Re: How do I determine subnet manually? in Networking http://www.dslreports.com/forum/r20451580 en Thu, 08 Jan 2009 02:54:19 EDT Thu, 08 Jan 2009 02:54:19 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20456937 tschmidt :
said by IamGimli See Profile :

I understand binary but how is hexadecimal any less confusing than decimal to represent network addresses? It may make them a bit shorter to write but I don't see any difference in the confusion aspect.
Dotted decimal represents a discontinuous counting method. If I add 1 to FFh I get 100h. If I add 1 to 0.255 (dotted decimal) I get 1.0. Life gets even more exciting if I start playing with masking lower order bits.

If I want to mask the 2 low order bits I have
1111 1100 b (binary)
FCh (hex)
252 Decimal

Perhaps I'm unique (after all this is not my day job) but I find it impossible to do arithmetic in dotted decimal.

/tom]]> http://www.dslreports.com/forum/remark,20456937 Fri, 09 May 2008 12:52:12 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20456100 IamGimli :
said by tschmidt See Profile :

If you stick with binary or hex it is a lot less confusing.
I understand binary but how is hexadecimal any less confusing than decimal to represent network addresses? It may make them a bit shorter to write but I don't see any difference in the confusion aspect.]]> http://www.dslreports.com/forum/remark,20456100 Fri, 09 May 2008 10:22:29 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20454234 NetFixer :
said by verifone411 See Profile :

(one for network, one for broadcast, and one for gateway [optional]).

I am guessing...

network = .0
broadcast= .255
gateway =.1

?

thanks
This is also a popular combination for a /24 network:

network = .0
broadcast= .255
gateway = .254
--
We can never have enough of nature.
We need to witness our own limits transgressed, and some life pasturing freely where we never wander.
Test your firewall.]]> http://www.dslreports.com/forum/remark,20454234 Thu, 08 May 2008 22:12:47 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20453928 verifone411 : (one for network, one for broadcast, and one for gateway [optional]).

I am guessing...

network = .0
broadcast= .255
gateway =.1

?

thanks]]> http://www.dslreports.com/forum/remark,20453928 Thu, 08 May 2008 21:18:14 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20453696 tschmidt : Each octet is 8-bits. A /24 address means 24-bits of address are reserved for network leaving 8 for host. This assumes you are talking about IPv4 with a 32-bit address range.

An Octet is FFh or 256 (in decimal notion) However, since the lowest address is reserved as network address and highest for broadcast that leaves 254 available host addresses.

Dotted Decimal was a convenient way to represent IP address when Class-full addressing was used since Class boundaries were also octet boundaries. With the advent of classless routing where split between network and host address can occur anywhere dotted decimal is a little like using Roman numerals to build a house.

If you stick with binary or hex it is a lot less confusing.

/tom ]]> http://www.dslreports.com/forum/remark,20453696 Thu, 08 May 2008 20:18:51 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20453349 verifone411 : Thanks for the infomation. Sticking with the first class b number (does that change things since you decided to move the example to a class c?)

How do you know and how do you determine that...

/24 = 256 IPs = 254 Usable ? i know the diagrams say that.

Is it because the last 2 sets of octects represent 8 bits each for a total of 16. Each 16 bit placeholder can have 2 different values so that is why you raise it to the power of 2? to get 65536- 2=65534 usable?]]> http://www.dslreports.com/forum/remark,20453349 Thu, 08 May 2008 18:46:15 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20451815 IamGimli : In binary, each position is an exponent of 2 (instead of an exponent of 10 in the decimal system). The first position is therefore worth 1 (2^0), the second position is worth 2 (2^1), the third 4 (2^2), etc. From that it's a bit of mental work to determine the whole number. For 141, for example, what whole exponents of 2 are represented? You got 128 (2^8) or 10000000 in binary, plus 8 (2^4) or 1000, plus 4 (2^3) or 100, plus 1 (2^0) or 1.

bin 10000000 + 1000 + 100 + 1 = 10001101
dec 128 + 8 + 4 + 1 = 141
hex 80 + 8 + 4 + 1 = 8D (disregard that, I'm just trying to confuse you with hexadecimal i.e. base 16 now ;))

BTW even the Windows calculator can easily translate from decimal to binary values. Just switch it to scientific mode and you'll see the Hex, Dec, Oct and Bin radio button on the left side. Enter a number while the radio button is on Dec, click on the Bin button and see it represented in binary. Just remember to add the leading zero(s) if your value is less than 128 and you are representing an IP address. Each of the four bytes in the IP address needs to be represented by 8 binary positions or bits (hence why you represent 67 with 01000011 instead of just 1000011).]]> http://www.dslreports.com/forum/remark,20451815 Thu, 08 May 2008 13:46:03 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20451599 anon : 141.67.0.0 is represented as 10001101.01000011.00000000.0000000

How do you convert to binary without a calculator that I see all over the net?]]> http://www.dslreports.com/forum/remark,20451599 Thu, 08 May 2008 13:06:08 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20451580 IamGimli : Since you asked, here's the "how".

A subnet mask determines how many bits of your IP address are static for a particular subnet.

Let's use your class B as an example. 141.67.0.0 is represented as 10001101.01000011.00000000.0000000 in binary. The subnet mask of your class B is /16, because the first 16 bits of that address will never change. When you want to further separate that class B in smaller subnet you remove more bits that will not change for that subnet.

For example, koitsu's suggestion of /20 subnets to get the most subnets out of your class B while having at least 4000 addresses in each subnet makes your first subnet 10001101.01000011.0000xxxx.xxxxxxxx in binary. That means that subnet includes every addresses between 10001101.01000011.00000000.00000000 and 10001101.01000011.00001111.11111111 (141.67.0.0 to 141.67.15.255). Your second subnet would be 10001101.01000011.0001xxxx.xxxxxxxx (141.67.16.0 to 141.67.31.255), your third 10001101.01000011.0010xxxx.xxxxxxxx (141.67.32.0 to 141.67.47.255), etc.

The x represent every variable bit in the addresses for that subnet.

If you went with a /22 subnet mask instead, the mask would look like 10001101.01000011.000000xx.xxxxxxxx and the addressable range in your first subnet would be 10001101.01000011.00000000.00000000 to 10001101.01000011.00000011.11111111 (141.67.0.0 to 141.67.3.255).]]> http://www.dslreports.com/forum/remark,20451580 Thu, 08 May 2008 13:02:36 EDT Re: How do I determine subnet manually? http://www.dslreports.com/forum/remark,20450057 koitsu : I think this was posted to the wrong forum. Should be over in one of the Networking or Advanced Networking forums. :-)

Regardless, an attempt answer your question -- and it's hard to answer because I'm not completely sure I understand the question fully:

You'll want to split 141.67.0.0/16 into separate subnets that offer enough IP space to guarantee at least 4000 (usable) IPs per block. That's fairly easy to figure out. Let's start with a /24 (what you call a "Class C") and break it down from there:

/24 = 256 IPs
/23 = 512 IPs
/22 = 1024 IPs
/21 = 2048 IPs
/20 = 4096 IPs

So what you'd have is multiple /20 subnets, guaranteeing 4093 usable IPs out of each of those netblocks (lost 3 IPs go to gateway, broadcast, and network).

How many of those subnets would you have? Well, that's simple: A /16 gives you 65536 IPs, so do some simple math:

65536 / 4096 = 16

You'd have a total of 16 /20 netblocks to use. Thus, your segregated netblocks would be:

141.67.[[16*0 through 16*15]].0/20, e.g.:

141.67.0.0/20
141.67.16.0/20
141.67.32.0/20
141.67.48.0/20
141.67.64.0/20
141.67.80.0/20
141.67.96.0/20
141.67.112.0/20
141.67.128.0/20
141.67.144.0/20
141.67.160.0/20
141.67.176.0/20
141.67.192.0/20
141.67.208.0/20
141.67.224.0/20
141.67.248.0/20

Now, you lose 3 IPs per /20 (one for network, one for broadcast, and one for gateway [optional]). So ultimately you're "losing" 48 IPs due to the subnetting of the /16.

Hope this makes sense.

--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.]]> http://www.dslreports.com/forum/remark,20450057 Thu, 08 May 2008 06:34:11 EDT How do I determine subnet manually? http://www.dslreports.com/forum/remark,20450032 verifone411 : This is more of a question of how than what. Since there are calculators on how to do this on the internet.

If one has been assigned a Class B address (141.67.x.x), and you need to be able to designate up to 4,000 hosts per subnet. What subnet mask specification would allow for the most possible subnets? How did you determine your subnet mask specification?

Excluding reserved addresses, how many addresses would be available for host assignment in each subnet? In total for all subnets?

What would be the last assignable address of the first subnet, and what would be first assignable address of the last subnet?]]> http://www.dslreports.com/forum/remark,20450032 Thu, 08 May 2008 06:00:13 EDT