Quick Subnetting Question.

[Saltin]

Hi guys, quick question. I'm pretty sure I've got the answer, but though I would come to you for confirmation. I can't afford to make any mistakes

The block of Ip's = w.x.y.0/26 (255.255.255.192 mask)
This allows 2^6-2 hosts for a total of 62 hosts.
What is the range of IP's I can employ here?

Thanks!

 

[cleverhandle]

Well, I wouldn't rely on my opinion, but I'd say...

If w.x.y.0 is the network portion of your addresses, you've got a Class C network with 2 bits of subnetting, allowing 2^2-2 = 2 subnets, and 2^6-2 = 62 hosts per subnet. For the subnet bits, you can't use 00 (the network), or 11 (the broadcast), so that leaves you 01 (w.x.y.64) or 10 (w.x.y.128) as the possible subnets. Which means possible host addresses are w.x.y.65-126 or w.x.y.129-190, assuming you're allowed to use either subnet.

Did I do good, guys?

Edit: second range ends at 190, as stated below. Arithmetic seems to be beyond me.

 

[GigaCluster]

11111111 11111111 11111111 11000000

2^6 = 64 possible IPs in the host section
.0 is the network ("wire") address of the current subnet.
.64 is the next subnet's network address
.63 is the current subnet's broadcast address.

Hence, the range of usable hosts on this subnet is from x.x.x.1 to x.x.x.62.

 

[GigaCluster]

Damn, I am in a CCNA class... I should know better. Sorry, had a brainfart.

Okay. Since you've not told us what IP class this is, I will assume that it's class C.

In that case, we have two subnetted bits, and since 00 and 11 cannot be used (one's network and one's broadcast), that leaves us with two subnets: 01 and 10.

The first IP range is 01000001 to 01111110. (x.x.x.65 to x.x.x.126)
The second IP range is 10000001 to 10111110. (x.x.x.129 to x.x.x.190)

That should be correct.

 

[Saltin]

Thanks for the confirmation.

 

[ScottMac]

The class of the address is NOT determined by the mask (dammit!).

The class of the address is determined by the leading (most significant) bits of the address. The mask you provided can be applied to ANY class address (well, A,B, or C anyway) and they would STILL be class A, B, or C.

01 = Class A (0-127) 00000000 - 01111111
10 = Class B (128-191) 10000000 - 10111111
11 = Class C (192-223) 11000000 - 11011111
111 = Class D (224-239) 11100000 - 11101111
1111 = Class E (240-255) 11110000 - 11111111

This is a Cisco test question.....learn it and remember well......

(BTW: There's a couple Miscrosoft "White Papers" on their website where the author(s) make some reference to the mask determining class, but you'd expect that kind of ignorance from Microsoft .... networking folks should know better).


FWIW

Scott

 

[Saltin]

I didnt mention the class because it's not important to the question. The fellas were obviously making educated guesses based on the subnet mask. No harm done there.

The range I came up with is w.x.y.1 >>>> w.x.y.63. Was a simple problem I had the answer to before hand, but wanted second opinions b/c I don't want to look like a tool at my new job

Like ScottM said, it's not actually a Class C addy, it's Class A. Regardless of the class though, the range available doesnt change.

Someone correct me if I am wrong.

Thanks for the input.

 

[cleverhandle]

<< Like ScottM said, it's not actually a Class C addy, it's Class A. Regardless of the class though, the range available doesnt change. >>



Again, not a pro, but I'm pretty that the class makes a big difference. From your description (not from the mask, ScottMac ) of your address block as w.x.y.0/26 I was assuming you had a Class C address, and thus 2 bits of subnetting. If you really have a Class A address, you have 18 bits of subnetting - that's a helluva lot of possible subnets (2^18-2), and thus a correspondingly helluva lot of possible host addresses. But I would imagine that if your setting up machines on a network of this size, you should have a subnet assigned to you, unless it's a private 10.x.x.x address, in which case someone probably picked a needlessly huge private network. If it's a 10.x.x.x network, you could use any host address whose last three octets, in decimal, are NOT a multiple of 64 or a multiple of 64 minus one, except for x.0.0.1-63 and x.255.255.192-255.

 

[Saltin]

<< If you really have a Class A address, you have 18 bits of subnetting - that's a helluva lot of possible subnets (2^18-2), and thus a correspondingly helluva lot of possible host addresses. >>



I'm not sure where you are getting these number's from.
It isnt a private 10. addy. This is a public Class A range /26
When you get assigned a block of IP's and they tell you the addy is w.x.y.0/26, you have (2^6)-2 hosts. The subnet mask HAS to be 255.255.255.192
I don't get to play with the first two bits in the fourth octet.

 

[cleverhandle]

I could continue to offer my opinion, but I'd prefer to let the pros handle this, since it has real world importance to you.

<Respectfully bowing out>

 

[Schoolies]

Cramsession has some good material on subnetting:

Quick and Dirty Guide to Subnetting
Leran to Subnet Pt.1
Learn to Subnet Pt.2

 

[GoldMember]

I need to learn this stuff guys.. Where can I learn it? Web resource? Book? I'd prefer to extract the data from the net.. but either way. I mean learn all of your network stuff about the sub nets and stuff with your 2^6 and what not..

 

[nightowl]

Saltin: Check out this link to verify what you have, but your usable addresses would be between w.x.y.1 and w.x.y.62. w.x.y.0 cannot be used since it is the network address and w.x.y.63 cannot be used since it is the broadcast address for the network. If the last octet can be subnetted by you then you will have 4 networks of 62 hosts. But, since you said w.x.y.0/26 I am assuming that is your network number and you need to determine the correct range of the node.

Also, the subnet does not determine the class of the address as Scottmac said. A class A address range can be subnetted into smaller network effectively making it a Class B, C, etc. address but it is still a Class A address.

 

[ScottMac]

I'm not nitpicking....this is more of an FYI in case some of the newer networking folks haven't seen / heard / read it:

The Zero subnet and the 255 subnets (and in many cases, the host address) are discarded by CONVENTION rather than a hard rule, regulation, or LAW.

The convention stems from the early days: AT&T Unix (System IV, I think) used to broadcast with an all-ones address, the BSD Unix systems used to broadcast all zeros.

There is no technical reason to not use the first & last subnet or host address (zero & 255), but it is the defacto standard, by convention (general agreement), and from that....damn-near set in concrete. Cisco routers can be set to use the Zero Subnet; many organizations use it for their WAN point-to-point links (with a /30 mask)....then they're not "throwing away" an address block that'll probably never see public light....

And, for those that are doomed to take a Cisco test anytime in the future, there are questions that ask "How Many hosts (or subnets) in this block?"...they mean TOTAL (includes the Zero and 255). If they don't want the TOTAL number, they'll ask for USABLE subnets (or host addresses.


Anyway, bottom line: You CAN use the zero and the 255, but it would be against the current convention (people would point and laugh at you if they found out).


FWIW

Scott

 

[GoldMember]

Please tell me where to learn all this fancy sub-netting stuff

 

[nightowl]

ScottMac, I was told that you could not use the 0 and 255 networks because RIPv1 did not support it and that any of the newer routing protocols that included the extended network prefix could handle it. That was the on restriction I have ever heard on not using those 2 networks. Other than that I have always thought that they were fair game unless you were using RIPv1 somewhere on your network.

Edit: As for where I have picked up my subnetting knowledge, I have really come to understand it by taking classes here at Purdue this past year. Also, I had a good deal of backround knowledge that made it easier for me to fully understand it this year.

 

[ScottMac]

It's true that some protocols have adopted the convention which would limit the use of zero and 255. RIPv1 doesn't do anything with net/subnet masks; I've never tried it, but it should work OK with a zero subnet. It almost certainly doesn'twork with the 255 because RIPv1 broadcasts with an all ones broadcast.

I *-think-* that routed CAN be configured either way so, conceivably, you could set up a UNIX station as a 255 subnet-compatible router.

As for learning this stuff, check out "Networking With TCP/IP" by Doug Comer (last I heard, he's a CS prof at Purdue). It's one of the "bibles" for TCP/IP networking. It's not a cheap book, but it's one of the better references available.

FWIW

Scott

 

[Mucman]

ScottMac - Thanks for hurting my feelings! We use zero subnets 'round here , not my fault... the Cisco 7200 had zero_zubnet enabled by default and our old admin kept it that way...

 

[GoldMember]

I have a TCP/IP MCSE training guide here.. subnetting be in there? Any good net sites that kinda summarize it?

 

[Xuttah]

Learning the theory is essential for Cisco tests, and handy in a bind. However, once you get out of class and get into managing IP space, this tool is a godsend: WildPackets

Not saying don't learn the stuff (at least at first). After a while it'll be second nature, but when trying give one group a /27 out of a /23 on the spot, you REALLY appreciate it.

 

[ScottMac]

Awww nuts muc....

Ya gotta remember I'm getting flakey in my old age (I'm not just over-the-hill....I can't even see the hill behind me...)

I just forgot to put in the part about how Cisco includes in its router-oriented class how to use the zero subnet, and actually encourages it. It's a neat thing. I use it m'self from time-to-time....

And Gawd knows how many 192.168.0.X networks are operating at the hands of the folks on the board here...we even got in a mildly heated discussion about it (about a year ago). My position was that it wasn't good to use 'em, because some products (many from the Great Northwest of these here U-nited States) have difficulty with a zero net/subnet.

We got the usual zillion posts saying "I'm doing it and it's working....so go kiss a cow.." kinda responses and the matter was dropped with the usual "My mind is made up, don't confuse me with the facts" kinda discussion ...and here we are.

Cisco is specifically rigged to handle zero subnets, should you decide to use 'em. It's OK. Heck, it's probably a Good Thing on a Cisco box (but not to a PC / workstation / server ...just router to router, router to switch , etc....).

FWIW

Scott

 

[Mucman]

ScottMac - Hey, no hard feelings! Didn't you see the winky? I too would prefer not to have it setup this way, but my boss (who setup the network) didn't know any better at the time and figured that since Cisco had it on default, you might as well use it . Our router speaks BGP to neighboring routers so I doubt any problems with RIP will interfere.

I am all about the standards! Heck, I even shown some of your posts to my boss to get my point across

I have another question to ask though... Who made these terms (class A,B,C,D,E) and what do they mean in the first place? Is there some sort of history behind it? I am not talking about what they are, but how they became to being. I think you clearly explain what a class A network address is, and so on....

 

[ScottMac]

The folks that developed TCP/IP did so because the government (and mostly government research agencies and contractors) had a bunch of different computer systems, made by bunches of different vendors, none of which spoke the same language. THey figured out that the only way to efficiently get information from one research group to another, they developed TCP/IP, and made it mandatory for participation in the programs (there were lots of mandatory things...like ADA as a programming language...)

Since DARPA (being a government agency) has no idea at the time that TCP/IP would ever be a public resource. AND, since it was not developed to be a marketable commodity AND since it was developed by engineering types (NOT driven by a marketing group)...they stretched their collective imaginations and came up with the classifications based on the wildest stretches of their imagination...."Class A, Class B, Class C ..."

The government has a history of imaginative labels..."Toilet Seat, White, Plastic" "Men's Shirt, Camoflaged, Green" and so on.

If the government had a marleting group driving the effort, we would have had something completely different (with a Class A being something like "New, Improved, Ultra Turbo Mega Block" (of Addresses).

That's my take on it,

FWIW

Scott