While perusing the textbook for my Intro to Microcomputers class, I came upon this gem, which will probably be on my test Friday :
“Each location on the Internet has a four-part numeric address called an IP (Internet Protocol) address. The first part of the IP address identifies the geographic region, the second part the company or organization, the their part the computer group, and the fourth and last part, the specific computer”
Say what?
I probably wouldn’t have noticed this if I hadn’t read a huge book on TCP/IP before I started school. According to the info I have, the first octet denotes the class of the network, and the other three either further identify the network, or identify the host, depending on which class the address is.
Now I fully acknowledge that I may be incorrect, and so may both of the TCP/IP books I consulted. I’m thinking that this isn’t the case, though.
My question : is my cruddy textbook correct in ANY way, shape, or form on this issue? If not, where the hell did they cull that wonderful bit of misinformation from, WHY is it being foisted on me, and don’t they have any competent proofreaders on hand to make sure that a generation of techs don’t walk off campus with associates degrees in hand, still thinking this tripe is true?
There’s a couple of threads around here on that topic. Short answer, the info you have is old. The rules change every couple of years as the number of connections squeeze the numbering systems.
Oh, and my TCP/IP books do mention CIDR, but geographic location, company name, etc, shouldn’t have anything to do with CIDR. It would be nice if it did, but this was something they had to develop to work alongside the old class IPs, right?
dylan, what are you thinking ? That’s like turning a firehose on a guy who’s asking for a glass of water. Besides, if too many people start reading RFCs, I’ll have to start checking my facts before shooting my mouth of and where would that leave me ?
Neutron star, you caught it: Your textbook is flat-out wrong. Inform whoever is responsible for book purchases.
And just to nit-pick: The class of network is not decided by the first octet, but by the first 1 to 4 bits, depending on class (A-E). As you’ve read, “class” is a somewhat outmoded concept and is now more or less used as a short-hand description of the size of given IP subnet, regardless of the first octet. It’s technically incorrect, but everybody’s doing it.
Example: The subnet 185.1.1.0 255.255.255.0 has 256 addresses, and would normally be called a “class C network”, although the classical, strict definition of a class C network is a network with 256 addresses and the first 3 bits of the first octet “110”.
Thank G-d for classless routing and variable subnet masks.
Every time a friend and I connect to the internet, and we live 600 km’s away from each other, and we dial different servers, we have an identical IP address, and it is THE SAME address every time we log on.
We keep getting kicked from mIRC chat rooms for cloning.
Do you have the same ISP ? - and do you get a registered or a private (192.168.x.x, 172.16-31.x.x or 10.x.x.x) address ?
I might perhaps imagine an ISP using these addresses for dial-in customers and tehn using adress translation with overloading (NAT/PAT) at a central location. That is, several users share one official IP address and the necessary bookkeeping is done by using UDP/TCP port numbers That could explain why you’d get kicked out from places that checks on the IP address only. But an ISP handing out the same IP address to two different customers would get some interesting internal routing problems.
Weird. Going from what you’ve told, it really shouldn’t work. Do you have any additional data ?
Thanks, Spiny Norman. The scary part is that my teacher never noticed this. They’ve been using this book for two years and each class only lasts a quarter, so that means eight classes have learned this erroneous info. He also teaches more advanced courses. I think I’m going to try my damnedest to find other teachers for those courses when I get to them.
Not to be anal about you being anal, but the network address and the broadcast address (.0 and .255, respectively, with the 24 bit subnet mask used in the example) are - as their name implies - addresses. Set aside for a special purpose, but addresses none the less. Together with the 254 possible host addresses, it adds up to - wait for it - 256 addresses in the subnet
As a middle ground - using (2^n-2), you will determine the USABLE number of subnet addresses. I don’t consider “This network” and broadcast to be usable addresses…
In any case, it is odd that you would both be showing the same address. If you enter IRC, before joining any channels, and do a “whois” on each other — do the IP addresses match? Or if you finger each other from a shell? Unless your ISP doesn’t know how to NAT, I can’t imagine what your problem would be.
Not to be anal about you being anal about Nannook being anal, but anymore the .0 address is pretty much universally usable as an address. It’s true that it was once bad and in some cases still is, but generally it’s usable. .255 is and probably always will be out because it’s the broadcast address.
Spiny is of course completely correct about neutron star’s criminal book and his assessment that the concept of classes is outdated.
Mad Hatter wrote
The odds against this are infitesimally small. There’s something else going on here.
Unless things have changed since the days when I did a lot of IRCing, a whois won’t give you an IP address. Type /dns* username* to obtain that information.
Are you sure you’re not mistaken, MadHatter? How would the IRC server know which user was which if you both had the same IP? Could it be that your IPs are in the same range, and that that entire range has been banned by the channel you frequent? Channels often do this to prevent banned users with dynamic IPs from disconnecting and reconnecting to their ISP to obtain a different IP, thereby skirting the ban of a single address.
