Right design (resources) for a nebula network

[TomFreudenberg]

Hi, first of all - thanks for that great piece of tool and to provide it as Open Source.

It works like a charm and I am impressed how got all fit together with my network concept.

The only part I am strugling with is about resource allocation for the nodes and lighthouses in case of CPU, RAM and traffic consumption.

For the network topology:

a. 4 lighthouses in different areas and states
b. 4 opnsense controlled gateways to control generel in and out traffic from and to public
c. x nodes running different services and applications
d. y nodes running as internal gateways
e. z nodes running as road warriors

Constraints:

f. the x nodes are mostly 90% running IPv6
g. the 4 lighthouses are running IPv4 and IPv6 and working as relays between
h. the 4 opnsense are running IPv4 and IPv6 to connect to the x nodes from public
i. the y nodes are running 50/50% IPv4 and IPv6 (there maybe a number of relaying traffic)
k. the z nodes are running mostly 90/10% IPv4 (there ist mostly relaying traffic)

The complete net is already up and running.

---

My questions are only about resource allocation and traffic speed:

1. Should I run the 4 lighthouses and 4 opnsenses on different VMS or should each opnsense also be a lighthouse?
2. In case of relaying traffic between IPv4 and IPv6 in case of performance should I increase the number of relay hosts?
3. In case of relaying traffic and prevent creating needle holes should I always try to upgrade the y gateways as it best also to IPv6? - or alternatively - get IPv4 on all of the x nodes?
4. How much CPU and network performace does it cost for the relays

Thanks in advance for any hint about the questions.

Best regards
Tom

[TomFreudenberg]

Hey Wade and Nate ( @wadey @nbrownus )

hopefully you have a short reply to my question?

Thanks in advance
Tom