Alberto wrote:
> Hi. I would like more info on the following: I want to make some
> calculus of subnetting and I am looking for info about the numbers of
> subnets available.
> RFC 950 says I should use n bits for addressing till 2^n -2 subnets, and
> I should avoid all 0's and all 1's subnets.
> But on "VLSM" RFC 1878 it says it is not necessary to avoid these
> subnets. What should I do ? What is current RFC being applied ???

I suspect that as you go up in RFC number you will see which supercede
which previous ones. In this case 950 and the concept of subnet
address has been superceded by (IIRC) CIDR - "class-less" IP
addresses.

> If I dont make VLSM (ie. all my subnets have the same number of hosts) ,
> could I use all 0's and all 1's subnets ? Or are reserved ?

Only if your hosts are running really old software that isn't up on
the current RFCs.

rick jones
--
The computing industry isn't as much a game of "Follow The Leader" as
it is one of "Ring Around the Rosy" or perhaps "Duck Duck Goose."
- Rick Jones
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

In article <_Whig.2328624$kp.12532393@telenews.teleline.es>, Alberto writes:
> Hi. I would like more info on the following: I want to make some
> calculus of subnetting and I am looking for info about the numbers of
> subnets available.
> RFC 950 says I should use n bits for addressing till 2^n -2 subnets, and
> I should avoid all 0's and all 1's subnets.
> But on "VLSM" RFC 1878 it says it is not necessary to avoid these
> subnets. What should I do ? What is current RFC being applied ???

RFC 1812, the "IPv4 router requirements" RFC from June, 1995 has somewhat
to say on the matter and appears to be the current applicable document.

RFC 1878 is little more than a bunch of convenient tables.


RFC 1812 includes the following:

4.2.1 INTRODUCTION

Routers MUST implement the IP protocol, as defined by [INTERNET:1].
They MUST also implement its mandatory extensions: subnets (defined
in [INTERNET:2]), IP broadcast (defined in [INTERNET:3]), and
Classless Inter-Domain Routing (CIDR, defined in [INTERNET:15]).

.... and ...

DISCUSSION
Previous versions of this document also noted that subnet numbers
must be neither 0 nor -1, and must be at least two bits in length.
In a CIDR world, the subnet number is clearly an extension of the
network prefix and cannot be interpreted without the remainder of
the prefix. This restriction of subnet numbers is therefore
meaningless in view of CIDR and may be safely ignored.

webriggs@encompasserve.org escribió:
> In article <_Whig.2328624$kp.12532393@telenews.teleline.es>, Alberto writes:
>> Hi. I would like more info on the following: I want to make some
>> calculus of subnetting and I am looking for info about the numbers of
>> subnets available.
>> RFC 950 says I should use n bits for addressing till 2^n -2 subnets, and
>> I should avoid all 0's and all 1's subnets.
>> But on "VLSM" RFC 1878 it says it is not necessary to avoid these
>> subnets. What should I do ? What is current RFC being applied ???
>
> RFC 1812, the "IPv4 router requirements" RFC from June, 1995 has somewhat
> to say on the matter and appears to be the current applicable document.
>
> RFC 1878 is little more than a bunch of convenient tables.
>
>
> RFC 1812 includes the following:
>
> 4.2.1 INTRODUCTION
>
> Routers MUST implement the IP protocol, as defined by [INTERNET:1].
> They MUST also implement its mandatory extensions: subnets (defined
> in [INTERNET:2]), IP broadcast (defined in [INTERNET:3]), and
> Classless Inter-Domain Routing (CIDR, defined in [INTERNET:15]).
>
> ... and ...
>
> DISCUSSION
> Previous versions of this document also noted that subnet numbers
> must be neither 0 nor -1, and must be at least two bits in length.
> In a CIDR world, the subnet number is clearly an extension of the
> network prefix and cannot be interpreted without the remainder of
> the prefix. This restriction of subnet numbers is therefore
> meaningless in view of CIDR and may be safely ignored.
well tnx for info, but what about if I dont use CIDR on my LAN ??? Would
be possible ?
I know 2days world IP is CIDR, but could I "mix" CIDR+classic subnetting
for my LAN addressing, or CIDR+VLSM ???

In article ,
Alberto wrote:

> well tnx for info, but what about if I dont use CIDR on my LAN ??? Would
> be possible ?
> I know 2days world IP is CIDR, but could I "mix" CIDR+classic subnetting
> for my LAN addressing, or CIDR+VLSM ???

As far as the devices are concerned, there's not really any difference.
Classic subnetting is just a special case of CIDR.

Pretty much any IP stack designed in the last decade should not consider
classful addresses specially, they should do everything with addresses
and masks.

--
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***

Alberto wrote:
> webriggs@encompasserve.org escribió:
> > In article <_Whig.2328624$kp.12532393@telenews.teleline.es>, Alberto writes:
> >> Hi. I would like more info on the following: I want to make some
> >> calculus of subnetting and I am looking for info about the numbers of
> >> subnets available.
> >> RFC 950 says I should use n bits for addressing till 2^n -2 subnets, and
> >> I should avoid all 0's and all 1's subnets.
> >> But on "VLSM" RFC 1878 it says it is not necessary to avoid these
> >> subnets. What should I do ? What is current RFC being applied ???
> >
> > RFC 1812, the "IPv4 router requirements" RFC from June, 1995 has somewhat
> > to say on the matter and appears to be the current applicable document.
> >
> > RFC 1878 is little more than a bunch of convenient tables.
> >
> >
> > RFC 1812 includes the following:
> >
> > 4.2.1 INTRODUCTION
> >
> > Routers MUST implement the IP protocol, as defined by [INTERNET:1].
> > They MUST also implement its mandatory extensions: subnets (defined
> > in [INTERNET:2]), IP broadcast (defined in [INTERNET:3]), and
> > Classless Inter-Domain Routing (CIDR, defined in [INTERNET:15]).
> >
> > ... and ...
> >
> > DISCUSSION
> > Previous versions of this document also noted that subnet numbers
> > must be neither 0 nor -1, and must be at least two bits in length.
> > In a CIDR world, the subnet number is clearly an extension of the
> > network prefix and cannot be interpreted without the remainder of
> > the prefix. This restriction of subnet numbers is therefore
> > meaningless in view of CIDR and may be safely ignored.
> well tnx for info, but what about if I dont use CIDR on my LAN ??? Would
> be possible ?

no

> I know 2days world IP is CIDR, but could I "mix" CIDR+classic subnetting

yes

classic subnetting is FLSM

CIDR, so , apply rule from RFC 1812 , use All 0 and All 1

> for my LAN addressing, or CIDR+VLSM ???

yes

similarly. use All 0 and All 1.

Barry Margolin wrote:
> In article ,
> Alberto wrote:
>
> > well tnx for info, but what about if I dont use CIDR on my LAN ??? Would
> > be possible ?
> > I know 2days world IP is CIDR, but could I "mix" CIDR+classic subnetting
> > for my LAN addressing, or CIDR+VLSM ???
>
> As far as the devices are concerned, there's not really any difference.
> Classic subnetting is just a special case of CIDR.
>
> Pretty much any IP stack designed in the last decade should not consider
> classful addresses specially, they should do everything with addresses
> and masks.
>
> --

given the network prefix terminology, "network prefix", "extended
network prefix". How do we as humans refer to the "bits of subnetting",
the difference between the "NW prefix" and "extended NW prefix". The
term subnet , I think since VLSM, tends to refer to the whole lot, and
is now meant to be obsoleted by 'network prefix'. So is there a term,
besides "bits of subnetting"?

CIDR seems to call anything after the "network prefix" the Host. which
is unhelpful and not really the area i'm referring to. I'm referring to
layer 2 in a "standard" 3 layer scheme that anybody always uses:
block,subnet,host

In article <1150326429.739656.23080@h76g2000cwa.googlegroups.c om>,
"q_q_anonymous@yahoo.co.uk" wrote:

> CIDR seems to call anything after the "network prefix" the Host. which
> is unhelpful and not really the area i'm referring to. I'm referring to
> layer 2 in a "standard" 3 layer scheme that anybody always uses:
> block,subnet,host

Who is this "anybody"? The advent of CIDR has made the distinction
between networks and subnets irrelevant. All any device cares about is
whether the address it's sending to is local or remote, i.e. whether it
can send directly to it or must go through a router. There's no special
significance to collecting subnets into a particular classful network.

Originally the all-1's subnet was reserved because they had the idea
that you might want to send a broadcast to ALL the subnets of your
network, and you would do that using the address {net, -1, -1}. This
turned out not to be a useful function, and few routers implemented it
(it's difficult to avoid broadcast loops). So the only special
significance of the network/subnet distinction was never really used,
and CIDR eliminated this special case for good.

--
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***

Barry Margolin wrote:
> In article <1150326429.739656.23080@h76g2000cwa.googlegroups.c om>,
> "q_q_anonymous@yahoo.co.uk" wrote:
>
> > CIDR seems to call anything after the "network prefix" the Host. which
> > is unhelpful and not really the area i'm referring to. I'm referring to
> > layer 2 in a "standard" 3 layer scheme that anybody always uses:
> > block,subnet,host
>
> Who is this "anybody"? The advent of CIDR has made the distinction
> between networks and subnets irrelevant. All any device cares about is
> whether the address it's sending to is local or remote, i.e. whether it
> can send directly to it or must go through a router. There's no special
> significance to collecting subnets into a particular classful network.
>

a human being might want to refer to it, since he is subnetting the
network. He won't be
creating subnets in the block part. There are still 3 layers there.

what you have shown is that a computer or router wouldn't care. But I
know that. A computer wouldn't be designing the subnetting scheme, or
describing where he subnetted the block or even which part is the
block.

> Originally the all-1's subnet was reserved because they had the idea
> that you might want to send a broadcast to ALL the subnets of your
> network, and you would do that using the address {net, -1, -1}. This
> turned out not to be a useful function, and few routers implemented it
> (it's difficult to avoid broadcast loops). So the only special
> significance of the network/subnet distinction was never really used,
> and CIDR eliminated this special case for good.
>

I know, thanks

In article <1150333213.554519.74900@y41g2000cwy.googlegroups.c om>,
"q_q_anonymous@yahoo.co.uk" wrote:

> Barry Margolin wrote:
> > In article <1150326429.739656.23080@h76g2000cwa.googlegroups.c om>,
> > "q_q_anonymous@yahoo.co.uk" wrote:
> >
> > > CIDR seems to call anything after the "network prefix" the Host. which
> > > is unhelpful and not really the area i'm referring to. I'm referring to
> > > layer 2 in a "standard" 3 layer scheme that anybody always uses:
> > > block,subnet,host
> >
> > Who is this "anybody"? The advent of CIDR has made the distinction
> > between networks and subnets irrelevant. All any device cares about is
> > whether the address it's sending to is local or remote, i.e. whether it
> > can send directly to it or must go through a router. There's no special
> > significance to collecting subnets into a particular classful network.
> >
>
> a human being might want to refer to it, since he is subnetting the
> network. He won't be
> creating subnets in the block part. There are still 3 layers there.

Not really. You're given a block of addresses. If you have multiple
networks at your location, you decide on an appropriate way to apportion
the block among them, and define the network masks for each one.
There's no need to think of it as 3 layers, you just define several
2-layer networks that happen to span the block you were given.

--
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***

Barry Margolin wrote:
> In article <1150333213.554519.74900@y41g2000cwy.googlegroups.c om>,
> "q_q_anonymous@yahoo.co.uk" wrote:
>
> > Barry Margolin wrote:
> > > In article <1150326429.739656.23080@h76g2000cwa.googlegroups.c om>,
> > > "q_q_anonymous@yahoo.co.uk" wrote:
> > >
> > > > CIDR seems to call anything after the "network prefix" the Host. which
> > > > is unhelpful and not really the area i'm referring to. I'm referring to
> > > > layer 2 in a "standard" 3 layer scheme that anybody always uses:
> > > > block,subnet,host
> > >
> > > Who is this "anybody"? The advent of CIDR has made the distinction
> > > between networks and subnets irrelevant. All any device cares about is
> > > whether the address it's sending to is local or remote, i.e. whether it
> > > can send directly to it or must go through a router. There's no special
> > > significance to collecting subnets into a particular classful network.
> > >
> >
> > a human being might want to refer to it, since he is subnetting the
> > network. He won't be
> > creating subnets in the block part. There are still 3 layers there.
>
> Not really. You're given a block of addresses. If you have multiple
> networks at your location, you decide on an appropriate way to apportion
> the block among them, and define the network masks for each one.
> There's no need to think of it as 3 layers, you just define several
> 2-layer networks that happen to span the block you were given.
>
> --

but if you were using CIDR+FLSM then in order to work out what your
subnets are, you have to think about those "bits of subnetting" as
distinct from the block.

certainly, to know how many subnets there are

furthermore, even to know if you're deling with subnet 0,1,2,3, or 4
(relative subnet number). To know the multiple, the bit values of
those bits of subnetting. To know the values of the interesting/split
octet

if you say we don't use FLSM. Surely FLSM is very convenient since it's
presumably quicker to set up than VLSM.. It's more wasteful but it can
still be used.

On 2006-06-14, Barry Margolin wrote:

> can send directly to it or must go through a router. There's no special
> significance to collecting subnets into a particular classful network.

There is in the network routing world, where sometimes it is highly
desirable to collect a contiguous set of subnets or the entirety of
a classful address range inside a routing "domain" such that the core
of the network has the minimum of network routing information needed
to push packets around. (This is of course a gross oversimplification of
that for the sake of this posting.). The process in the routing world
is called "summarization".

> Originally the all-1's subnet was reserved because they had the idea
> that you might want to send a broadcast to ALL the subnets of your
> network, and you would do that using the address {net, -1, -1}. This
> turned out not to be a useful function, and few routers implemented it

Excellent historical note, thank you.

/dmfh

----
__| |_ __ / _| |_ ____ __
dmfh @ / _` | ' \| _| ' \ _ / _\ \ /
\__,_|_|_|_|_| |_||_| (_) \__/_\_\
----

In article ,
DMFH wrote:

> On 2006-06-14, Barry Margolin wrote:
>
> > can send directly to it or must go through a router. There's no special
> > significance to collecting subnets into a particular classful network.
>
> There is in the network routing world, where sometimes it is highly
> desirable to collect a contiguous set of subnets or the entirety of
> a classful address range inside a routing "domain" such that the core
> of the network has the minimum of network routing information needed
> to push packets around. (This is of course a gross oversimplification of
> that for the sake of this posting.). The process in the routing world
> is called "summarization".

But you still don't need to think of it as a
triple.

Within the contiguous set of subnets you use the fine-grained network
mask and deal with . Outside that contiguous set you use
the summary network mask, and again deal with .

Nothing that deals with subnets needs to know about the aggregation, and
nothing that deals with the aggregate needs to know about the subnets,
so there's never any need to think of it as a 3-level network.

Of course, the guy who sets up the subnets needs to know what aggregate
he has available to split up.

--
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***

Barry Margolin wrote:

(snip)

> As far as the devices are concerned, there's not really any difference.
> Classic subnetting is just a special case of CIDR.

Except for non-contiguous masks classically allowed, but not for CIDR.

> Pretty much any IP stack designed in the last decade should not consider
> classful addresses specially, they should do everything with addresses
> and masks.

There are still old hosts around. I have seen some that won't allow
supernetting, because they believe the old classes. But yes, probably
more than 10 years old.

-- glen