Re: Multi-Networking in Linux - Multiple Subnets
Marc Haber <[email protected]> Fri, 31 Jul 2026 15:04:36 +0200
| Newsgroups | comp.os.linux.misc |
|---|---|
| Organization | private site, see http://www.zugschlus.de/ for details |
| Message-ID | <[email protected]> |
The Natural Philosopher <[email protected]> wrote: >The class ABC terminology persisted as a neat way to refer to specific >netmasks Only in Idiots, the rests uses /8, /16 and /24. >And the private networkls of those sizes were defined to never be on a >public network or routable by BGP But noone with a sane grip on networking ever configures 192.168.0.0/16. You use a /24 network out of that range. And if you're really smart, you pick a random third octet¹ to avoid conflicts. >In the end its the usual mix of pedantry, technical issues and common >speech. > >https://datatracker.ietf.org/doc/html/rfc1918 > >RFC 1918 may be 30 years old, but it is still in effect. And it refers >to classful networks Of course it does, and it doees so correctly. 10.0.0.0/8 IS a class A network, 172.16.0.0/12 IS a group of 16 class B network and 192.168.0.0/16 is a group of 256 class C networks. an IP address belonging to a class is determined by the first bits. IP addresses starting with binary 00 and 01 belong to Class A networks (0.0.0.0-127.0.0.0), IP addresses starting with binary 10 belong to class B networks (128.0.0.0-191.255.0.0), and IP addresses starting with 110 belong to Class C networks (192.0.0.0-223.255.255.0). There used to be a class D (beginning with 111), but thankfully noone remembers that. 172.16.10.0/24 is not a class C network despite having a /24 netmask, it's a /24 subnet of 172.16.0.0/16 which is a class B network. |[8/5177]mh@swivel:~ $ ipcalc 172.16.10.0/24 |Address: 172.16.10.0 10101100.00010000.00001010. 00000000 |Netmask: 255.255.255.0 = 24 11111111.11111111.11111111. 00000000 |Wildcard: 0.0.0.255 00000000.00000000.00000000. 11111111 |=> |Network: 172.16.10.0/24 10101100.00010000.00001010. 00000000 |HostMin: 172.16.10.1 10101100.00010000.00001010. 00000001 |HostMax: 172.16.10.254 10101100.00010000.00001010. 11111110 |Broadcast: 172.16.10.255 10101100.00010000.00001010. 11111111 |Hosts/Net: 254 Class B, Private Internet | |[9/5178]mh@swivel:~ $ ipcalc 172.16.0.0/16 |Address: 172.16.0.0 10101100.00010000. 00000000.00000000 |Netmask: 255.255.0.0 = 16 11111111.11111111. 00000000.00000000 |Wildcard: 0.0.255.255 00000000.00000000. 11111111.11111111 |=> |Network: 172.16.0.0/16 10101100.00010000. 00000000.00000000 |HostMin: 172.16.0.1 10101100.00010000. 00000000.00000001 |HostMax: 172.16.255.254 10101100.00010000. 11111111.11111110 |Broadcast: 172.16.255.255 10101100.00010000. 11111111.11111111 |Hosts/Net: 65534 Class B, Private Interne If you insist on using obsolete terminology, please at least get it right. >In a private network you can do WTF you like and have network spaces at >any one bit boundary of a netmask . You can probably create even weirder >netmasks than 2^n -1 boundaries, though only Farley Flud would try one. Of course. I have a /23 network for historical reasons. When you operate with public IPv4 addresses, you'll get quickly acquainted with /27 through /30 networks. >Most people stick to the simple cases. And split into 'class C sized' >subnets even in larger corporations because it limits the extent of >broadcast floods. Even though switches already limit traffic between >nodes to that which is desired only. You even should split smaller. Many companies have been so careless with the RFC1918 address space that they've run out of those addresses. On my network, I have /24 as network size, and I have like 16 of them. I was too lazy to use /26 or /26 networks which would have been a more sensible size. >Carlos has a point. If Starlink restrict his router to a precise local >IP and netmask, if he wants more he has no real option than to create >some routable subnet . It is still stupid to use such a HUGE network. Greetings Marc -- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " | Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402