[c-nsp] Current CCNA tests

Peter Rathlev peter at rathlev.dk
Wed Jan 9 13:29:00 EST 2008


On Wed, 2008-01-09 at 08:42 +0100, Gert Doering wrote:
> The answer to "how many subnets of size 16 can you stuff into a class C"
> is "15" or "16", depending on whether "ip subnet-zero" is in use...
> 
> I wasn't asking about "how many hosts" :-)

Sorry, my bad. It was late. :-)

> > Anyway, I still find it funny that even a lot of professionals can't
> > distinguish the 3.4.5.0/24 (a subnet) from 201.202.203.0/24 (a network).
>  
> There *is* *no* *difference*.
> 
> Class X networks do not exist anymore.  Networks are specified by a base 
> address and a netmask / CIDR bits.  "Subnet" and "Supernet" are terms from 
> the last century *that have no meaning anymore* in a CIDR environment.

The CIDR way of looking at routes does not distinguish, but there are
some bits in the MSB end that make a difference. But yes: For all
practical purposes there is no difference between {some class C}/24 and
{som subnet of class A}/24. Just like there is no CIDR difference
between 62.242.15.0/24 and 239.0.0.0/24 or even 253.72.88.0/24. :-)

> (If there is a difference between 3.4.5.0/24 and 201.202.203.0/24, it's
> because the *implementation* in a given router has not arrived in the
> 21st century, and has implicit assumptions on things based on two bits
> in front of the network address.  But that's broken implementations.)

So a router that treats multicast (class D) in a special way is "so last
century"? ;-)

(Sorry for being pedantic, I agree that people should generally stop
thinking in classful routing terms!)

Regards,
Peter



More information about the cisco-nsp mailing list