Could someone explain why GCE defaults to using /32 subnet masks when configuring interfaces on most GCE instances? I have noticed all of our Linux servers are using the /32 subnet masks. I have not found a clear answer in the docs as to why. I've also noticed that our Windows servers do not use /32 subnet masks and have the "MULTI_IP_SUBNET" Guest OS feature enabled which tells DHCP to use the subnet mask of the network.
Hi @smcevoy ,
When you create a VM instance, it gets a main internal IP address. This IP address uses a "/32 subnet mask," meaning it's treated like a single, isolated host.
Additionally, you can give your VM more IP addresses using something called alias IP ranges. Each of these extra IP addresses also gets a "/32 subnet mask." This setup is handy for various situations, like when you have multiple services or containers on the same VM and want them to have separate IP addresses.
Using "/32 subnet masks" ensures that each IP address is treated on its own, providing flexibility and isolation for different networking needs.
You may also find this related community discussion helpful.
Let me know if this addressed your questions. Thanks!
Thanks Marvin. That is helpful. I am still curious why Windows GCE images use the "MULTI_IP_SUBNET" guest OS setting by default? Does Windows not support alias IP's?