Information Technology Reference
In-Depth Information
The traffic flow from an NE to the management station is straightforward. The data from
the NE is sent toward the closest L1/L2 routers by following the attach bit (ATT bit) in the
L1 area. When the packet reaches the L1/L2 router, the CLNS default route, not the ATT
bit, is followed into the core autonomous system. The specific CLNS route to the manage-
ment station is used in the network core. The traffic flow from the management station can
follow a specific NSAP prefix back to the network element being managed.
BGP Next Hop for CLNS Prefixes
An interesting caveat comes with using BGP to transport CLNS prefix information. The
recursive routing functionality that BGP uses with respect to next-hop information is not
conducive to CLNS operation. IP addressing is per-link, whereas CLNS addressing is per-
node. This section looks at how the addressing paradigm for CLNS creates a complication
for using BGP.
The BGP protocol can establish a connection between the two peers because they share an
IP subnet, which means that each router knows how to reach the other router to which it is
directly connected. However, in CLNS, addressing is done on a per-node basis, not a per-
link basis. The result is that in a CLNS network, two directly connected routers do not share
any addressing information to provide unicast reachability between them without some
discovery mechanism. Figure A-3 shows an example.
CLNS Next-Hop Reachability
Figure A-3
47.5678.2222.2222.2222.00
47.1234.1111.1111.1111.00
.2
.1
10.1.1.0/30
R1
R2
As shown in Figure A-3, both routers have IP reachability through the 10.1.1.0/30 subnet.
However, both routers are in different CLNS areas because of the different area addresses.
R1 is in area 47.5678, and R2 is in area 47.1234.
In an IS-IS network, two nodes send out Hello packets and discover each other, resulting in
an adjacency's being formed. However, if in the DCN example the aggregation routers were
to form L2 adjacencies with the core routers, the result would be one very large L2 domain,
which is the initial problem that BGP is being used to solve.
A solution is to form an End System-to-Intermediate System (ES-IS) adjacency between
the aggregation router and the core router. An ES-IS adjacency is used to allow routers to
send data packets to ESs, or hosts. Using an ES-IS adjacency allows each router to learn
about the other router without merging the L2 routing domains.
This solution is not without complications. The core routers and aggregation routers are
all L2-only routers, which are not permitted to advertise an ES-IS adjacency according to
