maximum members of the workgroup

Any body knows how many machines ( max.) we can connect
in the one workgroup?

thx...ben

Should be 253.

"Benyamin" <(E-Mail Removed)> wrote in message
news:18e901c3baed$e40404f0$(E-Mail Removed)...
>
> Any body knows how many machines ( max.) we can connect
> in the one workgroup?
>
> thx...ben
>

In article <#(E-Mail Removed)>, "Carey Holzman"
<(E-Mail Removed)> wrote:
>"Benyamin" <(E-Mail Removed)> wrote in message
>news:18e901c3baed$e40404f0$(E-Mail Removed)...
>> Any body knows how many machines ( max.) we can connect
>> in the one workgroup?
>>
>> thx...ben
>
>Should be 253.

Where did you get that number, Carey? To the best of my knowledge,
there's no theoretical limit to the number of machines in a workgroup.

I haven't tested it, but I've read that the practical limit for good
network browsing performance in a workgroup is around 20 machines.
--
Best Wishes,
Steve Winograd, MS-MVP (Windows Networking)

Please post any reply as a follow-up message in the news group
for everyone to see. I'm sorry, but I don't answer questions
addressed directly to me in E-mail or news groups.

Microsoft Most Valuable Professional - Windows Networking
http://mvp.support.microsoft.com

Steve Winograd's Networking FAQ
http://www.bcmaven.com/networking/faq.htm

Based on IP addressing. Also, most home-based routers will auto-assign one
full address length. So, from 0-256, and you can't use 0 or 256, and the
router assigns itself 1, that leaves 253.

Also, reading the specs on routers from all the major manufacturers verifies
253 as the maximum number the router can assign.
(such as here: http://www.linksys.com/products/prod...id=142&scid=29)

IP addresses are separated into classes, let's have a look at how they're
separated, then i'll explain why:
Range Class

0-127 A

128-191 B

192-223 C

Now to find out which class an IP address is in, you look at the first byte
of the address:

213.56.112.65

The first byte is 213. This falls inside the class range 192-223 making this
a class C IP address.

24.34.124.213

The first byte is 24. This falls inside the class range 0-127 making this a
class A IP address.

When I first read about IP addresses being split up into classes, the author
failed to explain why it was done, so i assumed this wasn't very important,
that someone just put them into classes on a Sunday while they were bored or
something. Take a look at the table below:

Range Class N/H

0-127 A N.H.H.H

128-191 B N.N.H.H

192-223 C N.N.N.H

N = Network I.D and H = Host I.D. In a class A IP address the first byte is
part of the network I.D. The remaining three bytes are part of the host I.D.
This means in class A there can be fewer networks and many hosts.

There can be 255 networks.

There can be 255 x 255 x 255 = 16,581,375 theoretical hosts on each class A
network.

In class B the first two bytes are part of the network I.D and the remaining
two are part of the host I.D. This means in class B there can be a
reasonable amount of networks, and a reasonable amount of hosts on each one
of those networks.

255 x 255 = 65,025 networks.

255 x 255 = 65,025 theoretical hosts on each of those networks.

I'm sure you get the drift, class C:

255 x 255 x 255 = 16,581,375 networks.

255 theoretical hosts on each of those networks.

I say "theoretical" hosts on each class of network, because on each network
the first address and the last address cannot be used (and the router
assigns itself an address as the Gateway). So if you work out the
theoretical number of hosts then take away two you have the actual number of
host computers that can be on that same network.

IP Address: 213.54.112.221

This is a class C address.

213.54.112.0 is the NETWORK address. This cannot be used as a hosts IP
address.

213.54.112.255 is the BROADCAST address on this network. It cannot be used
as a hosts IP address.

Class C networks have a theoretical 255 hosts on each network, 255 - 2 =
253. So on each class C network you can have 253 host computers.

Applying This

Let's go over a couple of examples to make this stick.

IP: 203.222.34.57

Is a class C address, N.N.N.H

The network ID is 203.222.34.0

The network broadcast address is 203.222.34.255

There can be 255 Theoretical hosts on this network.

Thus there can be 255 - 2 = 253 host computers functioning on this network.

IP: 24.240.11.99

Is a class A address, N.H.H.H

The network ID is 24.0

The network broadcast address is 24.255.255.255

There can be 255 x 255 x 255 = 16,581,375 Theoretical hosts on this network.

Thus there can be 16,581,375 - 2 = 16,581,373 host computers functioning on
this network.



"Steve Winograd [MVP]" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> In article <#(E-Mail Removed)>, "Carey Holzman"
> <(E-Mail Removed)> wrote:
> >"Benyamin" <(E-Mail Removed)> wrote in message
> >news:18e901c3baed$e40404f0$(E-Mail Removed)...
> >> Any body knows how many machines ( max.) we can connect
> >> in the one workgroup?
> >>
> >> thx...ben
> >
> >Should be 253.
>
> Where did you get that number, Carey? To the best of my knowledge,
> there's no theoretical limit to the number of machines in a workgroup.
>
> I haven't tested it, but I've read that the practical limit for good
> network browsing performance in a workgroup is around 20 machines.
> --
> Best Wishes,
> Steve Winograd, MS-MVP (Windows Networking)
>
> Please post any reply as a follow-up message in the news group
> for everyone to see. I'm sorry, but I don't answer questions
> addressed directly to me in E-mail or news groups.
>
> Microsoft Most Valuable Professional - Windows Networking
> http://mvp.support.microsoft.com
>
> Steve Winograd's Networking FAQ
> http://www.bcmaven.com/networking/faq.htm

On Fri, 5 Dec 2003 01:25:24 -0700, "Carey Holzman" <(E-Mail Removed)>
wrote:

>Based on IP addressing. Also, most home-based routers will auto-assign one
>full address length. So, from 0-256, and you can't use 0 or 256, and the
>router assigns itself 1, that leaves 253.

<snip>

All of which is completely irrelevant to how many machines are
theoretically allowed in a workgroup. The limitations of a particular
router have no bearing on it. If it can't issue enough addresses then
use something else as a dhcp server or use static ip's. Also there is
no need for classful addressing. That is at about 10 years out of
date.


Jim.

In article <(E-Mail Removed)>, "Carey Holzman"
<(E-Mail Removed)> wrote:
>> >> Any body knows how many machines ( max.) we can connect
>> >> in the one workgroup?
>> >>
>> >> thx...ben
>> >
>> >Should be 253.
>>
>> Where did you get that number, Carey? To the best of my knowledge,
>> there's no theoretical limit to the number of machines in a workgroup.
>>
>> I haven't tested it, but I've read that the practical limit for good
>> network browsing performance in a workgroup is around 20 machines.
>
>Based on IP addressing. Also, most home-based routers will auto-assign one
>full address length. So, from 0-256, and you can't use 0 or 256, and the
>router assigns itself 1, that leaves 253.
>
>Also, reading the specs on routers from all the major manufacturers verifies
>253 as the maximum number the router can assign.
>(such as here: http://www.linksys.com/products/prod...id=142&scid=29)
>
>IP addresses are separated into classes, let's have a look at how they're
>separated, then i'll explain why:
>
>[remainder snipped]

I asked you what time it is, and you answered by telling me how a
grandfather clock works. ;-)

IP addressing doesn't limit the number of computers that can be in a
workgroup to 253, because:

1. As you know, a workgroup can use IPX/SPX or NetBEUI instead of
TCP/IP.

2. As you noted, a class A subnet can contain millions of computers.
Anyone can set up a class A subnet using the private 10.0.0.0/8 IP
address range.

3. The limitations of home broadband routers and their DHCP servers
are irrelevant to the question. Would you define the maximum possible
speed for a car based on how fast cars with 4-cylinder engines and
automatic transmissions can go?
--
Best Wishes,
Steve Winograd, MS-MVP (Windows Networking)

Please post any reply as a follow-up message in the news group
for everyone to see. I'm sorry, but I don't answer questions
addressed directly to me in E-mail or news groups.

Microsoft Most Valuable Professional - Windows Networking
http://mvp.support.microsoft.com

Steve Winograd's Networking FAQ
http://www.bcmaven.com/networking/faq.htm

Well, I made some assumptions.

I assumed any computer on a network would want Internet access and therefore
would need TCP/IP.

I also assumed this person is a consumer or has a small office, because any
large organization has professionals hired to handle this for them.

I then assumed that this person is more than likely using a router to share
files and printers as well as Internet access and is curious how many more
PC's they can add.

I work with end-users every day in a consumer environment and get this
question quite often. Perhaps my assumptions were wrong, or perhaps they
were right.

Would the original poster "Benyamin" clarify this for us?

One final question: If I have a two computers on a workgroup using ONLY
TCP/IP, one at 130.13.165.13 and the other at 130.13.65.22, how can they see
each other? Even though their workgroup name is the same, I have never
figured out a way to do it except to add IPX./SPX, which you constantly
suggest that people don't add. Of course, you could add NetBEUI, but with it
being phased out by Microsoft, it's just a matter of time before the next MS
OS doesn't offer it at all.

Any way, like I said, home or small office user, already has a router, wants
to share internet, files and printers and already owns a router.

If that is the correct scenario, isn't 253 the limit?

Carey

"Steve Winograd [MVP]" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> In article <(E-Mail Removed)>, "Carey Holzman"
> <(E-Mail Removed)> wrote:
> >> >> Any body knows how many machines ( max.) we can connect
> >> >> in the one workgroup?
> >> >>
> >> >> thx...ben
> >> >
> >> >Should be 253.
> >>
> >> Where did you get that number, Carey? To the best of my knowledge,
> >> there's no theoretical limit to the number of machines in a workgroup.
> >>
> >> I haven't tested it, but I've read that the practical limit for good
> >> network browsing performance in a workgroup is around 20 machines.
> >
> >Based on IP addressing. Also, most home-based routers will auto-assign
one
> >full address length. So, from 0-256, and you can't use 0 or 256, and the
> >router assigns itself 1, that leaves 253.
> >
> >Also, reading the specs on routers from all the major manufacturers
verifies
> >253 as the maximum number the router can assign.
> >(such as here:
http://www.linksys.com/products/prod...id=142&scid=29)
> >
> >IP addresses are separated into classes, let's have a look at how they're
> >separated, then i'll explain why:
> >
> >[remainder snipped]
>
> I asked you what time it is, and you answered by telling me how a
> grandfather clock works. ;-)
>
> IP addressing doesn't limit the number of computers that can be in a
> workgroup to 253, because:
>
> 1. As you know, a workgroup can use IPX/SPX or NetBEUI instead of
> TCP/IP.
>
> 2. As you noted, a class A subnet can contain millions of computers.
> Anyone can set up a class A subnet using the private 10.0.0.0/8 IP
> address range.
>
> 3. The limitations of home broadband routers and their DHCP servers
> are irrelevant to the question. Would you define the maximum possible
> speed for a car based on how fast cars with 4-cylinder engines and
> automatic transmissions can go?
> --
> Best Wishes,
> Steve Winograd, MS-MVP (Windows Networking)
>
> Please post any reply as a follow-up message in the news group
> for everyone to see. I'm sorry, but I don't answer questions
> addressed directly to me in E-mail or news groups.
>
> Microsoft Most Valuable Professional - Windows Networking
> http://mvp.support.microsoft.com
>
> Steve Winograd's Networking FAQ
> http://www.bcmaven.com/networking/faq.htm

On Fri, 5 Dec 2003 04:47:32 -0700, "Carey Holzman" <(E-Mail Removed)>
wrote:

>One final question: If I have a two computers on a workgroup using ONLY
>TCP/IP, one at 130.13.165.13 and the other at 130.13.65.22, how can they see
>each other?

You can change the subnet mask so that they are both on the same
network eg 255.255.0.0

Jim.

In article <(E-Mail Removed)>, "Carey Holzman"
<(E-Mail Removed)> wrote:
>Well, I made some assumptions.
>
>I assumed any computer on a network would want Internet access and therefore
>would need TCP/IP.
>
>I also assumed this person is a consumer or has a small office, because any
>large organization has professionals hired to handle this for them.
>
>I then assumed that this person is more than likely using a router to share
>files and printers as well as Internet access and is curious how many more
>PC's they can add.
>
>I work with end-users every day in a consumer environment and get this
>question quite often. Perhaps my assumptions were wrong, or perhaps they
>were right.

>Would the original poster "Benyamin" clarify this for us?

Regardless of whether your assumptions right or wrong, you should
have clearly stated your assumptions along with the answer.
Otherwise, your answer can mislead people.

>One final question: If I have a two computers on a workgroup using ONLY
>TCP/IP, one at 130.13.165.13 and the other at 130.13.65.22, how can they see
>each other?

Without knowing what the subnet mask is, it isn't possible to say
whether computers with those IP addresses can see each other.

However, those are public IP addresses. If they were assigned by an
Internet service provider and are connected to the Internet, it
probably isn't safe to use TCP/IP for file sharing in the first place.

>Even though their workgroup name is the same, I have never
>figured out a way to do it except to add IPX./SPX, which you constantly
>suggest that people don't add.

I tell people not to add IPX/SPX when their computers have private IP
addresses and they can't get TCP/IP to work. It's better to find and
fix the TCP/IP configuration problem than to add a protocol.

I tell people to use IPX/SPX when it's appropriate: their computers
have public IP addresses and they want to prevent other Internet users
from accessing their network. I quote from the section of the Windows
XP Help and Support file that recommends IPX/SPX in that specific
case, and I tell them how to un-bind file sharing from TCP/IP so that
only one protocol is used.

>Of course, you could add NetBEUI, but with it
>being phased out by Microsoft, it's just a matter of time before the next MS
>OS doesn't offer it at all.

I agree.

>Any way, like I said, home or small office user, already has a router, wants
>to share internet, files and printers and already owns a router.

My D-Link DI-604 router lets me specify the subnet mask. I just
changed it to 255.255.0.0, which would allow 65,533 computers to
connect with each other in a workgroup.

>If that is the correct scenario, isn't 253 the limit?

Possibly. And if all cars had 4-cylinder engines and automatic
transmissions, then the fastest possible speed for a car might be 95
mph.

>Carey
--
Best Wishes,
Steve Winograd, MS-MVP (Windows Networking)

Please post any reply as a follow-up message in the news group
for everyone to see. I'm sorry, but I don't answer questions
addressed directly to me in E-mail or news groups.

Microsoft Most Valuable Professional - Windows Networking
http://mvp.support.microsoft.com

Steve Winograd's Networking FAQ
http://www.bcmaven.com/networking/faq.htm

In article <(E-Mail Removed)>
"Carey Holzman" <(E-Mail Removed)> wrote:

Another clock. This is exactly what I stated earlier in the week. To
arbitrarily throw out information like "253" is opening the poor guy with
the question to all kinds of, maybe not misinformation as such, but at
least aggravation if he tries what you suggest and fails. Information
must be accurate and based on a working knowledge. Otherwise, it's just
useless noise as so much of the internet is.