Networks

[trehn]

# networks.py
networks = {
    'network1': {
        'ip4': '10.1.2.0/24',
        'ip6': '...',
        'metadata': {  # unlike node metadata, there is no merging of any sort going on here
            'gateway': '10.1.2.1',
        },
        'reserved_ip4': {  # or should we use dummy nodes?
            '10.1.2.1': 'floating router IP',
        },
    },
}

# nodes.py
nodes = {
    'switch1': {
        'needed_by_networks': {
            'network1',
    },
    'router1': {
        'networks': {
            'network1': {
                'ip4': {'10.1.2.2'},
            },
        },
    },
    'server1': {
            'network1': {
                'ip4': {'10.1.2.3'},
            },
        },
    },
}

With this, we could compute (and plot!) the shortest path between two nodes and provide a list of all nodes required for this path. Useful for host dependencies in monitoring. Also, free network diagrams.

This assumes a very simple routing setup and will not take actual routes into account. Good enough?

We could also auto-assign IPs from attached networks based on node.magic_number - but that would reshuffle all IPs in a network every time a node is added.

# bw net network1
10.1.2.1 floating router IP
10.1.2.2 router1
10.1.2.3 server2

# bw net --unused network1
10.1.2.4
10.1.2.5
10.1.2.6
...

[trehn]

A lot of the basics here can also be accomplished by using groups for networks. So we'd really be doing this for the plotting and host dependency features.