Figure 4.1: Passive RIP listener
The host H1 listen to RIP packets, but do not generate any routing announcement. To be used on a single connected workstation. A default route to the router R1 is set for safety which will be kept into the kernel routing table when gated exists (use of the retain statement). The logging in done into the file /tmp/gated.log, maximum size of 100kB, rotated once.
#
# This configuration runs RIP in quiet mode, it only listens to
# packets, no matter how many interfaces are configured.
#
traceoptions "/var/spool/gated/gatedlog" size 10k files 2 general;
rip yes {
nobroadcast ;
};
static {
0.0.0.0 gateway 198.108.0.1 retain;
};
Figure 4.2: RIP internal router
In this example a router with several interface runs only RIP. RIP will accept any RIP route from any router, and will export to any interface all RIP routes as well as attached networks.
#
# Just enable RIP
#
rip yes;
This this the same example as the previous one, except that a static default route is set and announced to the other RIP routers, and all subnets of 192.168.1 are aggregated into a natural class C network and announced to the other RIP routers.
rip yes;
# Define a default route
static {
default gateway 192.168.1.18; #ether address of WORLD router
};
# Create the aggregate 192.168.1.0/24 from the attached subnetwork
# 192.168.1.0/28 and 192.168.1.16/28
aggregate 192.168.1.0 mask 255.255.255.0 {
proto direct {
192.168.1.0 mask 255.255.255.240;
192.168.1.16 mask 255.255.255.240;
};
};
# RIP exportation
export proto rip metric 2 {
# The aggregate (class C net)
proto aggregate {
192.168.1.0 mask 255.255.255.0;
};
# The static route
proto static {
default;
};
};
The router R2 (in the autonomous system 1234) has a BGP session with the router R1 (in the autonomous system 2021). R2 will accept only the default route from R1, and will announce to R1 all it internal route got from RIP. R2 will also redistribute the default route to the internal routers through RIP.
#
# My autonomous system
#
autonomousystem 1234;
#
# RIP enable except on the DMZ (192.168.128)
#
rip yes {
interface le1 noripin noripout;
};
#
# BGP session with the ISP
#
bgp yes {
group type external peeras 2021 {
peer 192.168.128.1;
};
};
#
# Import only default from the ISP
#
import proto bgp {
default;
};
#
# Export our routes to the ISP
#
export proto bgp {
proto rip {
all;
};
};
#
# Redistribute default to our networks via RIP
#
export proto rip metric 2 {
proto bgp as 2021 {
default;
};
proto direct {
all;
};
};
This is the same example as the previous, but R2 will perform some aggregation of internal routes and announces only the aggregate to the ISP router.
#
# My autonomous system
#
autonomousystem 1234;
#
# RIP enable except on the DMZ (192.168.128)
#
rip yes {
interface le1 noripin noripout;
};
#
# BGP session with the ISP
#
bgp yes {
group type external peeras 2021 {
peer 192.168.128.1;
};
};
#
# Import only default from the ISP
#
import proto bgp {
default;
};
#
# Export our routes to the ISP
#
export proto bgp {
proto aggregate { all; };
proto rip { all; };
};
#
# Redistribute default to our networks via RIP
#
export proto rip metric 2 {
proto bgp {
default;
};
proto direct {all; };
};
#
# Aggregate our 4 class C nets into a /22 net. Two are directly connected nets.
#
aggregate 192.168.128.0 masklen 22 {
proto direct {
192.168.128.0 masklen 24 ;
192.168.129.0 masklen 24 ;
};
proto rip {
192.168.130.0 masklen 24 ;
192.168.131.0 masklen 24 ;
};
};
As with unicast routing protocols, there are a number of multicast routing protocols, each with its own advantages and disadvantages. In GateD, support for DVMRP, MOSPF, and Dense mode and Sparse mode PIM are anticipated. Currently, only Dense mode PIM is completed. All of these protocols rely on IGMP.
This configuration simply enables IGMP and PIM on all interfaces. The default mode for PIM is dense mode.
#
# Turn on IGMP
#
igmp yes ;
#
# Turn on PIM
#
pim yes ;
If you are running PIM in the presence of tunneled DVMRP and the GateD multicast kernel then PIM will try to run over the tunnel pseudo interfaces. They should be disabled as follows:
igmp yes ;
pim yes {
interface tu disable ;
};
An example of a DVMRP host with one tunnel. This host also runs RIP, and does a lot of extensive logging:
traceoptions "/tmp/gated.log" replace all;
rip yes;
kernel {
traceoptions all request;
};
igmp on {
traceoptions all packets;
};
dvmrp on {
#
# Tunnel definition. First remote address, then local address
#
tunnel 198.108.60.64 lcladdr 192.231.253.50;
traceoptions all packets;
};
static {
default gateway 192.231.253.2 retain;
};
