0

RFC 1918 “Address Allocation for Private Internets” (https://tools.ietf.org/html/rfc1918) specifies 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16 as private addresses suitable for unrestricted private internal use. Many home networks use the 192.168.1.0/24 address space.

Since 10.0.0.0/8 is a larger address space than 192.168.0.0/16, it would make sense to use 10.0.0.0/8 since there can be more IPs per network.

Why isn't 10.0.0.0/8 commonly used instead?

3
  • 8
    It is commonly used. It’s s mistake to think it isn’t.
    – Ron Trunk
    Jul 4 '19 at 17:25
  • 2
    Various 192.168.0.0/16 addresses are allocated by default on many domestic routers, which is why many domestic networks use them. 10.0.0.0/8 and 172.16.0.0/12 are in fact extremely common and used all over the place. It's just whatever the network designer chooses, for good reasons or poor reasons.
    – jonathanjo
    Jul 4 '19 at 18:06
  • You got me thinking, there might be a slight risk of conflict if your ISP does carrier-grade NAT (CGN) in the 10.* range, although the 100.64.0.0/10 range has been made available for that purpose.
    – Anonymous
    Jan 8 '20 at 21:31
8

It is merely a necessity constraint and personal preference.

If a Network is being built that would only have 10, or 20, or 50, or even 100 hosts, there is no reason not to use a /24 from 192.168.0.0/16. This is why home networks typically use the 192.168.0.0 range.

If a Network is being built that would (in all growth projections) only have up to 1000 or 2000 users, then picking a few /20's in the 172.16.0.0/12 range is perfectly reasonable.

If a Network is being built that would (in all growth projections) possibly exceed that, then the 10.0.0.0 IP space is ideal.

That said, there are no rules or regulations. Anyone is free to use the 10.0.0.0 range or the 192.168.0.0 range for whatever network(s) they want.

As an analogy... you can kill a fly with a flyswatter or a sledgehammer. But one of those is much simpler to handle, and just as effective.

2
  • 4
    You can even use all of the RFC1918 ranges together at the same time. ;-)
    – Zac67
    Jul 4 '19 at 18:25
  • 1
    I would argue that number of users is less important than number of subnets. I manage networks with ~100 users, but over three locations, and 50-100 seperate subnets. Going with 10.site.subnet.machine, e.g. 10.2.100.0/24 makes sense. The most important thing is to think trough it before deploying. Readdressing is a PITA... :)
    – vidarlo
    Jul 6 '19 at 22:05
2

As Eddie has very well explained, there's not much difference in which range you use.

One tiny difference there is though: since 192.168.0.0/16 is allocated from the former C-class range 192.0.0.0/3 many systems default the network mask to 255.255.255.0 or /24. Those same systems default a 172.16.0.0/12 subnet to B-class, /16, and and 10.0.0.0/8 subnet to /8. So, the use of a /24 may come more "natural" with 192.168.x.y.

In our company we use 172.16.0.0/12 for historical reasons, with /24 subnets. When setting up something new I very often have to adjust the default /16 mask to /24. As it seems, classful networking isn't quite dead yet - but it should be.

1
  • 3
    I would say in my experience, businesses/corporations tend to make more use of 10.0.0.0/8 and 172.16.0.0/12 more often than 192.168.0.0/16. Making consumer devices that use 192.168.0.0/16 for home devices makes sense as it helps to avoid any sort of IP space "overlap" when using features such as VPN.
    – YLearn
    Jul 4 '19 at 21:03
1

Its well beyond the need for most home networks but a site-to-site VPN will generally have issues if the IP network blocks on each end are not distinct and unique.

IPSEC would be fairly common VPN tech to set up between two site firewalls/routers and has broad intervendor compatibility. But the routing devices need to know if 10.0.0.x is local or remote.

The dirty fix is for each site to nominate a subnet of their block where hosts are, so siteA has 10/8 but their servers are in 10.0.0/25 and siteB has 10/8 with the hosts of interest at 10.0.0.192/26

The right fix is to use as small a subnet for each network as you can, while still allowing for future growth.

1

To add here, because I've not read lots on this aspect yet: there's absolutely no obligation to use the 10 range with a /8 netmask!
You're free to have your own 10.0.0.0/24 network, which has a pretty convenient side effect: IP allows to omit intermediate zeros in IP addresses, so instead of having to type ping 192.168.1.1, why not reduce that to ping 10.1 (and yes, the Windows ping utility is capable of this too)?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.