- Is it possible to have multiple logical segments (i.e. subnets) on the same physical segment (i.e. l2 broadcast domain)? Of course it is; is good or common network engineering practice? Probably not;
- Can this model be replicated in a cloud environment? Of course it can; would it make sense? Probably yes.
However, can this be done differently? I mean, if you created two logical networks A and B, with a single subnet S1 and S2 respectively wouldn't you achieve the same result? Is just the cost of creating the extra logical network that's stopping you? I think it's legitimate to lift this constraint, but then all sorts of issues may arise.
For example, if you now have Node A and Node B on the same logical network but sitting on different subnets they wouldn't be able to talk to each other if not via a router; this sounds like a complication that can be avoided in, at least in the logical space.
I would ask myself the following questions:
- Is it possible to have multiple logical segments (i.e. subnets) on the same physical segment (i.e. l2 broadcast domain)? Of course it is; is good or common network engineering practice? Probably not;
- Can this model be replicated in a cloud environment? Of course it can; would it make sense? Probably yes.
However, can this be done differently? I mean, if you created two logical networks A and B, with a single subnet S1 and S2 respectively wouldn't you achieve the same result? Is just the cost of creating the extra logical network that's stopping you? I think it's legitimate to lift this constraint, but then all sorts of issues may arise.
For example, if you now have Node A and Node B on the same logical network but sitting on different subnets they wouldn't be able to talk to each other if not via a router; this sounds like a complication that can be avoided in, at least in the logical space.