RFC 2740
OSPF for IPv6
3.4. Link State Advertisements
For IPv6, the OSPF LSA header has changed slightly, with the LS type field expanding and the Options field being moved into the body of appropriate LSAs. Also, the formats of some LSAs have changed somewhat (namely router-LSAs, network-LSAs and AS-external-LSAs), while the names of other LSAs have been changed (type 3 and 4 summary-LSAs are now inter-area-prefix-LSAs and inter-area-router-
LSAs respectively) and additional LSAs have been added (Link-LSAs and Intra-Area-Prefix-LSAs). Type of Service (TOS) has been removed from the OSPFv2 specification [Ref1], and is not encoded within OSPF for IPv6's LSAs. These changes will be described in detail in the following subsections.3.4.1. The LSA Header
In both IPv4 and IPv6, all OSPF LSAs begin with a standard 20 byte LSA header. However, the contents of this 20 byte header have changed in IPv6. The LS age, Advertising Router, LS Sequence Number, LS checksum and length fields within the LSA header remain unchanged, as documented in Sections 12.1.1, 12.1.5, 12.1.6, 12.1.7 and A.4.1 of [Ref1] respectively. However, the following fields have changed for IPv6: Options The Options field has been removed from the standard 20 byte LSA header, and into the body of router-LSAs, network-LSAs, inter- area-router-LSAs and link-LSAs. The size of the Options field has increased from 8 to 24 bits, and some of the bit definitions have changed (see Section A.2). In addition a separate PrefixOptions field, 8 bits in length, is attached to each prefix advertised within the body of an LSA. LS type The size of the LS type field has increased from 8 to 16 bits, with the top two bits encoding flooding scope and the next bit encoding the handling of unknown LS types. See Section A.4.2.1 for the current coding of the LS type field. Link State ID Link State ID remains at 32 bits in length, but except for network-LSAs and link-LSAs, Link State ID has shed any addressing semantics. For example, an IPv6 router originating multiple AS- external-LSAs could start by assigning the first a Link State ID of 0.0.0.1, the second a Link State ID of 0.0.0.2, and so on. Instead of the IPv4 behavior of encoding the network number within the AS-external-LSA's Link State ID, the IPv6 Link State ID simply serves as a way to differentiate multiple LSAs originated by the same router. For network-LSAs, the Link State ID is set to the Designated Router's Interface ID on the link. When a router originates a Link-LSA for a given link, its Link State ID is set equal to the router's Interface ID on the link.
3.4.2. The link-state database
In IPv6, as in IPv4, individual LSAs are identified by a combination of their LS type, Link State ID and Advertising Router fields. Given two instances of an LSA, the most recent instance is determined by examining the LSAs' LS Sequence Number, using LS checksum and LS age as tiebreakers (see Section 13.1 of [Ref1]). In IPv6, the link-state database is split across three separate data structures. LSAs with AS flooding scope are contained within the top-level OSPF data structure (see Section 3.1) as long as either their LS type is known or their U-bit is 1 (flood even when unrecognized); this includes the AS-external-LSAs. LSAs with area flooding scope are contained within the appropriate area structure (see Section 3.1.1) as long as either their LS type is known or their U-bit is 1 (flood even when unrecognized); this includes router-LSAs, network-LSAs, inter-area-prefix-LSAs, inter-area-router-LSAs, and intra-area-prefix-LSAs. LSAs with unknown LS type and U-bit set to 0 and/or link-local flooding scope are contained within the appropriate interface structure (see Section 3.1.2); this includes link-LSAs. To lookup or install an LSA in the database, you first examine the LS type and the LSA's context (i.e., to which area or link does the LSA belong). This information allows you to find the correct list of LSAs, all of the same LS type, where you then search based on the LSA's Link State ID and Advertising Router.3.4.3. Originating LSAs
The process of reoriginating an LSA in IPv6 is the same as in IPv4: the LSA's LS sequence number is incremented, its LS age is set to 0, its LS checksum is calculated, and the LSA is added to the link state database and flooded out the appropriate interfaces. To the list of events causing LSAs to be reoriginated, which for IPv4 is given in Section 12.4 of [Ref1], the following events and/or actions are added for IPv6: o The state of one of the router's interfaces changes. The router may need to (re)originate or flush its Link-LSA and one or more router-LSAs and/or intra-area-prefix-LSAs. o The identity of a link's Designated Router changes. The router may need to (re)originate or flush the link's network-LSA and one or more router-LSAs and/or intra-area-prefix-LSAs.
o A neighbor transitions to/from "Full" state. The router may need
to (re)originate or flush the link's network-LSA and one or more
router-LSAs and/or intra-area-prefix-LSAs.
o The Interface ID of a neighbor changes. This may cause a new
instance of a router-LSA to be originated for the associated area,
and the reorigination of one or more intra-area-prefix-LSAs.
o A new prefix is added to an attached link, or a prefix is deleted
(both through configuration). This causes the router to
reoriginate its link-LSA for the link, or, if it is the only
router attached to the link, causes the router to reoriginate an
intra-area-prefix-LSA.
o A new link-LSA is received, causing the link's collection of
prefixes to change. If the router is Designated Router for the
link, it originates a new intra-area-prefix-LSA.
Detailed construction of the seven required IPv6 LSA types is
supplied by the following subsections. In order to display example
LSAs, the network map in Figure 15 of [Ref1] has been reworked to
show IPv6 addressing, resulting in Figure 1. The OSPF cost of each
interface is has been displayed in Figure 1. The assignment of IPv6
prefixes to network links is shown in Table 1. A single area address
range has been configured for Area 1, so that outside of Area 1 all
of its prefixes are covered by a single route to 5f00:0000:c001::/48.
The OSPF interface IDs and the link-local addresses for the router
interfaces in Figure 1 are given in Table 2.
..........................................
. Area 1.
. + .
. | .
. | 3+---+1 .
. N1 |--|RT1|-----+ .
. | +---+ \ .
. | \ ______ .
. + \/ \ 1+---+
. * N3 *------|RT4|------
. + /\_______/ +---+
. | / | .
. | 3+---+1 / | .
. N2 |--|RT2|-----+ 1| .
. | +---+ +---+ .
. | |RT3|----------------
. + +---+ .
. |2 .
. | .
. +------------+ .
. N4 .
..........................................
Figure 1: Area 1 with IP addresses shown
Network IPv6 prefix
-----------------------------------
N1 5f00:0000:c001:0200::/56
N2 5f00:0000:c001:0300::/56
N3 5f00:0000:c001:0100::/56
N4 5f00:0000:c001:0400::/56
Table 1: IPv6 link prefixes for sample network
Router interface Interface ID link-local address
-------------------------------------------------------
RT1 to N1 1 fe80:0001::RT1
to N3 2 fe80:0002::RT1
RT2 to N2 1 fe80:0001::RT2
to N3 2 fe80:0002::RT2
RT3 to N3 1 fe80:0001::RT3
to N4 2 fe80:0002::RT3
RT4 to N3 1 fe80:0001::RT4
Table 2: OSPF Interface IDs and link-local addresses
3.4.3.1. Router-LSAs
The LS type of a router-LSA is set to the value 0x2001. Router-LSAs have area flooding scope. A router may originate one or more router- LSAs for a given area. Each router-LSA contains an integral number of interface descriptions; taken together, the collection of router-LSAs originated by the router for an area describes the collected states of all the router's interfaces to the area. When multiple router-LSAs are used, they are distinguished by their Link State ID fields. The Options field in the router-LSA should be coded as follows. The V6-bit should be set. The E-bit should be clear if and only if the attached area is an OSPF stub area. The MC-bit should be set if and only if the router is running MOSPF (see [Ref8]). The N-bit should be set if and only if the attached area is an OSPF NSSA area. The R-bit should be set. The DC-bit should be set if and only if the router can correctly process the DoNotAge bit when it appears in the LS age field of LSAs (see [Ref11]). All unrecognized bits in the Options field should be cleared To the left of the Options field, the router capability bits V, E and B should be coded according to Section 12.4.1 of [Ref1]. Bit W should be coded according to [Ref8]. Each of the router's interfaces to the area are then described by appending "link descriptions" to the router-LSA. Each link description is 16 bytes long, consisting of 5 fields: (link) Type, Metric, Interface ID, Neighbor Interface ID and Neighbor Router ID (see Section A.4.3). Interfaces in state "Down" or "Loopback" are not described (although looped back interfaces can contribute prefixes to Intra-Area-Prefix-LSAs). Nor are interfaces without any full adjacencies described. All other interfaces to the area add zero, one or more link descriptions, the number and content of which depend on the interface type. Within each link description, the Metric field is always set the interface's output cost and the Interface ID field is set to the interface's OSPF Interface ID. Point-to-point interfaces If the neighboring router is fully adjacent, add a Type 1 link description (point-to-point). The Neighbor Interface ID field is set to the Interface ID advertised by the neighbor in its Hello packets, and the Neighbor Router ID field is set to the neighbor's Router ID.
Broadcast and NBMA interfaces
If the router is fully adjacent to the link's Designated Router,
or if the router itself is Designated Router and is fully adjacent
to at least one other router, add a single Type 2 link description
(transit network). The Neighbor Interface ID field is set to the
Interface ID advertised by the Designated Router in its Hello
packets, and the Neighbor Router ID field is set to the Designated
Router's Router ID.
Virtual links
If the neighboring router is fully adjacent, add a Type 4 link
description (virtual). The Neighbor Interface ID field is set to
the Interface ID advertised by the neighbor in its Hello packets,
and the Neighbor Router ID field is set to the neighbor's Router
ID. Note that the output cost of a virtual link is calculated
during the routing table calculation (see Section 3.7).
Point-to-MultiPoint interfaces
For each fully adjacent neighbor associated with the interface,
add a separate Type 1 link description (point-to-point) with
Neighbor Interface ID field set to the Interface ID advertised by
the neighbor in its Hello packets, and Neighbor Router ID field
set to the neighbor's Router ID.
As an example, consider the router-LSA that router RT3 would
originate for Area 1 in Figure 1. Only a single interface must be
described, namely that which connects to the transit network N3. It
assumes that RT4 has been elected Designated Router of Network N3.
; RT3's router-LSA for Area 1
LS age = 0 ;newly (re)originated
LS type = 0x2001 ;router-LSA
Link State ID = 0 ;first fragment
Advertising Router = 192.1.1.3 ;RT3's Router ID
bit E = 0 ;not an AS boundary router
bit B = 1 ;area border router
Options = (V6-bit|E-bit|R-bit)
Type = 2 ;connects to N3
Metric = 1 ;cost to N3
Interface ID = 1 ;RT3's Interface ID on N3
Neighbor Interface ID = 1 ;RT4's Interface ID on N3
Neighbor Router ID = 192.1.1.4 ; RT4's Router ID
If for example another router was added to Network N4, RT3 would have
to advertise a second link description for its connection to (the now
transit) network N4. This could be accomplished by reoriginating the
above router-LSA, this time with two link descriptions. Or, a
separate router-LSA could be originated with a separate Link State ID (e.g., using a Link State ID of 1) to describe the connection to N4. Host routes no longer appear in the router-LSA, but are instead included in intra-area-prefix-LSAs.3.4.3.2. Network-LSAs
The LS type of a network-LSA is set to the value 0x2002. Network- LSAs have area flooding scope. A network-LSA is originated for every broadcast or NBMA link having two or more attached routers, by the link's Designated Router. The network-LSA lists all routers attached to the link. The procedure for originating network-LSAs in IPv6 is the same as the IPv4 procedure documented in Section 12.4.2 of [Ref1], with the following exceptions: o An IPv6 network-LSA's Link State ID is set to the Interface ID of the Designated Router on the link. o IPv6 network-LSAs do not contain a Network Mask. All addressing information formerly contained in the IPv4 network-LSA has now been consigned to intra-Area-Prefix-LSAs. o The Options field in the network-LSA is set to the logical OR of the Options fields contained within the link's associated link- LSAs. In this way, the network link exhibits a capability when at least one of the link's routers requests that the capability be asserted. As an example, assuming that Router RT4 has been elected Designated Router of Network N3 in Figure 1, the following network-LSA is originated: ; Network-LSA for Network N3 LS age = 0 ;newly (re)originated LS type = 0x2002 ;network-LSA Link State ID = 1 ;RT4's Interface ID on N3 Advertising Router = 192.1.1.4 ;RT4's Router ID Options = (V6-bit|E-bit|R-bit) Attached Router = 192.1.1.4 ;Router ID Attached Router = 192.1.1.1 ;Router ID Attached Router = 192.1.1.2 ;Router ID Attached Router = 192.1.1.3 ;Router ID
3.4.3.3. Inter-Area-Prefix-LSAs
The LS type of an inter-area-prefix-LSA is set to the value 0x2003. Inter-area-prefix-LSAs have area flooding scope. In IPv4, inter- area-prefix-LSAs were called type 3 summary-LSAs. Each inter-area- prefix-LSA describes a prefix external to the area, yet internal to the Autonomous System. The procedure for originating inter-area-prefix-LSAs in IPv6 is the same as the IPv4 procedure documented in Sections 12.4.3 and 12.4.3.1 of [Ref1], with the following exceptions: o The Link State ID of an inter-area-prefix-LSA has lost all of its addressing semantics, and instead simply serves to distinguish multiple inter-area-prefix-LSAs that are originated by the same router. o The prefix is described by the PrefixLength, PrefixOptions and Address Prefix fields embedded within the LSA body. Network Mask is no longer specified. o The NU-bit in the PrefixOptions field should be clear. The coding of the MC-bit depends upon whether, and if so how, MOSPF is operating in the routing domain (see [Ref8]). o Link-local addresses must never be advertised in inter-area- prefix-LSAs. As an example, the following shows the inter-area-prefix-LSA that Router RT4 originates into the OSPF backbone area, condensing all of Area 1's prefixes into the single prefix 5f00:0000:c001::/48. The cost is set to 4, which is the maximum cost to all of the prefix' individual components. The prefix is padded out to an even number of 32-bit words, so that it consumes 64-bits of space instead of 48 bits. ; Inter-area-prefix-LSA for Area 1 addresses ; originated by Router RT4 into the backbone LS age = 0 ;newly (re)originated LS type = 0x2003 ;inter-area-prefix-LSA Advertising Router = 192.1.1.4 ;RT4's ID Metric = 4 ;maximum to components PrefixLength = 48 PrefixOptions = 0 Address Prefix = 5f00:0000:c001 ;padded to 64-bits
3.4.3.4. Inter-Area-Router-LSAs
The LS type of an inter-area-router-LSA is set to the value 0x2004. Inter-area-router-LSAs have area flooding scope. In IPv4, inter-area-router-LSAs were called type 4 summary-LSAs. Each inter-area-router-LSA describes a path to a destination OSPF router (an ASBR) that is external to the area, yet internal to the Autonomous System. The procedure for originating inter-area-router-LSAs in IPv6 is the same as the IPv4 procedure documented in Section 12.4.3 of [Ref1], with the following exceptions: o The Link State ID of an inter-area-router-LSA is no longer the destination router's OSPF Router ID, but instead simply serves to distinguish multiple inter-area-router-LSAs that are originated by the same router. The destination router's Router ID is now found in the body of the LSA. o The Options field in an inter-area-router-LSA should be set equal to the Options field contained in the destination router's own router-LSA. The Options field thus describes the capabilities supported by the destination router. As an example, consider the OSPF Autonomous System depicted in Figure 6 of [Ref1]. Router RT4 would originate into Area 1 the following inter-area-router-LSA for destination router RT7. ; inter-area-router-LSA for AS boundary router RT7 ; originated by Router RT4 into Area 1 LS age = 0 ;newly (re)originated LS type = 0x2004 ;inter-area-router-LSA Advertising Router = 192.1.1.4 ;RT4's ID Options = (V6-bit|E-bit|R-bit) ;RT7's capabilities Metric = 14 ;cost to RT7 Destination Router ID = Router RT7's ID3.4.3.5. AS-external-LSAs
The LS type of an AS-external-LSA is set to the value 0x4005. AS- external-LSAs have AS flooding scope. Each AS-external-LSA describes a path to a prefix external to the Autonomous System. The procedure for originating AS-external-LSAs in IPv6 is the same as the IPv4 procedure documented in Section 12.4.4 of [Ref1], with the following exceptions:
o The Link State ID of an AS-external-LSA has lost all of its
addressing semantics, and instead simply serves to distinguish
multiple AS-external-LSAs that are originated by the same router.
o The prefix is described by the PrefixLength, PrefixOptions and
Address Prefix fields embedded within the LSA body. Network Mask
is no longer specified.
o The NU-bit in the PrefixOptions field should be clear. The coding
of the MC-bit depends upon whether, and if so how, MOSPF is
operating in the routing domain (see [Ref8]).
o Link-local addresses can never be advertised in AS-external-LSAs.
o The forwarding address is present in the AS-external-LSA if and
only if the AS-external-LSA's bit F is set.
o The external route tag is present in the AS-external-LSA if and
only if the AS-external-LSA's bit T is set.
o The capability for an AS-external-LSA to reference another LSA has
been included, by inclusion of the Referenced LS Type field and
the optional Referenced Link State ID field (the latter present if
and only if Referenced LS Type is non-zero). This capability is
for future use; for now Referenced LS Type should be set to 0 and
received non-zero values for this field should be ignored.
As an example, consider the OSPF Autonomous System depicted in Figure
6 of [Ref1]. Assume that RT7 has learned its route to N12 via BGP,
and that it wishes to advertise a Type 2 metric into the AS. Further
assume the the IPv6 prefix for N12 is the value 5f00:0000:0a00::/40.
RT7 would then originate the following AS-external-LSA for the
external network N12. Note that within the AS-external-LSA, N12's
prefix occupies 64 bits of space, to maintain 32-bit alignment.
; AS-external-LSA for Network N12,
; originated by Router RT7
LS age = 0 ;newly (re)originated
LS type = 0x4005 ;AS-external-LSA
Link State ID = 123 ;or something else
Advertising Router = Router RT7's ID
bit E = 1 ;Type 2 metric
bit F = 0 ;no forwarding address
bit T = 1 ;external route tag included
Metric = 2
PrefixLength = 40
PrefixOptions = 0
Referenced LS Type = 0 ;no Referenced Link State ID
Address Prefix = 5f00:0000:0a00 ;padded to 64-bits
External Route Tag = as per BGP/OSPF interaction
3.4.3.6. Link-LSAs
The LS type of a Link-LSA is set to the value 0x0008. Link-LSAs have
link-local flooding scope. A router originates a separate Link-LSA
for each attached link that supports 2 or more (including the
originating router itself) routers.
Link-LSAs have three purposes: 1) they provide the router's link-
local address to all other routers attached to the link and 2) they
inform other routers attached to the link of a list of IPv6 prefixes
to associate with the link and 3) they allow the router to assert a
collection of Options bits in the Network-LSA that will be originated
for the link.
A Link-LSA for a given Link L is built in the following fashion:
o The Link State ID is set to the router's Interface ID on Link L.
o The Router Priority of the router's interface to Link L is
inserted into the Link-LSA.
o The Link-LSA's Options field is set to those bits that the router
wishes set in Link L's Network LSA.
o The router inserts its link-local address on Link L into the
Link-LSA. This information will be used when the other routers on
Link L do their next hop calculations (see Section 3.8.1.1).
o Each IPv6 address prefix that has been configured into the router
for Link L is added to the Link-LSA, by specifying values for
PrefixLength, PrefixOptions, and Address Prefix fields.
After building a Link-LSA for a given link, the router installs the
link-LSA into the associate interface data structure and floods the
Link-LSA onto the link. All other routers on the link will receive
the Link-LSA, but it will go no further.
As an example, consider the Link-LSA that RT3 will build for N3 in
Figure 1. Suppose that the prefix 5f00:0000:c001:0100::/56 has been
configured within RT3 for N3. This will give rise to the following
Link-LSA, which RT3 will flood onto N3, but nowhere else. Note that
not all routers on N3 need be configured with the prefix; those not
configured will learn the prefix when receiving RT3's Link-LSA.
; RT3's Link-LSA for N3
LS age = 0 ;newly (re)originated
LS type = 0x0008 ;Link-LSA
Link State ID = 1 ;RT3's Interface ID on N3
Advertising Router = 192.1.1.3 ;RT3's Router ID
Rtr Pri = 1 ;RT3's N3 Router Priority
Options = (V6-bit|E-bit|R-bit)
Link-local Interface Address = fe80:0001::RT3
# prefixes = 1
PrefixLength = 56
PrefixOptions = 0
Address Prefix = 5f00:0000:c001:0100 ;pad to 64-bits
3.4.3.7. Intra-Area-Prefix-LSAs
The LS type of an intra-area-prefix-LSA is set to the value 0x2009.
Intra-area-prefix-LSAs have area flooding scope. An intra-area-
prefix-LSA has one of two functions. It associates a list of IPv6
address prefixes with a transit network link by referencing a
network- LSA, or associates a list of IPv6 address prefixes with a
router by referencing a router-LSA. A stub link's prefixes are
associated with its attached router.
A router may originate multiple intra-area-prefix-LSAs for a given
area, distinguished by their Link State ID fields. Each intra-area-
prefix-LSA contains an integral number of prefix descriptions.
A link's Designated Router originates one or more intra-area-prefix-
LSAs to advertise the link's prefixes throughout the area. For a link
L, L's Designated Router builds an intra-area-prefix-LSA in the
following fashion:
o In order to indicate that the prefixes are to be associated with
the Link L, the fields Referenced LS type, Referenced Link State
ID, and Referenced
Advertising Router are set to the corresponding fields in Link L's
network-LSA (namely LS type, Link State ID, and Advertising Router
respectively). This means that Referenced LS Type is set to
0x2002, Referenced Link State ID is set to the Designated Router's
Interface ID on Link L, and Referenced Advertising Router is set
to the Designated Router's Router ID.
o Each Link-LSA associated with Link L is examined (these are in the
Designated Router's interface structure for Link L). If the Link-
LSA's Advertising Router is fully adjacent to the Designated
Router, the list of prefixes in the Link-LSA is copied into the
intra-area-prefix-LSA that is being built. Prefixes having the
NU-bit and/or LA-bit set in their Options field should not be
copied, nor should link-local addresses be copied. Each prefix is
described by the PrefixLength, PrefixOptions, and Address Prefix
fields. Multiple prefixes having the same PrefixLength and Address
Prefix are considered to be duplicates; in this case their Prefix
Options fields should be merged by logically OR'ing the fields
together, and a single resulting prefix should be copied into the
intra-area-prefix-LSA. The Metric field for all prefixes is set to
0.
o The "# prefixes" field is set to the number of prefixes that the
router has copied into the LSA. If necessary, the list of prefixes
can be spread across multiple intra-area-prefix-LSAs in order to
keep the LSA size small.
A router builds an intra-area-prefix-LSA to advertise its own
prefixes, and those of its attached stub links. A Router RTX
would build its intra-area-prefix-LSA in the following fashion:
o In order to indicate that the prefixes are to be associated with
the Router RTX itself, RTX sets Referenced LS type to 0x2001,
Referenced Link State ID to 0, and Referenced Advertising Router
to RTX's own Router ID.
o Router RTX examines its list of interfaces to the area. If the
interface is in state Down, its prefixes are not included. If the
interface has been reported in RTX's router-LSA as a Type 2 link
description (link to transit network), its prefixes are not
included (they will be included in the intra-area-prefix-LSA for
the link instead). If the interface type is Point-to-MultiPoint,
or the interface is in state Loopback, or the interface connects
to a point-to-point link which has not been assigned a prefix,
then the site-local and global scope IPv6 addresses associated
with the interface (if any) are copied into the intra-area-
prefix-LSA, setting the LA-bit in the PrefixOptions field, and
setting the PrefixLength to 128 and the Metric to 0. Otherwise,
the list of site-local and global prefixes configured in RTX for
the link are copied into the intra-area-prefix-LSA by specifying
the PrefixLength, PrefixOptions, and Address Prefix fields. The
Metric field for each of these prefixes is set to the interface's
output cost.
o RTX adds the IPv6 prefixes for any directly attached hosts
belonging to the area (see Section C.7) to the intra-area-prefix-
LSA.
o If RTX has one or more virtual links configured through the area,
it includes one of its site-local or global scope IPv6 interface
addresses in the LSA (if it hasn't already), setting the LA-bit in
the PrefixOptions field, and setting the PrefixLength to 128 and
the Metric to 0. This information will be used later in the
routing calculation so that the two ends of the virtual link can
discover each other's IPv6 addresses.
o The "# prefixes" field is set to the number of prefixes that the
router has copied into the LSA. If necessary, the list of prefixes
can be spread across multiple intra-area-prefix-LSAs in order to
keep the LSA size small.
For example, the intra-area-prefix-LSA originated by RT4 for Network
N3 (assuming that RT4 is N3's Designated Router), and the intra-
area-prefix-LSA originated into Area 1 by Router RT3 for its own
prefixes, are pictured below.
; Intra-area-prefix-LSA
; for network link N3
LS age = 0 ;newly (re)originated
LS type = 0x2009 ;Intra-area-prefix-LSA
Link State ID = 5 ;or something
Advertising Router = 192.1.1.4 ;RT4's Router ID
# prefixes = 1
Referenced LS type = 0x2002 ;network-LSA reference
Referenced Link State ID = 1
Referenced Advertising Router = 192.1.1.4
PrefixLength = 56 ;N3's prefix
PrefixOptions = 0
Metric = 0
Address Prefix = 5f00:0000:c001:0100 ;pad
; RT3's Intra-area-prefix-LSA
; for its own prefixes
LS age = 0 ;newly (re)originated
LS type = 0x2009 ;Intra-area-prefix-LSA
Link State ID = 177 ;or something
Advertising Router = 192.1.1.3 ;RT3's Router ID
# prefixes = 1
Referenced LS type = 0x2001 ;router-LSA reference
Referenced Link State ID = 0
Referenced Advertising Router = 192.1.1.3
PrefixLength = 56 ;N4's prefix
PrefixOptions = 0
Metric = 2 ;N4 interface cost
Address Prefix = 5f00:0000:c001:0400 ;pad
When network conditions change, it may be necessary for a router to
move prefixes from one intra-area-prefix-LSA to another. For example,
if the router is Designated Router for a link but the link has no
other attached routers, the link's prefixes are advertised in an
intra-area-prefix-LSA referring to the Designated Router's router-
LSA. When additional routers appear on the link, a network-LSA is
originated for the link and the link's prefixes are moved to an
intra-area-prefix-LSA referring to the network-LSA.
Note that in the intra-area-prefix-LSA, the "Referenced Advertising
Router" is always equal to the router that is originating the intra-
area-prefix-LSA (i.e., the LSA's Advertising Router). The reason that
the Referenced Advertising Router field appears is that, even though
it is currently redundant, it may not be in the future. We may
sometime want to use the same LSA format to advertise address
prefixes for other protocol suites. In that event, the Designated
Router may not be running the other protocol suite, and so another of
the link's routers may need to send out the prefix-LSA. In that case,
"Referenced Advertising Router" and "Advertising Router" would be
different.
3.5. Flooding
Most of the flooding algorithm remains unchanged from the IPv4
flooding mechanisms described in Section 13 of [Ref1]. In particular,
the processes for determining which LSA instance is newer (Section
13.1 of [Ref1]), responding to updates of self-originated LSAs
(Section 13.4 of [Ref1]), sending Link State Acknowledgment packets
(Section 13.5 of [Ref1]), retransmitting LSAs (Section 13.6 of
[Ref1]) and receiving Link State Acknowledgment packets (Section 13.7
of [Ref1]) are exactly the same for IPv6 and IPv4.
However, the addition of flooding scope and handling options for
unrecognized LSA types (see Section A.4.2.1) has caused some changes
in the OSPF flooding algorithm: the reception of Link State Updates
(Section 13 in [Ref1]) and the sending of Link State Updates (Section
13.3 of [Ref1]) must take into account the LSA's scope and U-bit
setting. Also, installation of LSAs into the OSPF database (Section
13.2 of [Ref1]) causes different events in IPv6, due to the
reorganization of LSA types and contents in IPv6. These changes are
described in detail below.
3.5.1. Receiving Link State Update packets
The encoding of flooding scope in the LS type and the need to process unknown LS types causes modifications to the processing of received Link State Update packets. As in IPv4, each LSA in a received Link State Update packet is examined. In IPv4, eight steps are executed for each LSA, as described in Section 13 of [Ref1]. For IPv6, all the steps are the same, except that Steps 2 and 3 are modified as follows: (2) Examine the LSA's LS type. If the LS type is unknown, the area has been configured as a stub area, and either the LSA's flooding scope is set to "AS flooding scope" or the U-bit of the LS type is set to 1 (flood even when unrecognized), then discard the LSA and get the next one from the Link State Update Packet. This generalizes the IPv4 behavior where AS- external-LSAs are not flooded into/throughout stub areas. (3) Else if the flooding scope of the LSA is set to "reserved", discard the LSA and get the next one from the Link State Update Packet. Steps 5b (sending Link State Update packets) and 5d (installing LSAs in the link state database) in Section 13 of [Ref1] are also somewhat different for IPv6, as described in Sections 3.5.2 and 3.5.3 below.3.5.2. Sending Link State Update packets
The sending of Link State Update packets is described in Section 13.3 of [Ref1]. For IPv4 and IPv6, the steps for sending a Link State Update packet are the same (steps 1 through 5 of Section 13.3 in [Ref1]). However, the list of eligible interfaces out which to flood the LSA is different. For IPv6, the eligible interfaces are selected based on the following factors: o The LSA's flooding scope. o For LSAs with area or link-local flooding scoping, the particular area or interface that the LSA is associated with. o Whether the LSA has a recognized LS type. o The setting of the U-bit in the LS type. If the U-bit is set to 0, unrecognized LS types are treated as having link-local scope. If set to 1, unrecognized LS types are stored and flooded as if they were recognized.
Choosing the set of eligible interfaces then breaks into the
following cases:
Case 1
The LSA's LS type is recognized. In this case, the set of eligible
interfaces is set depending on the flooding scope encoded in the
LS type. If the flooding scope is "AS flooding scope", the
eligible interfaces are all router interfaces excepting virtual
links. In addition, AS-external-LSAs are not flooded out
interfaces connecting to stub areas. If the flooding scope is
"area flooding scope", the set of eligible interfaces are those
interfaces connecting to the LSA's associated area. If the
flooding scope is "link-local flooding scope", then there is a
single eligible interface, the one connecting to the LSA's
associated link (which, when the LSA is received in a Link State
Update packet, is also the interface the LSA was received on).
Case 2
The LS type is unrecognized, and the U-bit in the LS Type is set
to 0 (treat the LSA as if it had link-local flooding scope). In
this case there is a single eligible interface, namely, the
interface on which the LSA was received.
Case 3
The LS type is unrecognized, and the U-bit in the LS Type is set
to 1 (store and flood the LSA, as if type understood). In this
case, select the eligible interfaces based on the encoded flooding
scope as in Case 1 above. However, in this case interfaces
attached to stub areas are always excluded.
A further decision must sometimes be made before adding an LSA to a
given neighbor's link-state retransmission list (Step 1d in Section
13.3 of [Ref1]). If the LS type is recognized by the router, but not
by the neighbor (as can be determined by examining the Options field
that the neighbor advertised in its Database Description packet) and
the LSA's U-bit is set to 0, then the LSA should be added to the
neighbor's link-state retransmission list if and only if that
neighbor is the Designated Router or Backup Designated Router for the
attached link. The LS types described in detail by this memo, namely
router-LSAs (LS type 0x2001), network-LSAs (0x2002), Inter-Area-
Prefix-LSAs (0x2003), Inter-Area-Router-LSAs (0x2004), AS-External-
LSAs (0x4005), Link-LSAs (0x0008) and Intra-Area-Prefix-LSAs (0x2009)
are assumed to be understood by all routers. However, as an example
the group-membership-LSA (0x2006) is understood only by MOSPF routers
and since it has its U-bit set to 0, it should only be forwarded to a
non-MOSPF neighbor (determined by examining the MC-bit in the
neighbor's Database Description packets' Options field) when the
neighbor is Designated Router or Backup Designated Router for the
attached link. The previous paragraph solves a problem in IPv4 OSPF extensions such as MOSPF, which require that the Designated Router support the extension in order to have the new LSA types flooded across broadcast and NBMA networks (see Section 10.2 of [Ref8]).3.5.3. Installing LSAs in the database
There are three separate places to store LSAs, depending on their flooding scope. LSAs with AS flooding scope are stored in the global OSPF data structure (see Section 3.1) as long as their LS type is known or their U-bit is 1. LSAs with area flooding scope are stored in the appropriate area data structure (see Section 3.1.1) as long as their LS type is known or their U-bit is 1. LSAs with link-local flooding scope, and those LSAs with unknown LS type and U-bit set to 0 (treat the LSA as if it had link-local flooding scope) are stored in the appropriate interface structure. When storing the LSA into the link-state database, a check must be made to see whether the LSA's contents have changed. Changes in contents are indicated exactly as in Section 13.2 of [Ref1]. When an LSA's contents have been changed, the following parts of the routing table must be recalculated, based on the LSA's LS type: Router-LSAs, Network-LSAs, Intra-Area-Prefix-LSAs and Link-LSAs The entire routing table is recalculated, starting with the shortest path calculation for each area (see Section 3.8). Inter-Area-Prefix-LSAs and Inter-Area-Router-LSAs The best route to the destination described by the LSA must be recalculated (see Section 16.5 in [Ref1]). If this destination is an AS boundary router, it may also be necessary to re-examine all the AS-external-LSAs. AS-external-LSAs The best route to the destination described by the AS-external-LSA must be recalculated (see Section 16.6 in [Ref1]). As in IPv4, any old instance of the LSA must be removed from the database when the new LSA is installed. This old instance must also be removed from all neighbors' Link state retransmission lists.
3.6. Definition of self-originated LSAs
In IPv6 the definition of a self-originated LSA has been simplified from the IPv4 definition appearing in Sections 13.4 and 14.1 of [Ref1]. For IPv6, self-originated LSAs are those LSAs whose Advertising Router is equal to the router's own Router ID.3.7. Virtual links
OSPF virtual links for IPv4 are described in Section 15 of [Ref1]. Virtual links are the same in IPv6, with the following exceptions: o LSAs having AS flooding scope are never flooded over virtual adjacencies, nor are LSAs with AS flooding scope summarized over virtual adjacencies during the Database Exchange process. This is a generalization of the IPv4 treatment of AS-external-LSAs. o The IPv6 interface address of a virtual link must be an IPv6 address having site-local or global scope, instead of the link- local addresses used by other interface types. This address is used as the IPv6 source for OSPF protocol packets sent over the virtual link. o Likewise, the virtual neighbor's IPv6 address is an IPv6 address with site-local or global scope. To enable the discovery of a virtual neighbor's IPv6 address during the routing calculation, the neighbor advertises its virtual link's IPv6 interface address in an Intra-Area-Prefix-LSA originated for the virtual link's transit area (see Sections 3.4.3.7 and 3.8.1). o Like all other IPv6 OSPF interfaces, virtual links are assigned unique (within the router) Interface IDs. These are advertised in Hellos sent over the virtual link, and in the router's router- LSAs.3.8. Routing table calculation
The IPv6 OSPF routing calculation proceeds along the same lines as the IPv4 OSPF routing calculation, following the five steps specified by Section 16 of [Ref1]. High level differences between the IPv6 and IPv4 calculations include: o Prefix information has been removed from router-LSAs, and now is advertised in intra-area-prefix-LSAs. Whenever [Ref1] specifies that stub networks within router-LSAs be examined, IPv6 will instead examine prefixes within intra-area-prefix-LSAs.
o Type 3 and 4 summary-LSAs have been renamed inter-area-prefix-LSAs
and inter-area-router-LSAs (respectively).
o Addressing information is no longer encoded in Link State IDs, and
must instead be found within the body of LSAs.
o In IPv6, a router can originate multiple router-LSAs within a
single area, distinguished by Link State ID. These router-LSAs
must be treated as a single aggregate by the area's shortest path
calculation (see Section 3.8.1).
For each area, routing table entries have been created for the area's
routers and transit links, in order to store the results of the
area's shortest-path tree calculation (see Section 3.8.1). These
entries are then used when processing intra-area-prefix-LSAs, inter-
area-prefix-LSAs and inter-area-router-LSAs, as described in Section
3.8.2.
Events generated as a result of routing table changes (Section 16.7
of [Ref1]), and the equal-cost multipath logic (Section 16.8 of
[Ref1]) are identical for both IPv4 and IPv6.
3.8.1. Calculating the shortest path tree for an area
The IPv4 shortest path calculation is contained in Section 16.1 of
[Ref1]. The graph used by the shortest-path tree calculation is
identical for both IPv4 and IPv6. The graph's vertices are routers
and transit links, represented by router-LSAs and network-LSAs
respectively. A router is identified by its OSPF Router ID, while a
transit link is identified by its Designated Router's Interface ID
and OSPF Router ID. Both routers and transit links have associated
routing table entries within the area (see Section 3.3).
Section 16.1 of [Ref1] splits up the shortest path calculations into
two stages. First the Dijkstra calculation is performed, and then the
stub links are added onto the tree as leaves. The IPv6 calculation
maintains this split.
The Dijkstra calculation for IPv6 is identical to that specified for
IPv4, with the following exceptions (referencing the steps from the
Dijkstra calculation as described in Section 16.1 of [Ref1]):
o The Vertex ID for a router is the OSPF Router ID. The Vertex ID
for a transit network is a combination of the Interface ID and
OSPF Router ID of the network's Designated Router.
o In Step 2, when a router Vertex V has just been added to the
shortest path tree, there may be multiple LSAs associated with the
router. All Router-LSAs with Advertising Router set to V's OSPF
Router ID must processed as an aggregate, treating them as
fragments of a single large router-LSA. The Options field and the
router type bits (bits W, V, E and B) should always be taken from
"fragment" with the smallest Link State ID.
o Step 2a is not needed in IPv6, as there are no longer stub network
links in router-LSAs.
o In Step 2b, if W is a router, there may again be multiple LSAs
associated with the router. All Router-LSAs with Advertising
Router set to W's OSPF Router ID must processed as an aggregate,
treating them as fragments of a single large router-LSA.
o In Step 4, there are now per-area routing table entries for each
of an area's routers, instead of just the area border routers.
These entries subsume all the functionality of IPv4's area border
router routing table entries, including the maintenance of virtual
links. When the router added to the area routing table in this
step is the other end of a virtual link, the virtual neighbor's IP
address is set as follows: The collection of intra-area-prefix-
LSAs originated by the virtual neighbor is examined, with the
virtual neighbor's IP address being set to the first prefix
encountered having the "LA-bit" set.
o Routing table entries for transit networks, which are no longer
associated with IP networks, are also modified in Step 4.
The next stage of the shortest path calculation proceeds similarly to
the two steps of the second stage of Section 16.1 in [Ref1]. However,
instead of examining the stub links within router-LSAs, the list of
the area's intra-area-prefix-LSAs is examined. A prefix advertisement
whose "NU-bit" is set should not be included in the routing
calculation. The cost of any advertised prefix is the sum of the
prefix' advertised metric plus the cost to the transit vertex (either
router or transit network) identified by intra-area-prefix-LSA's
Referenced LS type, Referenced Link State ID and Referenced
Advertising Router fields. This latter cost is stored in the transit
vertex' routing table entry for the area.
3.8.1.1. The next hop calculation
In IPv6, the calculation of the next hop's IPv6 address (which will
be a link-local address) proceeds along the same lines as the IPv4
next hop calculation (see Section 16.1.1 of [Ref1]). The only
difference is in calculating the next hop IPv6 address for a router
(call it Router X) which shares a link with the calculating router. In this case the calculating router assigns the next hop IPv6 address to be the link-local interface address contained in Router X's Link- LSA (see Section A.4.8) for the link. This procedure is necessary since on some links, such as NBMA links, the two routers need not be neighbors, and therefore might not be exchanging OSPF Hellos.3.8.2. Calculating the inter-area routes
Calculation of inter-area routes for IPv6 proceeds along the same lines as the IPv4 calculation in Section 16.2 of [Ref1], with the following modifications: o The names of the Type 3 summary-LSAs and Type 4 summary-LSAs have been changed to inter-area-prefix-LSAs and inter-area-router-LSAs respectively. o The Link State ID of the above LSA types no longer encodes the network or router described by the LSA. Instead, an address prefix is contained in the body of an inter-area-prefix-LSA, and a described router's OSPF Router ID is carried in the body of an inter-area- router-LSA. o Prefixes having the "NU-bit" set in their Prefix Options field should be ignored by the inter-area route calculation. When a single inter-area-prefix-LSA or inter-area-router-LSA has changed, the incremental calculations outlined in Section 16.5 of [Ref1] can be performed instead of recalculating the entire routing table.3.8.3. Examining transit areas' summary-LSAs
Examination of transit areas' summary-LSAs in IPv6 proceeds along the same lines as the IPv4 calculation in Section 16.3 of [Ref1], modified in the same way as the IPv6 inter-area route calculation in Section 3.8.2.3.8.4. Calculating AS external routes
The IPv6 AS external route calculation proceeds along the same lines as the IPv4 calculation in Section 16.4 of [Ref1], with the following exceptions: o The Link State ID of the AS-external-LSA types no longer encodes the network described by the LSA. Instead, an address prefix is contained in the body of an AS- external-LSA.
o The default route is described by AS-external-LSAs which advertise
zero length prefixes.
o Instead of comparing the AS-external-LSA's Forwarding address
field to 0.0.0.0 to see whether a forwarding address has been
used, bit F of the external-LSA is examined. A forwarding address
is in use if and only if bit F is set.
o Prefixes having the "NU-bit" set in their Prefix Options field
should be ignored by the inter-area route calculation.
When a single AS-external-LSA has changed, the incremental
calculations outlined in Section 16.6 of [Ref1] can be performed
instead of recalculating the entire routing table.
3.9. Multiple interfaces to a single link
In OSPF for IPv6, a router may have multiple interfaces to a single
link. All interfaces are involved in the reception and transmission
of data traffic, however only a single interface sends and receives
OSPF control traffic. In more detail:
o Each of the multiple interfaces are assigned different Interface
IDs. In this way the router can automatically detect when
multiple interfaces attach to the same link, when receiving Hellos
from its own Router ID but with an Interface ID other than the
receiving interface's.
o The router turns off the sending and receiving of OSPF packets
(that is, control traffic) on all but one of the interfaces to the
link. The choice of interface to send and receive control traffic
is implementation dependent; as one example, the interface with
the highest Interface ID could be chosen. If the router is
elected DR, it will be this interface's Interface ID that will be
used as the network-LSA's Link State ID.
o All the multiple interfaces to the link will however appear in the
router-LSA. In addition, a Link-LSA will be generated for each of
the multiple interfaces. In this way, all interfaces will be
included in OSPF's routing calculations.
o If the interface which is responsible for sending and receiving
control traffic fails, another will have to take over, reforming
all neighbor adjacencies from scratch. This failure can be
detected by the router itself, when the other interfaces to the
same link cease to hear the router's own Hellos.
References
[Ref1] Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998. [Ref2] McKenzie, A., "ISO Transport Protocol specification ISO DP 8073", RFC 905, April 1984. [Ref3] McCloghrie, K. and F. Kastenholz, "The Interfaces Group MIB using SMIv2", RFC 2233, November 1997. [Ref4] Fuller, V., Li, T, Yu, J. and K. Varadhan, "Classless Inter- Domain Routing (CIDR): an Address Assignment and Aggregation Strategy", RFC 1519, September 1993. [Ref5] Varadhan, K., Hares, S. and Y. Rekhter, "BGP4/IDRP for IP--- OSPF Interaction", RFC 1745, December 1994 [Ref6] Reynolds, J. and J. Postel, "Assigned Numbers", STD 2, RFC 1700, October 1994. [Ref7] deSouza, O. and M. Rodrigues, "Guidelines for Running OSPF Over Frame Relay Networks", RFC 1586, March 1994. [Ref8] Moy, J., "Multicast Extensions to OSPF", RFC 1584, March 1994. [Ref9] Coltun, R. and V. Fuller, "The OSPF NSSA Option", RFC 1587, March 1994. [Ref10] Ferguson, D., "The OSPF External Attributes LSA", unpublished. [Ref11] Moy, J., "Extending OSPF to Support Demand Circuits", RFC 1793, April 1995. [Ref12] Mogul, J. and S. Deering, "Path MTU Discovery", RFC 1191, November 1990. [Ref13] Rekhter, Y. and T. Li, "A Border Gateway Protocol 4 (BGP-4)", RFC 1771, March 1995. [Ref14] Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification", RFC 2460, December 1998. [Ref15] Hinden, R. and S. Deering, "IP Version 6 Addressing Architecture", RFC 2373, July 1998.
[Ref16] Conta, A. and S. Deering, "Internet Control Message Protocol (ICMPv6) for the Internet Protocol Version 6 (IPv6) Specification" RFC 2463, December 1998. [Ref17] Narten, T., Nordmark, E. and W. Simpson, "Neighbor Discovery for IP Version 6 (IPv6)", RFC 2461, December 1998. [Ref18] McCann, J., Deering, S. and J. Mogul, "Path MTU Discovery for IP version 6", RFC 1981, August 1996. [Ref19] Kent, S. and R. Atkinson, "IP Authentication Header", RFC 2402, November 1998. [Ref20] Kent, S. and R. Atkinson, "IP Encapsulating Security Payload (ESP)", RFC 2406, November 1998.
(next page on part 3)
