Oddities in the software, routing?

Question

Network administration

Good afternoon!
We have the following situation: a server on Windows 2008R2, clients on windows 7.
Specialized software is installed on the server and a hardware protection flash drive (Guardant key) is plugged in, a protection software layer interacts with this flash drive, being between the server and the client and working via TCP / IP according to a specific TCP port, respectively, the client on Windows 7 can start its client software only if there is a security key on the server.
The server and clients are on the same subnet, connected through the same unmanaged switch.
Network settings on the server (one network card):
192.168.1.15, mask /24, gateway 192.168.1.1
On clients:
192.168.1.3X, mask /24, gateway 192.168.1.1
The gateway is a regular zyxel keenetic router.
All network settings on the server and clients are registered statically, the router does not distribute them via DHCP.
Now we take and simply turn off the gateway, i.e. 192.168.1.1 becomes unavailable, and Internet access is lost accordingly. Clients from the server continue to ping normally, like the server from the client side, the local network continues to work (which is logical).
BUT the software stops running on the client side, swearing at the lack of a security key on the server...
We remove the default gateway (192.168.1.1) in the server settings, leave only the IP address and subnet mask, after that the software starts working fine from the client side. .
I don’t understand in the end, are these some kind of Windows jokes or is it hardware key protection?

Answer 1

Apparently this is a crooked key driver on the client/server. Contact the software or hardware key manufacturer for support.

Answer 2

Most likely, it crawls somewhere on the Internet to check something. When the gateway is not specified in the system, it breaks off quickly, since the system knows that there are no routes to remote networks. When the gateway is simply disabled, it waits for a certain timeout of the call, and after its expiration it breaks off. Probably, the timeout in the software is less than the TCP / IP timeout, and the software is simply not able to process this situation correctly.
Run a traffic sniffer and see if the client breaks somewhere on the Internet when you run this secure software.