Network Working Group T. Li
Request for Comments: 2966 Procket Networks
Category: Informational T. Przygienda
Redback
H. Smit
Procket Networks
October 2000
Domain-wide Prefix Distribution with Two-Level IS-IS
Status of this Memo
This memo provides information for the Internet community. It does
not specify an Internet standard of any kind. Distribution of this
memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2000). All Rights Reserved.
Abstract
This document describes extensions to the Intermediate System to
Intermediate System (IS-IS) protocol to support optimal routing
within a two-level domain. The IS-IS protocol is specified in ISO
10589, with extensions for supporting IPv4 (Internet Protocol)
specified in RFC 1195 [2].
This document extends the semantics presented in RFC 1195 so that a
routing domain running with both level 1 and level 2 Intermediate
Systems (IS) [routers] can distribute IP prefixes between level 1 and
level 2 and vice versa. This distribution requires certain
restrictions to insure that persistent forwarding loops do not form.
The goal of this domain-wide prefix distribution is to increase the
granularity of the routing information within the domain.
An IS-IS routing domain (a.k.a., an autonomous system running IS-IS)
can be partitioned into multiple level 1 (L1) areas, and a level 2
(L2) connected subset of the topology that interconnects all of the
L1 areas. Within each L1 area, all routers exchange link state
information. L2 routers also exchange L2 link state information to
compute routes between areas.
Informational [Page 1]
RFC 2966 Domain-wide Prefix Distribution October 2000
RFC 1195 [2] defines the Type, Length and Value (TLV) tuples that are
used to transport IPv4 routing information in IS-IS. RFC 1195 also
specifies the semantics and procedures for interactions between
levels. Specifically, routers in a L1 area will exchange information
within the L1 area. For IP destinations not found in the prefixes in
the L1 database, the L1 router should forward packets to the nearest
router that is in both L1 and L2 (i.e., an L1L2 router) with the
"attached bit" set in its L1 Link State Protocol Data Unit (LSP).
Also per RFC 1195, an L1L2 router should be manually configured with
a set of prefixes that summarizes the IP prefixes reachable in that
L1 area. These summaries are injected into L2. RFC 1195 specifies
no further interactions between L1 and L2 for IPv4 prefixes.
The mechanisms specified in RFC 1195 are appropriate in many
situations, and lead to excellent scalability properties. However,
in certain circumstances, the domain administrator may wish to
sacrifice some amount of scalability and distribute more specific
information than is described by RFC 1195. This section discusses
the various reasons why the domain administrator may wish to make
such a tradeoff.
One major reason for distributing more prefix information is to
improve the quality of the resulting routes. A well know property of
prefix summarization or any abstraction mechanism is that it
necessarily results in a loss of information. This loss of
information in turn results in the computation of a route based upon
less information, which will frequently result in routes that are not
optimal.
A simple example can serve to demonstrate this adequately. Suppose
that a L1 area has two L1L2 routers that both advertise a single
summary of all prefixes within the L1 area. To reach a destination
inside the L1 area, any other L2 router is going to compute the
shortest path to one of the two L1L2 routers for that area. Suppose,
for example, that both of the L1L2 routers are equidistant from the
L2 source, and that the L2 source arbitrarily selects one L1L2
router. This router may not be the optimal router when viewed from
the L1 topology. In fact, it may be the case that the path from the
selected L1L2 router to the destination router may traverse the L1L2
router that was not selected. If more detailed topological
information or more detailed metric information was available to the
L2 source router, it could make a more optimal route computation.
Informational [Page 2]
RFC 2966 Domain-wide Prefix Distribution October 2000
This situation is symmetric in that an L1 router has no information
about prefixes in L2 or within a different L1 area. In using the
nearest L1L2 router, that L1L2 is effectively injecting a default
route without metric information into the L1 area. The route
computation that the L1 router performs is similarly suboptimal.
Besides the optimality of the routes computed, there are two other
significant drivers for the domain wide distribution of prefix
information.
When a router learns multiple possible paths to external destinations
via BGP, it will select only one of those routes to be installed in
the forwarding table. One of the factors in the BGP route selection
is the IGP cost to the BGP next hop address. Many ISP networks
depend on this technique, which is known as "shortest exit routing".
If a L1 router does not know the exact IGP metric to all BGP speakers
in other L1 areas, it cannot do effective shortest exit routing.
The third driver is the current practice of using the IGP (IS-IS)
metric as part of the BGP Multi-Exit Discriminator (MED). The value
in the MED is advertised to other domains and is used to inform other
domains of the optimal entry point into the current domain. Current
practice is to take the IS-IS metric and insert it as the MED value.
This tends to cause external traffic to enter the domain at the point
closest to the exit router. Note that the receiving domain may,
based upon policy, choose to ignore the MED that is advertised.
However, current practice is to distribute the IGP metric in this way
in order to optimize routing wherever possible. This is possible in
current networks that only are a single area, but becomes problematic
if hierarchy is to be installed into the network. This is again
because the loss of end-to-end metric information means that the MED
value will not reflect the true distance across the advertising
domain. Full distribution of prefix information within the domain
would alleviate this problem as it would allow accurate computation
of the IS-IS metric across the domain, resulting in an accurate value
presented in the MED.
The disadvantage to performing the domain-wide prefix distribution
described above is that it has an impact to the scalability of IS-IS.
Areas within IS-IS help scalability in that LSPs are contained within
a single area. This limits the size of the link state database, that
in turn limits the complexity of the shortest path computation.
Informational [Page 3]
RFC 2966 Domain-wide Prefix Distribution October 2000
Further, the summarization of the prefix information aids scalability
in that the abstraction of the prefix information removes the sheer
number of data items to be transported and the number of routes to be
computed.
It should be noted quite strongly that the distribution of prefixes
on a domain wide basis impacts the scalability of IS-IS in the second
respect. It will increase the number of prefixes throughout the
domain. This will result in increased memory consumption,
transmission requirements and computation requirements throughout the
domain.
It must also be noted that the domain-wide distribution of prefixes
has no effect whatsoever on the first aspect of scalability, namely
the existence of areas and the limitation of the distribution of the
link state database.
Thus, the net result is that the introduction of domain-wide prefix
distribution into a formerly flat, single area network is a clear
benefit to the scalability of that network. However, it is a
compromise and does not provide the maximum scalability available
with IS-IS. Domains that choose to make use of this facility should
be aware of the tradeoff that they are making between scalability and
optimality and provision and monitor their networks accordingly.
Normal provisioning guidelines that would apply to a fully
hierarchical deployment of IS-IS will not apply to this type of
configuration.
This document defines the syntax of how to advertise level 2 routes
in level 1 LSPs. The encoding is an extension of the encoding in RFC
1195.
To some extent, in IS-IS the level 2 backbone can be seen as a
separate area itself. RFC 1195 defines that L1L2 routers can
advertise IP routes that were learned via L1 routing into L2. These
routes can be regarded as inter-area routes. RFC 1195 defines that
these L1->L2 inter-area routes must be advertised in L2 LSPs in the
"IP Internal Reachability Information" TLV (TLV 128). Intra-area L2
routes are also advertised in L2 LSPs in an "IP Internal Reachability
Information" TLV. Therefore, L1->L2 inter-area routes are
indistinguishable from L2 intra-area routes.
RFC 1195 does not define L2->L1 inter-area routes. A simple
extension would be to allow a L1L2 router to advertise routes learned
via L2 routing in its L1 LSP. However, to prevent routing-loops,
L1L2 routers must never advertise L2->L1 inter-area routes that they
Informational [Page 4]
RFC 2966 Domain-wide Prefix Distribution October 2000
learn via L1 routing, back into L2. Therefore, there must be a way
to distinguish L2->L1 inter-area routes from L1 intra-area routes.
Draft-ietf-isis-traffic-01.txt defines the "up/down bit" for this
purpose. RFC 1195 defines TLVs 128 and 130 to contain IP routes.
TVLs 128 and 130 have a metric field that consists of 4 TOS metrics.
The first metric, the so-called "default metric", has the high-order
bit reserved (bit 8). Routers must set this bit to zero on
transmission, and ignore it on receipt.
This document redefines this high-order bit in the default metric
field in TLVs 128 and 130 to be the up/down bit. L1L2 routers must
set this bit to one for prefixes that are derived from L2 routing and
are advertised into L1 LSPs. The bit must be set to zero for all
other IP prefixes in L1 or L2 LSPs. Prefixes with the up/down bit
set that are learned via L1 routing, must never be advertised by L1L2
routers back into L2.
RFC 1195 defines two TLVs for carrying IP prefixes. TLV 128 is
defined as "IP Internal Reachability Information", and should be used
to carry IP prefixes that are directly connected to IS-IS routers.
TLV 130 is defined as "IP External Reachability Information", and
should be used to carry routes learned from outside the IS-IS domain.
RFC 1195 documents TLV type 130 only for level 2 LSPs.
RFC 1195 also defines two types of metrics. Metrics of the internal
metric-type should be used when the metric is comparable to metrics
used to weigh links inside the ISIS domain. Metrics of the external
metric-type should be used if the metric of an IP prefix cannot be
directly compared to internal metrics. External metric-type can only
be used for external IP prefixes. A direct result is that metrics of
external metric-type should never be seen in TLV 128.
To prevent confusion, this document states again that when a router
computes IP routes, it must give the same preference to IP routes
advertised in an "IP Internal Reachability Information" TLV and IP
routes advertised in an "IP External Reachability Information" TLV.
RFC 1195 states this quite clearly in the note in paragraph 3.10.2,
item 2c). This document does not alter this rule of preference.
NOTE: Internal routes (routes to destinations announced in the
"IP Internal Reachability Information" field), and external
routes using internal metrics (routes to destinations announced
in the "IP External Reachability Information" field, with a
metric of type "internal") are treated identically for the
purpose of the order of preference of routes, and the Dijkstra
calculation.
Informational [Page 5]
RFC 2966 Domain-wide Prefix Distribution October 2000
However, IP routes advertised in "IP External Reachability
Information" with external metric-type must be given less preference
than the same IP routes advertised with internal-metric type,
regardless of the value of the metrics.
While IS-IS routers must not give different preference to IP prefixes
learned via "IP Internal Reachability Information" and "IP External
Reachability Information" when executing the Dijkstra calculation,
routers that implement multiple IGPs are free to use this distinction
between internal and external routes when comparing routes derived
from different IGPs for inclusion in their global RIB.
RFC 1195 does not define the "IP External Reachability Information"
TLV for L1 LSPs. However, there is no reason why an IS-IS
implementation could not allow for redistribution of external routes
into L1. Some IS-IS implementations already allow network
administrators to do this. This document loosens the restrictions in
RFC 1195, and allows for the inclusion of the "IP External
Reachability Information" TLV in L1 LSPs.
RFC 1195 defines that IP routes learned via L1 routing must always be
advertised in L2 LSPs in a "IP Internal Reachability Information"
TLV. Now that this document allows "IP External Reachability
Information" TLVs in L1 LSPs, and allows for the advertisement of
routes learned via L2 routing into L1, the above rule needs a
extensions.
When a L1L2 router advertises a L1 route into L2, where that L1 route
was learned via a prefix advertised in a "IP External Reachability
Information" TLV, that L1L2 router should advertise that prefix in
its L2 LSP within an "IP External Reachability Information" TLV. L1
routes learned via an "IP Internal Reachability Information" TLV
should still be advertised within a "IP Internal Reachability
Information" TLV. These rules should also be applied when
advertising IP routes derived from L2 routing into L1. Of course in
this case also the up/down bit must be set.
RFC 1195 defines that if a router sees the same external prefix
advertised by two or more routers with the same external metric, it
must select the route that is advertised by the router that is
closest to itself. It should be noted that now that external routes
can be advertised from L1 into L2, and vice versa, that the router
that advertises an external prefix in its LSP might not be the router
that originally injected this prefix into the IS-IS domain.
Therefore, it is less useful to advertise external routes with
external metrics into other levels.
Informational [Page 6]
RFC 2966 Domain-wide Prefix Distribution October 2000
RFC 1195 and this document defines several ways of advertising IP
routes in IS-IS. There are four variables involved.
1) The level of the LSP in which the route is advertised. There are
currently two possible values: level 1 and level 2
2) The route-type, which can be derived from the type of TLV in which
the prefix is advertised. Internal routes are advertised in IP
Internal Reachability Information TLVs (TLV 128), and external
routes are advertised in IP External Reachability Information TLVs
(TLV 130).
3) The metric-type: Internal or External. The metric-type is derived
from the Internal/External metric-type bit in the metric field
(bit 7).
4) The fact whether this route is leaked down in the hierarchy, and
thus can not be advertised back up. This information can be
derived from the newly defined up/down bit in the default metric
field.
The combination IP Internal Reachability Information and external
metric-type is not allowed. Also the up/down bit is never set in L2
LSPs. This leaves us with 8 different types of IP advertisements in
IS-IS. However, there are more than 8 reasons for IP prefixes to be
advertised in IS-IS. The following tables describe the types of IP
prefixes and how they are encoded.
1) L1 intra-area routes
These are advertised in L1 LSPs, in TLV 128.
The up/down bit is set to zero, metric-type is internal metric.
These IP prefixes are directly connected to the advertising router.
2) L1 external routes
These are advertised in L1 LSPs, in TLV 130.
The up/down bit is set to zero, metric-type is internal metric.
These IP prefixes are learned from other IGPs, and are usually not
directly connected to the advertising router.
Informational [Page 7]
RFC 2966 Domain-wide Prefix Distribution October 2000
3) L2 intra-area routes
These are advertised in L2 LSPs, in TLV 128.
The up/down bit is set to zero, metric-type is internal metric.
These IP prefixes are directly connected to the advertising router.
These prefixes can not be distinguished from L1->L2 inter-area
routes.
4) L2 external routes
These are advertised in L2 LSPs, in TLV 130.
The up/down bit is set to zero, metric-type is internal metric.
These IP prefixes are learned from other IGPs, and are usually not
directly connected to the advertising router. These prefixes can
not be distinguished from L1->L2 inter-area external routes.
5) L1->L2 inter-area routes
These are advertised in L2 LSPs, in TLV 128.
The up/down bit is set to zero, metric-type is internal metric.
These IP prefixes are learned via L1 routing, and were derived
during the L1 SPF computation from prefixes advertised in L1 LSPs in
TLV 128. These prefixes can not be distinguished from L2 intra-area
routes.
6) L1->L2 inter-area external routes
These are advertised in L2 LSPs, in TLV 130.
The up/down bit is set to zero, metric-type is internal metric.
These IP prefixes are learned via L1 routing, and were derived
during the L1 SPF computation from prefixes advertised in L1 LSPs in
TLV 130. These prefixes can not be distinguished from L2 external
routes.
7) L2->L1 inter-area routes
These are advertised in L1 LSPs, in TLV 128.
The up/down bit is set to one, metric-type is internal metric.
These IP prefixes are learned via L2 routing, and were derived
during the L2 SPF computation from prefixes advertised in TLV 128.
8) L2->L1 inter-area external routes
These are advertised in L1 LSPs, in TLV 130.
The up/down bit is set to one, metric-type is internal metric.
These IP prefixes are learned via L2 routing, and were derived
during the L2 SPF computation from prefixes advertised in L2 LSPs in
TLV 130.
Informational [Page 8]
RFC 2966 Domain-wide Prefix Distribution October 2000
9) L1 external routes with external metric
These are advertised in L1 LSPs, in TLV 130.
The up/down bit is set to zero, metric-type is external metric.
These IP prefixes are learned from other IGPs, and are usually not
directly connected to the advertising router.
10) L2 external routes with external metric
These are advertised in L2 LSPs, in TLV 130.
The up/down bit is set to zero, metric-type is external metric.
These IP prefixes are learned from other IGPs, and are usually not
directly connected to the advertising router. These prefixes can
not be distinguished from L1->L2 inter-area external routes with
external metric.
11) L1->L2 inter-area external routes with external metric
These are advertised in L2 LSPs, in TLV 130.
The up/down bit is set to zero, metric-type is external metric.
These IP prefixes are learned via L1 routing, and were derived
during the L1 SPF computation from prefixes advertised in L1 LSPs in
TLV 130 with external metrics. These prefixes can not be
distinguished from L2 external routes with external metric.
12) L2->L1 inter-area external routes with external metric
These are advertised in L1 LSPs, in TLV 130.
The up/down bit is set to one, metric-type is external metric.
These IP prefixes are learned via L2 routing, and were derived
during the L1 SPF computation from prefixes advertised in L2 LSPs in
TLV 130 with external metrics.
Unfortunately IS-IS cannot depend on metrics alone for route
selection. Some types of routes must always preferred over others,
regardless of the costs that were computed in the Dijkstra
calculation. One of the reasons for this is that inter-area routes
can only be advertised with a maximum metric of 63. Another reason
is that this maximum value of 63 does not mean infinity (e.g. like a
hop count of 16 in RIP denotes unreachable). Introducing a value for
infinity cost in IS-IS inter-area routes would introduce counting-
to-infinity behavior via two or more L1L2 routers, which would have a
bad impact on network stability.
The order of preference of IP routes in IS-IS is based on a few
assumptions.
Informational [Page 9]
RFC 2966 Domain-wide Prefix Distribution October 2000
- RFC 1195 defines that routes derived from L1 routing are preferred
over routes derived from L2 routing.
- The note in RFC 1195 paragraph 3.10.2, item 2c) defines that
internal routes with internal metric-type and external prefixes
with internal metric-type have the same preference.
- RFC 1195 defines that external routes with internal metric-type are
preferred over external routes with external metric type.
- Routes derived from L2 routing are preferred over L2->L1 routes
derived from L1 routing.
Based on these assumptions, this document defines the following route
preferences.
1) L1 intra-area routes with internal metric
L1 external routes with internal metric
2) L2 intra-area routes with internal metric
L2 external routes with internal metric
L1->L2 inter-area routes with internal metric
L1->L2 inter-area external routes with internal metric
3) L2->L1 inter-area routes with internal metric
L2->L1 inter-area external routes with internal metric
4) L1 external routes with external metric
5) L2 external routes with external metric
L1->L2 inter-area external routes with external metric
6) L2->L1 inter-area external routes with external metric
Paragraphs 4.1 and 4.2 enumerate all used IP route types in IS-IS.
Besides these defined route types, the encoding used would allow for
a few more potential combinations. One of them is the combination of
"IP Internal Reachability Information" and external metric type.
This combination should never be used when building an LSP. Upon
receipt of an IP prefix with this combination, routers must ignore
this prefix.
Another issue would be the usage of the up/down bit in L2 LSPs.
Because IS-IS is currently defined with two levels of hierarchy,
there should never be a need to set the up/down bit in L2 LSPs.
However, if IS-IS would ever be extended with more than two levels of
hierarchy, L2-only (or L1L2) routers will need to be able to accept
L2 IP routes with the up/down bit set. Therefore, it is recommended
that implementations ignore the up/down bit in L2 LSPs, and accept
the prefixes in L2 LSPs regardless whether the up/down bit is set.
This will allow for simpler migration once more than two levels of
hierarchy are defined.
Informational [Page 10]
RFC 2966 Domain-wide Prefix Distribution October 2000
Another detail that implementors should be aware of is the fact that
L1L2 routers should only advertise in their L2 LSP those L1 routes
that they use for forwarding themselves. They should not
unconditionally advertise into L2 all prefixes from LSPs in the L1
database.
Not all prefixes need to be advertised up or down the hierarchy.
Implementations might allow for additional manual filtering or
summarization to further bring down the number of inter-area prefixes
they advertise in their LSPs. It is also recommended that the
default configuration of L1L2 routers is to not advertise any L2
routes into L1 (see also paragraph 5.0).
The solution in this document is not fully compatible with RFC 1195.
It is an extension to RFC 1195. If routers do not use the new
functionality of external L1 routes, nor L2->L1 inter-area routes,
older implementations that strictly follow RFC 1195 will be
compatible with newer implementations that follow this document.
Implementations that do not accept the "IP External Reachability
Information" TLV in L1 LSPs will not be able to compute external L1
routes. This could cause routing loops between L1-only routers that
do understand external L1 routes for a particular destination, and
L1-only routers that use the default route pointing the closest
attached L1L2 router for that destination.
Implementations that follow RFC 1195 should ignore bit 8 in the
default metric field when computing routes. Therefore, even older
implementations that do not know of the up/down bit should be able to
accept the new L2->L1 inter-area routes. These older implementations
will install the new L2->L1 inter-area routes as L1 intra-area
routes, but that in itself does not cause routing loops among L1-only
routers.
However, it is vital that the up/down bit is recognized by L1L2
routers. As has been stated before, L1L2 routers must never
advertise L2->L1 inter-area routes back into L2. Therefore, if L2
routes are advertised down into L1 area, it is required that all L1L2
routers in that area run software that understands the new up/down
bit. Older implementations that follow RFC 1195 and do not
understand the new up/down bit will threat the L2->L1 inter-area
routes as L1 intra-area routes, and they will advertise these routes
back into L2. This can cause routing loops, sub-optimal routing or
extra routing instability. For this reason it is recommended that
Informational [Page 11]
RFC 2966 Domain-wide Prefix Distribution October 2000
implementations by default do not advertise any L2 routes into L1.
Implementations should force the network administrator to manually
configure L1L2 routers to advertise any L2 routes into L1.
In [3], a new TLV is defined to transport IP prefix information.
This TLV format also defines an up/down bit to allow for L2->L1
inter-area routes. [3] also defines a new TLV to describe links.
Both TLVs have wider metric space, and have the possibility to define
sub-TLVs to advertise extra information belonging to the link or
prefix. The wider metric space in IP prefix TLVs allows for more
granular metric information about inter-area path costs. To make
full use of the wider metric space, network administrators must
deploy both new TLVs at the same time.
Deployment of [3] requires an upgrade of all routers in the network
and a transition to the new TLVs. Such a network-wide upgrade and
transition might not be an easy task. In this case, the solution
defined in this document, which requires only an upgrade of L1L2
routers in selected areas, might be a good alternative to the
solution defined in [3].
[1] ISO 10589, "Intermediate System to Intermediate System Intra-
Domain Routing Exchange Protocol for use in Conjunction with the
Protocol for Providing the Connectionless-mode Network Service
(ISO 8473)". [Also republished as RFC 1142.]
[2] Callon, R., "Use of OSI IS-IS for routing in TCP/IP and dual
environments", RFC 1195, December 1990.
[3] Smit, H. and T. Li, "IS-IS Extensions for Traffic Engineering",
Work in Progress.
Informational [Page 12]
RFC 2966 Domain-wide Prefix Distribution October 2000
Tony Li
Procket Networks
1100 Cadillac Court
Milpitas, CA 95035-3025
EMail: tli@procket.com
Tony Przygienda
Redback
350 Holger Way
San Jose, CA 95134
EMail: prz@redback.com
Henk Smit
Procket Networks
1100 Cadillac Court
Milpitas, CA 95035-3025
EMail: henk@procket.com
Informational [Page 13]
RFC 2966 Domain-wide Prefix Distribution October 2000
Copyright (C) The Internet Society (2000). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS 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.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
Informational [Page 14]