We know that subnetting before CIDR involved sub-dividing classful networks So subnets of class A, B and C. But with CIDR we can have arbitrary prefixes. For example, 192.0.0.0/7. Could we have done that before CIDR, or were the subnets restricted to their respective "classes of networks". If I had put an arbitrary subnet mask on say, old version of Windows, would it stop me?
2
-
1Before CIDR, there were no subnets. Classful means classful - the address dictates the netmask, you don't even get to enter the netmask. Was it possible before CIDR? Maybe, but your network stack was technically broken.– RickyCommented Jun 21 at 4:31
-
That’s not true. Subnets existed as far back as ‘85. You were allocated a classful network and then internally subnetted it.– user242114Commented Jun 21 at 12:58
Add a comment
|
2 Answers
Before the introduction of CIDR in 1993, VLSM provided a somewhat similar method of subnetting.
While supernetting classful networks might have been possible (depending on implementation), it wouldn't have been practical since all allocations were issued by class. Freely usable private addressing didn't exist yet (where you could use 256 class-C or 16 class-B networks).
You'd have to find a very old version of Windows to predate CIDR capabilities (perhaps WfW 3.11+Wolverine or NT 3.1) but host implementations and configurations aren't on topic here.
-
My question isn’t about private vs public of even connectivity to the internet. But if TCP/IP stack of routers allowed it. For example, putting a bigger mask on a class C address. When the ISP gave you, say, four class C addresses, could you have supernetted it on your routers and used 255.255.252.0. Or would the software stop you and say that class C addresses don’t have masks bigger than 255.255.255.0. Thanks Commented Jun 21 at 12:47
-
1Possibly, depending on implementation. I was trying to refer to actually being able to use multiple x-class networks to supernet, which you commonly wouldn't with public addresses.– Zac67 ♦Commented Jun 21 at 12:58
The concept of subnetmask itself isn't part of IP header. The subnet mask, however, is currently being used by networking vendors to confine the limits of your local subnets. I would say, before the concept of CIDR there was no concept of subnet mask and hence all devices in network would reach each other (Note this is a vendor implementation. It is not defined in IP/TCP stack so each vendor would choose what networks can be reached and how. but having no concept of subnet mask would make all IPs reachable to reach other same way).
I can read the word "mask" in first CIDR RFC: https://datatracker.ietf.org/doc/html/rfc1519
Hope this helps.
-
ok, but there are RFCs before 1519 that mention subnets. for example, RFC 917 Commented Jun 26 at 17:09
-
Thanks for sharing RFC. I wasn't aware about this RFC. It also mentions variable-width subnets in this RFC before CIDIR RFC I mentioned above. "If a self-encoding scheme is not used, it is clear that a variable-width subnet field is appropriate. Since there must in any case be some per-network "flag"<....omitted text....>. The advantage of using a variable-width subnet field is that it allows each organization to choose the best way to allocate relatively scarce bits of local address to subnet and host numbers." Commented Jun 27 at 13:31