Question about RIPv2 classless IP ranges

[Roberthas]

This is my first post, I have searched and attempted to find the answer to this and could not, therefore, the post...

It is my understanding that RIPv2 advertises the interface assigned IP addresses, as if they were classfull.

According to RFC 1918:

A = 10.0.0.0 /8  ,

B = 172.16.0.0 /12

C = 192.168.0.0 /16

I seek clarification on the subject of the last class and RIPv2 advertising it.  An address of 192.168.1.1/24 would be treated as if it were 192.168.x.x/16, would it not?

I ask because I have seen MANY individuals writing as if the 192.168.x.x is a /24 address range AND a class C and have even encountered the problem on practice tests ( which I believe are in error, according to the rules).

Quite simply, according to RIPv2, is an interface with ip address 192.168.1.1/24 advertised, based on it being a /24 or /16 network?

EXAMPLE:

Router(config-router)#network 10.0.0.0 

Router(config-router)#network 192.168.0.0

Router(config-router)#network 172.31.0.0

Router(config)# Int G1/0

Router(config-int)#ip address 10.2.1.1 255.255.255.0

Router(config)# Int G1/1

Router(config-int)#ip address 192.168.1.1 255.255.255.0

Router(config)# Int G1/2

Router(config-int)#ip address 172.16.2.1 255.255.255.0

In this example, interfaces G1/0 and G1/1 would get advertised and not G1/2, correct?

I hope my wording is comprehend-able as my immediate family think I speak gibberish, or gibrii'... :)

I thank you in advance for responses that shed light on this subject.

Roberhas

[DAVID YARASHUS]

In short, no.  When the RFC1918 range of 192.168.0.0/16 is specified, that's just shorthand for saying "all addresses in this range (192.168.0.0-192.168.255.255," but that range is entirely within the classful Class C space, and that's how other systems will generally interpret any given part of it in the absence of other specific mask information for the specific /24 classful network within that /16 address block.

In particular, with routing, while you can specify a network range of a supernetted CIDR block (and people often do), if you do not provide a mask, the traditional classful treatment will generally be default behavior.

In your example, Gi1/1 and Gi1/2 would not be matched, because those network statements are interpreted classfully unless you provide mask information along with them, and so all of 10.0.0.0/8 is matched, but your network statement specifies 192.168.0.0 (with an implied /24) while your Gi1/1 interface is in a different classful network (192.168.1.0/24). Similarly, your Gi1/2 interface is in the subnet 172.16.2.0/24 with no matching major network specified - since your network statement is for 172.31.0.0 (as an implied /16).

Basically, any time you are specifying a network for an address, you need to assume things will treat it classfully unless you override that by providing supernet or VLSM mask info.

In practice, RIPv2 advertisements generally include the actual configured mask of the interface, as described in RFC 2543. 

[Roberthas]

I am studying for my first time taking the ICND1 and your answer has a lot to interpret.   I have read it several times and am trying to determine what all it is saying.  

I thank you for your response and will go look to RFC 2543 for further clarification.

In a practice test with a similar question to the example I provided, it stated that ONLY the 10. network would be advertised and I thought that the 192. should be also...  Is that incorrect?

Thanks,

Roberthas

[DAVID YARASHUS · Accepted solution]

The practice question's answer is correct.
Network statements are interpreted as covering their respective classful ranges unless you override that by explicitly providing a supernet mask (such as 172.16.0.0/12 or 192.168.0.0/16).

[Roberthas]

Thank you very much!  I clearly understand that description and am now inspired to go find out what the term "supernet mask" means and its uses....

Thanks,

Roberthas

[DAVID YARASHUS]

The mainstream classful masks are:
255.0.0.0 or /8 for class A networks
255.255.0.0 or /16 for class B networks, and
255.255.255.0 or /24 for class C networks
(we don't talk about classes D & E much).
A subnet mask is when you use more bits in the mask than the default classful mask.
A supernet mask is when you use fewer bits in the mask than the default classful mask.
For example, if you use a 255.255.255.252 mask with, well, anything, there's a subnet involved, obviously, because all classful networks have shorter masks than that.
If you use 255.240.0.0 (/12) there may or may not be a supernet involved. There's a subnet involved if it's a class A, but it's a supernet if you're referring to addresses in the traditional class B or C space. Hope that summary is useful as you're starting your studies.

[Roberthas]

VERY east to understand! I thank you!.  I have never thought of the term "subnet" having a relational value to it that would later be joined with "supernet".  It is very easy to comprehend and I do recall seeing something to the effect of supernet, (or an abbreviated version of the statement) in the practice test questions at times.

I have been taking various practice tests for ICDN1 for almost 2 weeks now and am ready to go do it for real anytime now.

I thank you for the sharing of easy to understand information!

Thanks,

Roberthas