BobSutan
[H]F Junkie
- Joined
- Apr 5, 2000
- Messages
- 12,226
Is there any documentation that explains why 32 bit subnet masks work on Loopback interfaces on Cisco routers? How about any rules governing when they could (or should) be used?
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
32 bit subnet masks
- Thread starter BobSutan
- Start date
- Joined
- Jun 27, 2001
- Messages
- 15,832
32 bit mask would be 255.255.255.255 right?
this is a weird one... because ive actually seen a lan where the subnet is 255.255.255.255 and i have no clue how/why it was working...
in theory it shouldnet have worked at all... but magically all the machines on the lan were working with that subnet
this is a weird one... because ive actually seen a lan where the subnet is 255.255.255.255 and i have no clue how/why it was working...
in theory it shouldnet have worked at all... but magically all the machines on the lan were working with that subnet
foreignkid
Gawd
- Joined
- Jun 18, 2002
- Messages
- 521
Well, its pretty simple.... all the traffic goes through the default gateway. As long as the gateway didnt have that same subnet mask. This would be great if you had a bunch of computers that you wanted to monitor all the outgoing traffic on.
- Joined
- Jun 27, 2001
- Messages
- 15,832
the router would have to have the same subnet mask, if not then you couldent get to it anyway...
At work we use host addressing for network probes that we have deployed on critical customer connections. We do it this way because those probes need to act as their own network, do not require a gateway, etc. All management traffic is sent from these probes, across dedicated DLCI's to border routers which in turn distribute this information through BGP. So, having a subnet mask of 255.255.255.255 does not necessarily mean a device would not be reachable, although on a LAN it would be a much different scenario.
foreignkid
Gawd
- Joined
- Jun 18, 2002
- Messages
- 521
FLECOM, I think you're wrong.... you realize that the only thing that a subnet mask does is to help the device decide whether or not to send the packet to the default gateway. And it also lets it figure out the broadcast and network addresses of the subnet.
I'm thinking that you'd do this if you wanted all traffic to go through one central point (router, filter, computer with ethereal, etc..) before it got to its destination. Simple as that.
I'm thinking that you'd do this if you wanted all traffic to go through one central point (router, filter, computer with ethereal, etc..) before it got to its destination. Simple as that.
Ok, some of you probably already know this, but a subnetting is borrowing bits from the host portion of the subnet mask to create subnetworks. In a class c ip address, the last octet is the host. If you were to borrow all the bits, 1111 1111, ie 255, you would have 64 subnets, and 4 hosts per subnet. You would have 2 useable hosts per subnet in this situation then since the first and last are in use. The last is used for broadcast, and the first is the new network number. It is possible to have a subnet mask of 255.255.255.255.
edit: maybe doesnt answer the question exactly now that i reread it, also this is my second day in datacom we have been talking aobut subnetting
edit: maybe doesnt answer the question exactly now that i reread it, also this is my second day in datacom we have been talking aobut subnetting
foreignkid
Gawd
- Joined
- Jun 18, 2002
- Messages
- 521
So i guess the question is, when there's only one address in a subnet, is that address going to be usable, or is it going to be a broadcast address (last in subnet) or is it going to be a network address (first in subnet, usually .0 on bigger subnets)? Does this depend on how TCPIP is implemented in the operating system? Somebody needs to experiment.
I think that as long as the router, etc, whatever is going to send packets to the computer with that snet mask, doesnt have a snet mask of 255.255.255.255, itll be ok. THe computer sending it will just send it to the default gateway, and itll be dealt with appropriately there (at that point the snet mask wont matter).
Derrtor, you're describing a snet mask of 255.255.255.252 or in other words a /30 address. And, for all practical purposes, the names class a b and c are obselete, replaced by CIDR.
I think that as long as the router, etc, whatever is going to send packets to the computer with that snet mask, doesnt have a snet mask of 255.255.255.255, itll be ok. THe computer sending it will just send it to the default gateway, and itll be dealt with appropriately there (at that point the snet mask wont matter).
Derrtor, you're describing a snet mask of 255.255.255.252 or in other words a /30 address. And, for all practical purposes, the names class a b and c are obselete, replaced by CIDR.
BobSutan
[H]F Junkie
- Joined
- Apr 5, 2000
- Messages
- 12,226
I do believe this thread will help shed some light on the matter.
The core of it is that on cisco routers all 1's in the host field means its the the broadcast for the subnet. However, when you have all 1's in the subnet mask the router interprets it as meaning just look at the one single host.
Again, is there any documentation available that explains why this is? I've found nothing on cisco's site, google, or even in TCP/IP that could shed any light on the matter.
The core of it is that on cisco routers all 1's in the host field means its the the broadcast for the subnet. However, when you have all 1's in the subnet mask the router interprets it as meaning just look at the one single host.
Again, is there any documentation available that explains why this is? I've found nothing on cisco's site, google, or even in TCP/IP that could shed any light on the matter.