rfc1142.txt

RFC 的详细文档！

and a LAN Level 2 Designated Intermediate System purges 
Level 2 Link State PDUs.  An Intermediate system which 
has resigned as both Level 1 and Level 2 Designated Inter
mediate System shall purge both sets of LSPs.
 
When an Intermediate system declares itself as designated 
Intermediate system and it is in possession of a Link State 
PDU of the same level issued by the previous Designated 
Intermediate System for that circuit (if any), it shall initiate 
a network wide purge of that (or those) Link State PDU(s) 
as above.
7.2.4 Links
Two Intermediate systems are not considered neighbours 
unless each reports the other as directly reachable over one 
of their SNPAs. On a Connection-oriented subnetwork 
(either point-to-point or general topology), the two Interme
diate systems in question shall ascertain their neighbour re
lationship when a connection is established and hello PDUs 
exchanged. A malfunctioning IS might, however, report an
other IS to be a neighbour when in fact it is not. To detect 
this class of failure the decision process checks that each 
link reported as up in a LSP is so reported by both Inter
mediate systems. If an Intermediate system considers a link 
down it shall not mention the link in its Link State PDUs.
On broadcast subnetworks, this class of failure shall be de
tected by the designated IS, which has the responsibility to 
ascertain the set of Intermediate systems that can all com
municate on the subnetwork. The designated IS shall in
clude these Intermediate systems (and no others) in the 
Link State PDU it generates for the pseudonode represent
ing the broadcast subnetwork.
7.2.5 Multiple LSPs for the same system
The Update process is capable of dividing a single logical 
LSP into a number of separate PDUs for the purpose of 
conserving link bandwidth and processing (see 7.3.4).  The 
Decision Process, on the other hand, shall regard the LSP 
with LSP Number zero in a special way. If the LSP with 
LSP Number zero and remaining lifetime > 0, is not present 
for a particular system then the Decision Process shall not 
process any LSPs with non-zero LSP Number which may 
be stored for that system. 
The following information shall be taken only from the LSP 
with LSP Number zero. Any values which may be present 
in other LSPs for that system shall be disregarded by the 
Decision Process.
a)The setting of the LSP Database Overload bit. 
b)The value of the IS Type field. 
c)The Area Addresses option.
7.2.6 Routeing Algorithm Overview
The routeing algorithm used by the Decision Process is a 
shortest path first (SPF) algorithm. Instances of the algo
rithm are run independently and concurrently by all Inter
mediate systems in a routeing domain. Intra-Domain route
ing of a PDU occurs on a hop-by-hop basis: that is, the al
gorithm determines only the next hop, not the complete 
path, that a data PDU will take to reach its destination. To 
guarantee correct and consistent route computation by 
every Intermediate system in a routeing domain, this Inter
national Standard depends on the following properties:
 
a)All Intermediate systems in the routeing domain con
verge to using identical topology information; and
b)Each Intermediate system in the routeing domain gen
erates the same set of routes from the same input to
pology and set of metrics.
The first property is necessary in order to prevent inconsis
tent, potentially looping paths. The second property is nec
essary to meet the goal of determinism stated in 6.6.
A system executes the SPF algorithm to find a set of legal 
paths to a destination system in the routeing domain. The 
set may consist of:
a)a single path of minimum metric sum: these are 
termed minimum cost paths;
b)a set of paths of equal minimum metric sum: these are 
termed equal minimum cost paths; or
c)a set of paths which will get a PDU closer to its desti
nation than the local system: these are called down
stream paths.
Paths which do not meet the above conditions are illegal 
and shall not be used.
The Decision Process, in determining its paths, also ascer
tains the identity of the adjacency which lies on the first 
hop to the destination on each path. These adjacencies are 
used to form the Forwarding Database,  which the forward
ing process uses for relaying PDUs.
Separate route calculations are made for each pairing of a 
level in the routeing hierarchy (i.e. L1 and L2) with a sup
ported routeing metric. Since there are four routeing metrics 
and two levels some systems may execute multiple in
stances of the SPF algorithm. For example,
-if an IS is a L2 Intermediate system which supports all 
four metrics and computes minimum cost paths for all 
metrics, it would execute the SPF calculation eight 
times.
-if an IS is a L1 Intermediate system which supports all 
four metrics, and additionally computes downstream 
paths, it would execute the algorithm  4 W (number of 
neighbours + 1) times.
Any implementation of an SPF algorithm meeting both the 
static and dynamic conformance requirements of clause 12 
of this International Standard may be used. Recommended 
implementations are described in detail in Annex C.
7.2.7 Removal of Excess Paths
When there are more than max

i

mum

Path

Splits legal 
paths to a destination, this set shall be pruned until only 
max

i

mum

Path

Splits remain. The Intermediate system 
shall discriminate based upon:
NOTE - The precise precedence among the paths is speci
fied in order to meet the goal of determinism defined in 6.6.
 
-adjacency type: Paths associated with End system or 
level 2 reachable address prefix adjacencies are re
tained in preference to other adjacencies
-metric sum: Paths having a lesser metric sum are re
tained in preference to paths having a greater metric 
sum. By metric sum is understood the sum of the 
metrics along the path to the destination.
-neighbour ID: where two or more paths are associ
ated with adjacencies of the same type, an adjacency 
with a lower neighbour ID is retained in preference to 
an adjacency with a higher neighbour id.
-circuit ID: where two or more paths are associated 
with adjacencies of the same type, and same neigh
bour ID, an adjacency with a lower circuit ID is re
tained in preference to an adjacency with a higher cir
cuit ID, where circuit ID is the value of:
7ptPtCircuitID for non-broadcast circuits, 
7l1CircuitID for broadcast circuits when running 
the Level 1 Decision Process, and 
7l2CircuitID for broadcast circuits when running 
the Level 2 Decision Process. 
-lANAddress: where two or more adjacencies are of 
the same type, same neighbour ID, and same circuit 
ID (e.g. a system with multiple LAN adapters on the 
same circuit) an adjacency with a lower lANAddress 
is retained in preference to an adjacency with a higher 
lANAddress.
7.2.8 Robustness Checks
7.2.8.1 Computing Routes through Overloaded 
Intermediate systems
The Decision Process shall not utilise a link to an Interme
diate system neighbour from an IS whose LSPs have the 
LSP Database Overload indication set. Such paths may in
troduce loops since the overloaded IS does not have a com
plete routeing information base. The Decision Process shall, 
however utilise the link to reach End system neighbours 
since these paths are guaranteed to be non-looping.
7.2.8.2 Two-way connectivity check
The Decision Process shall not utilise a link between two 
Intermediate Systems unless both ISs report the link. 
NOTE - the check is not applicable to links to an End Sys
tem.
Reporting the link indicates that it has a defined value for at 
least the default routeing metric. It is permissible for two 
endpoints to report different defined values of the same 
metric for the same link. In this case, routes may be asym
metric.
 
7.2.9 Construction of a Forwarding Database
The information that is needed in the forwarding database 
for routeing metric k is the set of adjacencies for each sys
tem N.
7.2.9.1 Identification of Nearest Level 2 IS by a 
Level 1 IS
Level 1 Intermediate systems need one additional piece of 
information per routeing metric: the next hop to the nearest 
level 2 Intermediate system according to that routeing met
ric. A level 1 IS shall ascertain the set, R, of attached 
level 2 Intermediate system(s) for metric k such that the to
tal cost to R for metric k is minimal. 
If there are more adjacencies in this set than max

i

mum

 
Path

Splits, then the IS shall remove excess adjacencies as 
described in 7.2.7.
7.2.9.2 Setting the Attached Flag in Level 2 
Intermediate Systems
If a level 2 Intermediate system discovers, after computing 
the level 2 routes for metric k, that it cannot reach any other 
areas using that metric, it shall:
-set AttachedFlag for metric k to False;
-regenerate its Level 1 LSP with LSP number zero; and
-compute the nearest level 2 Intermediate system for 
metric k for insertion in the appropriate forwarding 
database, according to the algorithm described in 
7.2.9.1 for level 1 Intermediate systems.
NOTE - AttachedFlag for each metric k is examined by the 
Update Process, so that it will report the value in the ATT 
field of its Link State PDUs. 
If a level 2 Intermediate system discovers, after computing 
the level 2 routes for metric k, that it can reach at least one 
other area using that metric, it shall
-set AttachedFlag for metric k to True;
-regenerate its Level 1 LSP with LSP number zero; and 
-set the level 1 forwarding database entry for metric k 
which corresponds to nearest level 2 Intermediate 
system to Self.
7.2.10 Information for Repairing Partitioned 
Areas
An area may become partitioned as a result of failure of one 
or more links in the area. However, if each of the partitions 
has a connection to the level 2 subdomain, it is possible to 
repair the partition via the level 2 subdomain, provided that 
the level 2 subdomain itself is not partitioned. This is illus
trated in Figure 4.
All the systems A  I, R and P are in the same area n. 
When the link between D and E is broken, the area be
 
comes partitioned. Within each of the partitions the Parti
tion Designated Level 2 Intermediate system is selected 
from among the level 2 Intermediate systems in that parti
tion. In the case of partition 1 this is P, and in the case of 
partition 2 this is R. The level 1 repair path is then estab
lished between between these two level 2 Intermediate sys
tems. Note that the repaired link is now between P and R, 
not between D and E.
The Partition Designated Level 2 Intermediate Systems re
pair the partition by forwarding NPDUs destined for other 
partitions of the area through the level 2 subdomain. They 
do this by acting in their capacity as Level 1 Intermediate 
Systems and advertising in their Level 1 LSPs adjacencies 
to each Partition Designated Level 2 Intermediate System 
in the area. This adjacency is known as a Virtual Adja
cency or Virtual Link. Thus other Level 1 Intermediate 
Systems in a partition calculate paths to the other partitions 
through the Partition Designated Level 2 Intermediate Sys
tem. A Partition Designated Level 2 Intermediate System 
forwards the Level 1 NPDUs through the level 2 subdomain 
by encapsulating them in 8473 Data NPDUs with its Virtual 
Network Entity Title as the source NSAP and the adja
cent Partition Designated Level 2 Intermediate System's 
Virtual Network Entity Title as the destination NSAP. The 
following sub-clauses describe this in more detail.
7.2.10.1 Partition Detection and Virtual Level 1 
Link Creation
Partitions of a Level 1 area are detected by the Level 2 In
termediate System(s) operating within the area.  In order to 
participate in the partition repair process, these Level 2 In
termediate systems must also act as Level 1 Intermediate 
systems in the area. A partition of a given area exists when
ever two or more Level 2 ISs located in that area are re
ported in the L2 LSPs as being a Partition Designated 
Level 2 IS. Conversely, when only one Level 2 IS in an 
area is reported as being the Partition Designated Level 2 
 
IS, then that area is not partitioned.  Partition repair is ac
complished by the Partition Designated Level 2 IS.  The 
election of the Partition Designated Level 2 IS as described 
in the next subsection must be done before the detection 
and repair process can begin.
In order to repair a partition of a Level 1 area, the Partition 
designated Level 2 IS creates a Virtual Network Entity to 
represent the partition.  The Network Entity Title for this 
virtual network entity shall be constructed from the first 
listed area address from its Level 2 Link State PDU, and the 
ID of the Partition Designated Level 2 IS.  The IS shall also 
construct a virtual link (represented by a new Virtual Adja
cency managed object) to each Partition Designated Level 2 
IS in the area, with the NET of the partition recorded in the 
Identifier attribute.  The virtual links are the repair paths for 
the partition.  They are reported by the Partition Designated 
Level 2 IS into the entire Level 1 area by adding the ID of 
each adjacent Partition Designated Level 2 IS to the In
termediate System Neighbours field of its Level 1 Link 
State PDU.  The Virtual Flag shall be set True for these 
Intermediate System neighbours.  The metric value for this 
virtual link shall be the default metric value d(N) obtained 
from this system's Level 2 PATHS database, where N is the 
adjacent Partition Designated Level 2 IS via the Level 2 
subdomain.
An Intermediate System which operates as the Partition 
Designated Level 2 Intermediate System shall perform the 
following steps after completing the Level 2 shortest path 
computation in order to detect partitions in the Level 1 area 
and create repair paths: 
a)Examine Level 2 Link State PDUs of all