.
- Change the configuration of the LSPs that need the second metric, and create new ones if needed.
Step 2 should not disrupt traffic, as long as the IGP and TE metrics are equal on every link at that moment.
First give every router the new switch, then flip all trips to the regular ruler, then change the second ruler's numbers, and only then move the trips that need it.
§4Security Considerations
The practice raises no specific security issues beyond those of existing traffic engineering. The RFC refers readers to the security sections of the traffic engineering requirements RFC and of the RSVP-TE and CR-LDP specifications.
Using the second ruler does not make the roads any less safe than before.