Can the default gateway equal the network address?

Question

Computer networks

I'm slowly getting ready to pass the MCSE for 2012. Looking through the dumps, I came across something strange.
Given network 10.10.0.0/24 Machine address 10.10.0.15 Gateway address, attention, 10.10.0.0 I
thought it was an error, I tried to set such settings in Windows 8 and to my surprise, the OS did not swear.
To be honest, the picture of the world began to slowly crumble. I was tormented all day, but I did not come up with an answer why such a configuration is allowed and why to do this. I know that in the case of /31 networks, in which there are only 2 addresses, it is allowed to use the network address as the ip address of the device. But here /24. I gave up asking the community for help.

Answer 1

In the classical approach, in theory, no. However, a number of equipment (not all) support the /31 mask, in which the first address is set by both the gateway, and the broadcast, and the network address. It works.

Answer 2

I'm not a network expert a bit, but for myself I made the following conclusion:
The network address and the broadcast address are just IP addresses, one of many.
And the reaction of network devices depends on the proximity of the implementation to the standard.
Nothing prevents, except for artificial restrictions, to put these addresses as machine addresses. Another thing is how other equipment on the network will react when sending packets to this address.
You can read some discussion here

Answer 3

And the gateway from the machine is not available through the tunnel, purely by chance? :-) There can be any addresses in the tunnel, and everything will work.

Answer 4

I will express my opinion based on my own experience. If you suddenly hurt someone's religious feelings ("the gospel of Olifer decided to contradict the heretic!"), I'm sorry.
The term "network address" itself is a legacy of the dashing times of classful addressing. From there, the terms "network" and "subnet" in the sense of IP addressing. I propose to forget about these terms, due to their weak applicability to the realities of 2015, and operate with the concept of a prefix , i.e. groups of addresses with the same leading bits.
In your case, I will assume that the OS distinguishes between the route to 10.10.0.0/32 (to the router) and to 10.10.0.0/24 (the route to all other hosts on the network). This does not mean that everything will work. Certainly, nuances with the implementation of ARP, etc. may come up.
There is also the question of how, when and why the "null" address was banned from being used as a host address. Without having the source code of old *nix in front of you, it’s hard to say, but the hypotheses are as follows:
1) earlier, the “zero” address was used not as a network address, but as a broadcast one, probably, the roots grow from there
2) the nuances of implementing routing and traffic forwarding. For example, an entry for a route to a classful network was stored without a mask/prefix (the meaning of classful addressing), it was probably not possible to distinguish between a route to a network with address ccc0 and a route to a host with address ccc0 . Then, probably, this approach ("whatever happens", or fear-uncertainty-doubt in developed democracies) spread to subnets, and then to prefixes with classless addressing.
Generally speaking, it seems useful to try to solve the following problem.
You have an L3 network device (router, because almost any server or computer is engaged in routing its own traffic), with several ethernet interfaces and with a certain number of routes. How to organize the routing table view? How, given the destination IPv4 address, to determine how and where to forward the packet? Try to make an approximate algorithm for yourself, a lot will become clear.