one final belaborment of the point. ( where is Bob Vance, whose thoroughness
would be quite welcome here? )
observe the following from a lab setup. all interfaces in the pod are /29
subnets of the 37.0.0.0 network. note the RIP routes that are in the routing
table. auto summarization is turned off, but even with it turned on, there
is no effect on the routes advertised. other vendors may have different ways
of doing it.
How is it that RIP, a classful protocol, can advertise non classful subnets?
Gateway of last resort is 37.1.1.26 to network 0.0.0.0
C 171.171.0.0/16 is directly connected, Loopback0
37.0.0.0/29 is subnetted, 7 subnets
R 37.1.1.32 [120/1] via 37.1.1.26, 00:00:10, Serial0
R 37.1.1.40 [120/1] via 37.1.1.26, 00:00:10, Serial0
R 37.1.1.48 [120/2] via 37.1.1.26, 00:00:10, Serial0
R 37.1.1.0 [120/3] via 37.1.1.26, 00:00:10, Serial0
R 37.1.1.8 [120/2] via 37.1.1.26, 00:00:10, Serial0
C 37.1.1.16 is directly connected, Ethernet0
C 37.1.1.24 is directly connected, Serial0
R* 0.0.0.0/0 [120/3] via 37.1.1.26, 00:00:10, Serial0
the answer is that classfulness really has nothing to do with routing
protocol behaviour, unless certain circumstances dictate. when everyone
agrees on a prefix length, i.e. when all interfaces in the RIP domain have
the same mask, all the proper routes are advertised.
remember that the RIP packet contains fields for advertised networks. each
field is 32 bits in length. there is no subnet mask information. it matters
not one iota to RIP what the advertised networks are. 37.0.0.0 fits into the
field as does 37.1.1.16 etc.
BTW, the directed broadcast addresses of each of the subnets in the table
are respectively 37.1.1.7, .15, .23, .31, .39, .47, and .55 classful
protocol or not, this is the way it works.
sorry to be so obsessive about this, but a thorough understanding of
protocol behaviour can be helpful in many circumstances. for example, I have
a problem I am working on with a customer that revolves around RIP
behaviour. I'm not going to convince them to migrate to OSPF, or even RIPv2
as a fix. I have to live within their existing constraints. My ability to
understand how RIP really works allows me to create a design that serves the
customer needs.
good learning to all
Chuck
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of
Leigh Anne Chisholm
Sent: Wednesday, July 25, 2001 10:15 AM
To: [EMAIL PROTECTED]
Subject: RE: ccna challenge question [7:13565]
Ed, what you're forgetting is this: a directed broadcast would either be
generated by a device 172.18.2.0 network that has been configured with a
subnet mask of 255.255.254.0 or by and end-system specifically trying to
reach an end-host configured as part of that subnet (for example, a
user-initiated "ping" sent to 172.18.3.255 trying to illicit a response from
all end-systems on the 172.18.2.0 /22 subnet). If RouterB does not have an
interface directly connected to the 172.18.0.0 network, yes it will see
network 172.18.2.0 as 172.18.0.0 and to it it will think the directed
broadcast is 172.18.255.255 but why would RouterB need to GENERATE a
directed broadcast to this network? RouterB is likely to only RECEIVE a
directed broadcast sent to 172.18.3.255 and won't CARE about the IP address
of that packet. It takes the packet and forwards it to the 172.18.0.0
network.
As a side note, if RouterB DOES have an interface that is part of the
172.18.0.0 network, that interface must be configured with the same subnet
mask as the 172.18.2.0 network as variable-length subnet masking is not
supported by RIP. In other words, the subnet mask used in conjunction with
any subnet from the 172.18.0.0 network must be configured with the subnet
mask 255.255.252.0. Then RouterB would know that 172.18.3.255 is the
directed broadcast address.
Let's look at the question again and look it from a different perspective:
Using classful assumptions, what is the directed broadcast address for
172.18.2.0 with the mask 255.255.254.0?
Now put that on a PC. What will the directed broadcast address be if this
was on a PC? RouterB really doesn't even enter into the picture.
There's really nothing wrong with the question, if you understand the
concept of classful and classless routing.
-- Leigh Anne
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of
Ed Moss
Sent: Wednesday, July 25, 2001 9:30 AM
To: [EMAIL PROTECTED]
Subject: Re: ccna challenge question [7:13565]
Everyone has some good ideas on this... and one of the replies made me
rethink my previous answer.... Look at it from the aspect of the ROUTER...
assuming that this is implemented on a router.
An interface with the address of 172.18.2.0 and a mask of 255.255.254.0
will send directed broadcasts to 172.18.3.255. Answer B is correct.
However the above router sends a RIP version 1 update (which is classless)
to router B. Router B gets the router 172.18.2.0 and no mask information.
It assumes a classful boundary. ( I know, there are other rules that affect
the advertisement and receipt of the route) Router B will then send
directed broadcasts to 172.18.255.255.
I think everyone would agree that the question is poorly worded... If I
came across the question I would still select 'C' as my answer,
172.18.255.255. My clue in the question is 'using classful assumptions'
Ed
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=13827&t=13565
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
