Iprange and netmask explanation question

We were in the situation that there were too many computers and not
enough ipaddresses at school. We asked the network administrator too
expand the range of addresses and he answered it would not be possible.
Instead he added a whole new subnet to our network. It's like this...

We had addresses 1.1.1.66 until 1.1.1.127 this had a netmask of
255.255.255.192. Now the network administrator added 1.1.1.128 until
1.1.1.192 again with a netmask of 255.255.255.192. So I'm thinking just
change the netmask to 255.255.255.127 if I'm correct so we will have one
bigger subnet instead of 2 small ones. Since the ipaddresses are not
private addresses I'm guessing I'm wrong and thinking too simple and the
network administrator is right, right. But can someone explain why this
administrator cannot expand our current range but can add another subnet
which numbering follows the first subnet.

Arnaud.

A. Loonstra wrote:
> We were in the situation that there were too many computers and not
> enough ipaddresses at school. We asked the network administrator too
> expand the range of addresses and he answered it would not be possible.
> Instead he added a whole new subnet to our network. It's like this...
>
> We had addresses 1.1.1.66 until 1.1.1.127 this had a netmask of
> 255.255.255.192. Now the network administrator added 1.1.1.128 until
> 1.1.1.192 again with a netmask of 255.255.255.192. So I'm thinking just
> change the netmask to 255.255.255.127 if I'm correct so we will have one
> bigger subnet instead of 2 small ones. Since the ipaddresses are not
> private addresses I'm guessing I'm wrong and thinking too simple and the
> network administrator is right, right. But can someone explain why this
> administrator cannot expand our current range but can add another subnet
> which numbering follows the first subnet.
>
> Arnaud.

Almost, but not quite.

The subnet mask you are using (255.255.255.192) chops 1.1.1.x
into 4 groups: 0-63, 64-127, 128-191, 192-255.

A subnet mask of 255.255.255.128 (not 127) would chop 1.1.1.x
into two groups: 0-127 and 128-255.

I assume that he cannot simply expand your subnet from 64-127 to
0-127 because 0-63 are already used elsewhere in the network.

It is also possible that the routing protocol in use cannot cope
with the concept of not all subnets being the same size (e.g.
RIPv1 or IGRP).

The addresses you have been assigned, 64-127 and 128-192, do not
merge until you use a mask of 255.255.255.0, But then you lose
sight of both the other departments that are the other side of
the router (0-63 and 192-255) because they would appear to be
local, and your machines would not try to use the router to
connect to them. (Actually, enabling "proxy arp" on the router
would allow this, but it is only available on some models of router).

Steve

Steve Horsley wrote:
[snip]
>
> Almost, but not quite.
>
> The subnet mask you are using (255.255.255.192) chops 1.1.1.x into 4
> groups: 0-63, 64-127, 128-191, 192-255.
>
> A subnet mask of 255.255.255.128 (not 127) would chop 1.1.1.x into two
> groups: 0-127 and 128-255.
>
> I assume that he cannot simply expand your subnet from 64-127 to 0-127
> because 0-63 are already used elsewhere in the network.

Yes that's the case. 0-63 are used for personel. Students should not be
able to access that range. But from what I read from your response it is
not possible to have a 0-127 range and a 128-192 range since the netmask
of the first range forces the complete range split in two. So he cannot
do 1.1.1.0-127 with a netmask of 255.255.255.127 and a second range of
1.1.1.128-191 with a netmask of 255.255.255.192.

>
> It is also possible that the routing protocol in use cannot cope with
> the concept of not all subnets being the same size (e.g. RIPv1 or IGRP).
>
> The addresses you have been assigned, 64-127 and 128-192, do not merge
> until you use a mask of 255.255.255.0, But then you lose sight of both
> the other departments that are the other side of the router (0-63 and
> 192-255) because they would appear to be local, and your machines would
> not try to use the router to connect to them. (Actually, enabling "proxy
> arp" on the router would allow this, but it is only available on some
> models of router).
>
> Steve
>
Our school has the complete range for the ipaddresses and more ranges.
The network administrator said it needed to order a new subnet from the
supplier of the ipaddresses in order to fix our problem. But when they
already own the complete iprange they can do whatever they want with how
many subnets they place on their network. Just as long as they stay
within the range provided. Or is even the subnet division provided by
the supplier of the iprange?

Thanks for your response, it was really helpful.

Arnaud.

"A. Loonstra" <arnaud@[nospam]sphaero.org> wrote in news:11itm27rr1ghn66
@corp.supernews.com:

> We were in the situation that there were too many computers and not
> enough ipaddresses at school. We asked the network administrator too
> expand the range of addresses and he answered it would not be possible.
> Instead he added a whole new subnet to our network. It's like this...
>
> We had addresses 1.1.1.66 until 1.1.1.127 this had a netmask of
> 255.255.255.192. Now the network administrator added 1.1.1.128 until
> 1.1.1.192 again with a netmask of 255.255.255.192. So I'm thinking just
> change the netmask to 255.255.255.127 if I'm correct so we will have one
> bigger subnet instead of 2 small ones.

No. A subnet mask 255.255.255.128 applied to the address 1.1.1.66 is telling
your TCP stack that the subnet is starting at 1.1.1.0 through 1.1.1.127. i.e
usable addresses 1.1.1.1 through 1.1.1.126. The addresses 1.1.1.1 - 1.1.1.62
are probably already assigned to another subnet. So what the admin did is
correct. You can't have a subnet 1.1.1.64 - 1.1.1.191, you would have to make
this two subnets, each with a 255.255.255.192 mask. A single subnet with a
mask of 255.255.255.128 has to start on a .0 or .128 boundary. You could have
completely renumbered the subnet to 1.1.1.128 - 1.1.1.254 using the
255.255.255.128 mask, provided the higher addreses from 1.1.1.192 - 1.1.1.254
are not already in use.

> Since the ipaddresses are not
> private addresses I'm guessing I'm wrong and thinking too simple and the
> network administrator is right, right. But can someone explain why this
> administrator cannot expand our current range but can add another subnet
> which numbering follows the first subnet.

Subnet masks are binary bit masks e.g 255.255.255.128 =
11111111.11111111.11111111.10000000. The mask is logically anded with the ip
address for the ip router to determine the subnet that the address belongs
to. Networkers would write the above mask as /25 and your 255.255.255.192
mask as /26. The number after / tells you how many binary ones are in the
mask. The more ones the smaller the subnet. e.g:

/8 = a class A subnet mask 255.0.0.0 (16777214 possible addresses).
/16 = a class B subnet mask 255.255.0.0 (65534 possible addresses).
/24 = a class C subnet mask 255.255.255.0 (254 possible addresses).

Subnets are always 2^n addresses but you can't use the first and last address
so the number of usable addresses is 2^ n - 2. A subnet always has to start
on a natural boundary. For example a /28 subnet can't start with a .8 address
or a .24 etc, it has to start on a multiple of 16, so starting at .0, .16,
..32 etc would be ok. A /25 subnet cannot start at .64 which is why your admin
added a second /26 subnet instead.

BTW the use of IANA reserved block 1.0.0.0/8 is a bit unusual. You should
refer you admin to the following RFC:

http://www.faqs.org/rfcs/rfc1918.html

Klazmon.





>
> Arnaud.

arnaud <arnaud@[nospam]!sphaero.org> wrote in news:11iuffhmrrjkp07
@corp.supernews.com:

> Steve Horsley wrote:
> [snip]
>>
>> Almost, but not quite.
>>
>> The subnet mask you are using (255.255.255.192) chops 1.1.1.x into 4
>> groups: 0-63, 64-127, 128-191, 192-255.
>>
>> A subnet mask of 255.255.255.128 (not 127) would chop 1.1.1.x into two
>> groups: 0-127 and 128-255.
>>
>> I assume that he cannot simply expand your subnet from 64-127 to 0-127
>> because 0-63 are already used elsewhere in the network.
>
> Yes that's the case. 0-63 are used for personel. Students should not be
> able to access that range. But from what I read from your response it is
> not possible to have a 0-127 range and a 128-192 range since the netmask
> of the first range forces the complete range split in two. So he cannot
> do 1.1.1.0-127 with a netmask of 255.255.255.127 and a second range of
> 1.1.1.128-191 with a netmask of 255.255.255.192.

1.1.1.0-127 would be a netmask 255.255.255.128. A netmask 255.255.255.127
is nonsensical. Write it out in binary.


>
>>
>> It is also possible that the routing protocol in use cannot cope with
>> the concept of not all subnets being the same size (e.g. RIPv1 or IGRP).
>>
>> The addresses you have been assigned, 64-127 and 128-192, do not merge
>> until you use a mask of 255.255.255.0, But then you lose sight of both
>> the other departments that are the other side of the router (0-63 and
>> 192-255) because they would appear to be local, and your machines would
>> not try to use the router to connect to them. (Actually, enabling "proxy
>> arp" on the router would allow this, but it is only available on some
>> models of router).
>>
>> Steve
>>
> Our school has the complete range for the ipaddresses and more ranges.
> The network administrator said it needed to order a new subnet from the
> supplier of the ipaddresses in order to fix our problem. But when they
> already own the complete iprange they can do whatever they want with how
> many subnets they place on their network. Just as long as they stay
> within the range provided. Or is even the subnet division provided by
> the supplier of the iprange?

I am assuming that the 1.x.x.x are not the real addresses then. The
1.0.0.0/8 block is reserved by IANA.

Klazmon.






>
> Thanks for your response, it was really helpful.
>
> Arnaud.
>

Llanzlan Klazmon wrote:
[snip]
> BTW the use of IANA reserved block 1.0.0.0/8 is a bit unusual. You should
> refer you admin to the following RFC:
>
> http://www.faqs.org/rfcs/rfc1918.html
>
> Klazmon.

Thanks for your explanation. The 1.0.0.0/8 block I just made up. The
range used at school is a different one but I didn't want to use that in
my explanation. The situation is exactly the same.

Arnaud.