ctlow wrote:
Thanks, MrHoffman,
Your assumptions are correct. I looked at the MacOS firewall, …
The built-in macOS firewall can be left at its default.
My ISP is a major cellphone-network company with a poor history of customer service, which doesn't seem to have affected its profitability, so I can try that, but I don't think I will get very far.
Major companies can serve different customers including themselves, and sometimes less so their subscribers.
They recently provided me with this brand-spanking-new cellular broadband modem, not by Ubiquiti, for a price I couldn't turn down, SSH functionality or not, so my only options I think are to return it and go back to my (expensive) cable-modem service, or live with it.
I usually use the ISP device in its bridged mode or replace it entirely where permissible and feasible, and then install gear appropriate to the tasks. ISP gear focuses on basic functions and low cost of support. Pretty much everything g else is a distant third, at best.
Interestingly, to me, the modem configuration website does list "port forwarding", but it is poorly described in the documentation, and they list in their help files that it is not possible with this model. The port-forwarding page looks like this:
Here is a generic description ofhttps://en.wikipedia.org/wiki/List_of_TCP_and_UDP_port_numbers.
WAN port is the port facing the internet, and LAN port is the port in the internal net the traffic will get forwarded to, and the internal client is which internal host is involved. Between those two sides of the boxis usually NAT, network address translation. Because of some design fallout, NAT usually blocks all incoming IPv4 traffic. What happens with IPv6 varies. It may pass through, or it may be blocked.
Here is what the Ubiquiti port-forwarding equivalent looks like:

This is very similar ro what you are looking at, though the grouping allows the rule to operate with groups or whatever local ports or whatever local hosts together, rather than dealing with one port or one host at a time.
I don't know what most of that means. I know what a WAN and a LAN are, but not what to do with them here. Probably irrelevant to my question.
You will need a port-forwarding WAN-side rule for TCP port 22 (or another port, if you want to use a non-default port) to the internal LAN-side TCP port 22 on whichever computer is offering the ssh server on the internal network.
Here are the ports used by Apple software: TCP and UDP ports used by Apple software products - Apple Support
Here are the reserved ports, and you will usually want to recognise or (depending on what you are doing) avoid using these, if you are (for instance) picking your own WAN-side port to forward to some LAN-side host and TCP 22 LAN-side: https://en.wikipedia.org/wiki/List_of_TCP_and_UDP_port_numbers.
TCP port 8022 is not assigned, so you can conceivably “borrow” that for your external port, and set your ssh command to use (for instance) ssh -p 8022 which will then get port-forwarded to TCP 22 on whichever internal computer, once the port-forwarding rule is established. Or you can forward TCP 22 to TCP 22, but every gremlin on the internet will poke at that port. And poke lots. I usually use a VPN server and not ssh here, not the least of which is because that can forward some or all of the client app traffic using built-in tools. And I usually have firewalls with embedded VPN server capabilities.
I have more to learn about traceroute, but this command (names changed):
traceroute server.account@ddns_account.dynamic_dns_server.com
It’s usually traceroute to just the host name, without an account.
When you are making up a domain name, the domains example.org, example,com, and example.net are available, and are reserved for this and related uses.
ran for many minutes.
The first line reads (names changed):
www.company-wifi.com (192.168.x.y) 3.359 ms 2.765 ms 2.746 ms
It got up to 64 lines of:
* * *
It’s not gettig anywhere, though I;d usually use traceroute yourfirewall.example.com as the parameter, using whatever host name is temporarily yours thanks to DDNS.
and then went back to the command-line.
I don't know of any firewall-embedded VPN server. I didn't see anything like that when I looked at the MacOS one in System Settings.
To determine what features this ISP-provided box supports, you’d have to check the documentation for your ISP-provided firewall / gateway / router / NAT box.
A VPN server in the firewall is usually the best spot, as it deals with NAT, one of the scourges of IPv4.
Once the VPN is connected, your client would be effectively be another host on the same target internal network.
If the firewall lacks VPN server capabilities, it’s possible to run a VPN server on an internal host and set up port forwarding to access that, but that’s more effort and can meet some issues. Particularly around NAT, a scourge of IPv4.