what does DUP! means? 2

[luciodandrea]

hi, i'm on a lan which is subnetted with /28, and the address scheme is a class C which begins with 192.168.1.1 , so that the broadcast address is .15...
so, when i ping to the .15 address, i get this:

64 bytes from 192.168.1.1: icmp_seq=0 ttl=64 time=1.345 ms (DUP!)
64 bytes from 192.168.1.10: icmp_seq=0 ttl=255 time=2.554 ms (DUP!)
64 bytes from 192.168.1.3: icmp_seq=0 ttl=64 time=18.569 ms (DUP!)

what does the DUP! means, and why sometimes in certain LANs when i ping to the broadcast address, i recieve responses from all of the hosts, and im some other LANs not????
thanks!

 

[manarth]

i'm on a lan which is subnetted with /28, and the address scheme is a class C

It's either subnetted /28, or a class C. Not both.

what does the DUP! mean

A quick google search will help you find the answer.

why...when i ping the broadcast address, i recieve responses from all of the hosts, and...other LANs not...

Again, google is your friend. - particularly the top post from Microsoft.

<marc>
New to Tek-Tips? Get better answers - faq581-3339

 

[luciodandrea]

you're right, i've expressed in a bad way...what i was trying to say was that it was subnetted using a class C address (192.168.....)
now i'll check those links!
thanks!!

 

[ChrisAC]

It's either subnetted /28, or a class C. Not both.

Not so! It can be a Class C address with a /28 mask. Class C does *not* mean that it's a /24!

Chris.

**********************
Chris A.C, CCNA, CCSA
**********************

 

[manarth]

It's either subnetted /28, or a class C. Not both.

Not so! It can be a Class C address with a /28 mask. Class C does *not* mean that it's a /24!

RFC 1236 includes a definition of a Class C network:[tt]

Class C:

The three highest-order bits are set to 1-1-0.
21-bits define the network number.
8-bits define the local address.[/tt]

Binary Representation[tt]
nnnnnnnn.nnnnnnnn.nnnnnnnn.hhhhhhhh
11000000.00000000.00000000.xxxxxxxx
11011111.11111111.11111111.xxxxxxxx[/tt]

Therefore, a Class C network has to be a /24 network with a network address in between 192.168.0 and 223.255.255



Other references:[ul]
[li]RFC 1366 refers on numerous occassions to the public Class C networks.[/li]
[li]Webopedia[/li]
[li]TCP IP Guide[/li][/ul]

The RFC on Private Addressing (RFC 1918) doesn't refer to classes at all, and in fact the private addressing scheme of 192.168 refers to a /16 network.

It's a common misconception that 192.168.x.x refers to a Class C network; one that I try to correct where I can!

Whilst in many cases it would be correct, it allows room for misunderstanding - for example, classful routers cannot route a /28 subnet correctly (they only handle /8, /16 and /24 networks). Referring to a /28 as a 'Class C' could lead to the belief that a classful router would be an appropriate choice for that setting.

---
Marcus
better questions get better answers - faq581-3339
accessible web design - zioncore.com

 

[ChrisAC]

The class is defined by the first octet (ie. three highest-order bits are set to 110). /24 is the *default* mask for a class C but does not define the class.

However, your last statement is quite correct. The problem is that most believe that the mask is what defines the class when in fact it's the first octet of the IP address.

As a matter of point I try avoid using the term 'class A/B/C/D etc..' in favour of simply stating the mask length.

I think that everyone understands now.

Chris.

**********************
Chris A.C, CCNA, CCSA
**********************

 

[manarth]

The class is defined by the first octet (ie. three highest-order bits are set to 110). /24 is the *default* mask for a class C but does not define the class.

Where did you hear this?

Everything I've ever learned about IP addressing, and according to RFC 791 and RFC 796, the class of network specifies both the High Order Bits and the 'bits of net':

[tt]
Address Formats:

High Order Bits Format Class
--------------- ------------------------------- -----
0 7 bits of net, 24 bits of host a
10 14 bits of net, 16 bits of host b
110 21 bits of net, 8 bits of host c
111 escape to extended addressing mode

[/tt]

[tt] The third type (or class c) of address has a 21-bit network number
and a 8-bit local address.[/tt]

---
Marcus
better questions get better answers - faq581-3339
accessible web design - zioncore.com

 

[cyberspace]

I thought the Octets in a class C address were N:N:N:H

- So how is this a 21bit network number? What happens to the other three bits?

'When all else fails.......read the manual'

 

[cyberspace]

Actually to answer my own question it's obviously because of the three high order bits (110)

Never thought of it as a 21bit network number though..hence my confusion

'When all else fails.......read the manual'

 

[cyberspace]

Just another question:

Where did you hear this?

Everything I've ever learned about IP addressing, and according to RFC 791 and RFC 796, the class of network specifies both the High Order Bits and the 'bits of net':

I am not sure what point you are making here?

If the high order bits of an address are:

0 - You know it is a Class A address
10 - You know it is a Class B address
110 - You know it is a Class C address

What ChrisAC said, is the same that I have learned through the CCNA programme.


'When all else fails.......read the manual'

 

[manarth]

If the high order bits of an address are:

0 - You know it is a Class A address
10 - You know it is a Class B address
110 - You know it is a Class C address

That's not exactly the case: to be a Class C, it must use a /24 subnet. That is the definition of a Class C address.

A high order bit of 110 and a subnet of /28 does not make a Class C network address.

ChrisAC is also a CCNA/CCSA - I suspect that this may be taught by some within the CCNA program, but it is not the entire story, as the RFCs detail.

---
Marcus
better questions get better answers - faq581-3339
accessible web design - zioncore.com

 

[cyberspace]

Thanks for that info Manarth, didn't know that!

So I presume that referring to the high order bits to explicitly define the class of a given address should only be done from a classful addressing perspective?

'When all else fails.......read the manual'

 

[manarth]

indeed, referring to a Class A/B/C network is only appropriate in classful addressing.

When engineers are referred to a Class x network, they're usually expecting a corresponding subnet mask. Finding out post-installation that a new server won't connect because the network is in fact a subnetted/supernetted Class x can be quite annoying (the amount of annoyance being directly proportional to time spent trying to diagnose the 'fault'!)

Hence CIDR being Classless Inter-Domain Routing ;-)

---
Marcus
better questions get better answers - faq581-3339
accessible web design - zioncore.com