Page 1 of 2

Posted: Fri Feb 01, 2002 6:06 pm
by combo
Hi, could any of these be valid ip addresses?

I belive they can not have all 0's 1's, end in 255 or end in 0. Is this correct?

197.25.0.255
150.25.0.0
1.1.1.1
0.0.0.0
129.1.1.0
127.188.18.5

Posted: Fri Feb 01, 2002 6:19 pm
by RandomNull
well im not good at networking but my "guess" and let me emphasize guess would be that all of em could be valid cept for the 0.0.0.0.
I may b wrong I may be right either way I will find out as will you when Forge or some other know-it-all person comes in.

Posted: Fri Feb 01, 2002 7:00 pm
by DiMaestro
If your building a home network, follow the 192.168.*.* or 10.*.*.* rule.

Posted: Fri Feb 01, 2002 7:04 pm
by combo
This is all theory, so Im not going to be using any of these. I just want to get the rules correct of what could be valid ip on the public internet.

Posted: Fri Feb 01, 2002 11:58 pm
by SecretSquirrel
197.25.0.255 is a valid IP. It is the broadcast address (goes to all hosts) for the Class C subnet 197.25.0.0

150.25.0.0 is a valid IP. It is the network address (used for routing) for the Class B network 150.25.0.0

1.1.1.1 is a valid IP. It belongs to the Class A subnet 1.0.0.0

129.1.1.0 is a valid IP. It belongs to the Class B subnet 129.1.0.0.

127.188.18.5 is a valid IP, but not valid on the internet. It belongs to the Class A subnet 127.0.0.0 which is reserved for traffic local to your host.

---------------------

The only one not valid for the internet is the 127.188.18.5 address. Following are the ip ranges that are not valid on the open internet:

10.0.0.0 - 10.255.255.255
172.16.0.0 - 172.31.255.255
192.168.0.0 - 192.168.255.255
127.0.0.0 - 127.255.255.255
224.0.0.0 - 255.255.255.255

The first three are designated as private networks. The fourth is the loopback network and the fifth set is reserved for multicast and experimental purposes.

Posted: Sat Feb 02, 2002 1:50 am
by combo
Thanks SecretSquirrel!!

Posted: Sat Feb 02, 2002 4:02 am
by Bruce
Squirrel, you forgot the 169.254.0.0/16 netblock - also ingress and egress filtered by most ISP's.

Bruce

Posted: Sat Feb 02, 2002 5:02 am
by SecretSquirrel
Non of the docs I had hany mentioned that netblock. Guess I'll have to update the docs. As for ISP filtering... well they are still valid IPs even if the ISPs don't like them.

Posted: Sat Feb 02, 2002 4:30 pm
by Bruce
Squirrel: in terms of docs I quite like this one. The 192.168.0.0/16 et al are also valid, they're just not to be routed outside an AS.

If you wanted to be pedantic you could say that 197.25.0.255 is not a valid IP address - it's a network broadcast address and not a unicast IP address.

Bruce

Posted: Sun Feb 03, 2002 1:43 am
by SecretSquirrel
On 2002-02-02 15:30, Bruce wrote:
Squirrel: in terms of docs I quite like this one. The 192.168.0.0/16 et al are also valid, they're just not to be routed outside an AS.

If you wanted to be pedantic you could say that 197.25.0.255 is not a valid IP address - it's a network broadcast address and not a unicast IP address.

Bruce


No arguements there. It centers around the meaning of valid. The original Q was valid on the internet I believe. Which the 192.168.0.0/16 addresses are not. As far as the broadcast and network address thing goes, I pondered over it for a bit while I was writing my original reply. I decided that since you could send and packet via IP to those addresses and expect something to happen to it (even if its just dropped) that they qualified as valid adresses. They are not however valid node addresses, I agree.

Posted: Sun Mar 03, 2002 10:41 pm
by Speed
Let's look at each of these:

197.25.0.255 -- Broadcast to 197.25.0.0/24
150.25.0.0 -- This is a network, not a single IP address.
1.1.1.1 -- This is an IP address.
0.0.0.0 -- This is the entire network.
129.1.1.0 -- Another network.
127.188.18.5 -- An IP address.

Notice that if it ends in 0 that it's not an IP <i>address</i>. And if it ends in 255 (all ones) that it's a <i>broadcast</i> address. A broadcast address really means "send to all hosts on this subnet" more than anything else. It's a virtual address, so you can't put a host at that address. A network number and mask (like 197.25.0.0/255.255.255.0 or 197.25.0.0/24 in shorthand) specifies a range of IP addresses in a network. 0.0.0.0 is used in routing to signify the default route, where all packets go unless there's another rule. These days, 0.0.0.0 is synonymous for the Internet.

Posted: Mon Mar 04, 2002 4:11 am
by Bruce
Not an accurate description: 10.10.10.0 is an IP address, if no subnet mask is specified. Just because it ends in zero doesn't make it a network address.

Posted: Mon Mar 04, 2002 5:51 am
by Speed
Sorry Bruce, but that zero <i>does</i> make it a network and not an address. The .0 and .255 suffixes are reserved. Read the RFCs, or take an IP networking course for details.

Posted: Mon Mar 04, 2002 6:05 am
by Bruce
I'll stand by my statement. In the example I gave (10.0.0.0/8), the ONLY network broadcast address is 10.255.255.255 and the ONLY network address is 10.0.0.0. If you subnet 10.0.0.0/8 then yes, you may have subnet broadcast addresses - but in this example I excluded subnetting.

You can also end up with weird scenarios if you use supernetting.

Feel free to correct me, but please provide a reference.

Bruce

Posted: Mon Mar 04, 2002 6:34 am
by Speed
Bruce, I can't discuss this with you as long as you don't understand the difference between a network address and a network. There are plenty of good tutorials on IP networking out there. Check them out.

Posted: Mon Mar 04, 2002 7:23 am
by Bruce
I've been dealing with IP for a few years now (on Ciscos. For an ISP.) - like I said, I'll stick with my answer, please provide a reference for what you're saying.

Posted: Mon Mar 04, 2002 8:33 am
by Despite
I too have worked with IP for a long time, and I'm going to agree with you, Bruce. The example you stated does make sense, both mathematically & logically. However, I have not read the RFCs, so I will concede that it <i>may</i> be convention to never use .0 in any octet of a host address, even if it would work logically. Therefore, I too am calling on Speed to provide the evidence for his statement.

Posted: Mon Mar 04, 2002 8:43 am
by Bruce
Whenever you see an exam question like "How many valid IP addresses can be assigned using the network 202.49.46.0" - and no further information given - you assume that the network is not subnetted. The answer to the question is 2^hostbits - 2, or (in this particular example) 254 valid IP addresses.

In the case of 10.0.0.0 you assume a /8 netmask, and there are (2^24)-2 valid IP addresses, or 16,777,214. The only way this can work is if, say, 10.10.10.0 and 10.10.10.255 are valid IP addresses.

Now to clarify my original point: just because the last octet is .0 or .255 does NOT necessarily mean that it's not a valid, routable IP address - you have to know the netmask associated with it. The network address has all the host bits set to '0'. The broadcast address has all the host bits set to '1'.

Or have I missed something? Am I losing it at the ripe old age of 27?!?

Bruce

Posted: Mon Mar 04, 2002 6:23 pm
by Speed
Sorry guys, but I'm not here for pissing contests. If you don't recognize the significance of a 4th octet full of zeroes or ones, that's not my failing. Like I said, the RFCs or a good networking course will teach you what you need to know. And since the RFCs are the ultimate authority on the matter, I don't have anything to offer that they don't.


<font size=-1>[ This Message was edited by: Speed on 2002-03-04 17:24 ]</font>

Posted: Mon Mar 04, 2002 8:03 pm
by Despite
I suspected as much.

Posted: Tue Mar 05, 2002 4:09 am
by Bruce
I'm not here for the pissing contest either. Either I've misunderstood something (and would love to be corrected), or we're talking about the same thing from different angles. Shame you can't even specify which RFC you're referring to, either by name or number.

Posted: Tue Mar 05, 2002 5:52 am
by Speed
Bruce, I can specify several RFCs that cover the subject. But I have no interest in teaching courses for free. (There are better ways to ask, people.) Here are a few to look over:

RFC 0790 - ASSIGNED NUMBERS
RFC 0791 - INTERNET PROTOCOL
RFC 0796 - ADDRESS MAPPINGS
RFC 0814 - NAME, ADDRESSES, PORTS, AND ROUTES
RFC 0917 - INTERNET SUBNETS
RFC 0922 - BROADCASTING INTERNET DATAGRAMS IN THE PRESENCE OF SUBNETS
RFC 0932 - A SUBNETWORK ADDRESSING SCHEME
RFC 0936 - Another Internet Subnet Addressing Scheme
RFC 0940 - Toward an Internet Standard Scheme for Subnetting
RFC 0950 - Internet Standard Subnetting Procedure

The documents may be found at http://www.ietf.org/rfc/ This is only in the first 1000, so you may find more of interest, but you're going to have to learn to find data for yourself.

Posted: Tue Mar 05, 2002 7:19 am
by Bruce
Maybe I should have been more clear - I'm very familiar with the RFC's, I just wanted to know which you wanted me to read.

I went and eyeballed RFC 1878 after my earlier post, which I think is the current RFC detailing IPv4 subnetting. Have a look at the link (especially the first entry in Table 1-1) and see if you can see why I hold my point of view - 32766 hosts per subnet means you will end up with valid IP addresses where the last octet is 255.

BTW, I'm not after a course (free or otherwise). When information conflicts with my understanding of something I'd like to know why - I don't like being wrong, my work depends on being able to give people accurate information. My apologies if I came across harsh - I justed wanted more than an RTFM answer, when I think I understand the FM fairly well. You've obviously been around IT for a while, so I'd value your input.

Posted: Tue Mar 05, 2002 8:15 am
by Speed
Bruce, I'm sorry if I come off as being evasive. But as you can see, the subject is quite involved. There's enough material there to fill up a semester course. I simply have no interest in going into that detail over such a trivial matter.

I have yet to see a valid host number that isn't divisible by one. The early RFCs say why.

Posted: Tue Mar 05, 2002 8:29 am
by Despite
I can't believe you're being so polite to this guy, Bruce. As far as I'm concerned, you're right on this deal. He comes in and tells you that you're wrong, but won't supply the evidence to prove his assertion. The onus is fully his. It is my belief that he hasn't quoted us the relevant piece of the RFC because it doesn't exist. I personally won't waste the time needed to scour through the RFCs. (Y)Our side of the argument is complicated by the fact that what we need to prove is that something (the part of the RFC proscribing use of all 1s or 0s in an octet) does NOT exist. As any scientist will tell you, that's a really tall order. Therefore, I will once again call on Speed to quote to us the relevant portion of the proper RFC.

Posted: Tue Mar 05, 2002 8:33 am
by Bruce
OK, I'll bite - time to do some reading. I'll get back to you when I've digested the superseded RFC's - I've generally skipped them in favour of the ones they've been superseded by.

BTW, as for my level of knowledge: probably between CCNA and CCNP. I've read Comer, Stevens vol 1, most of Todd Lammle's Cisco books. I've got patchy knowledge (eg, I've used BGP but not OSPF) but good understanding of what I've dealt with. So it's frustrating to think I've misunderstood something this basic.

Bruce

Posted: Tue Mar 05, 2002 8:36 am
by Bruce
Despite: maybe I'm just a nice guy :smile:

And as Speed has been around IT for a while (see here - the earliest Linux kernel I've ever seen is 1.3.something) I'll give him the benefit of the doubt.

<font size=-1>[ This Message was edited by: Bruce on ]</font>

Posted: Tue Mar 05, 2002 8:52 am
by Despite
Bruce, see you in a week or two! And hey, when you finally slog through all that text and don't find what he's referring to, I hope you come back here and stick it to him but good for this! By the way, would you like to take a bet that Speed will just claim that you missed the part he was talking about?

Posted: Tue Mar 05, 2002 9:01 am
by Speed
LOL... I'll have to tell you about my early experiences of writing code on a genuine Teletype terminal with paper tape and a real bell. Or the joys of hauling around Hollerith cards. I've actually done those things for real. Man I feel old!

Bruce, I give you credit for being a gentleman. Again, I'm not making this a pissing contest (which would only annoy Despite). I'm certain that a host address must be >0, and I remember getting that info from an RFC around 900, and that's all I remember. I could be wrong. But I have bigger fish to fry right now.

BTW, how is the website performing for you? Any lags or errors?


<font size=-1>[ This Message was edited by: Speed on ]</font>

Posted: Tue Mar 05, 2002 10:12 am
by Bruce
I work with a guy who's been in IT for 30+ years - used to work for ICL. He's also worked with Hollerith cards, and has some interesting stories to tell.

As for lags - a bit. No errors, but occasionally I'll wait 30-odd seconds for a page to load.

Bruce