old legacy equipment which has it's own routing, bridges, spanning trees and management networks.
Bridges pass Ethernet frames. They don't care about IPv4 vs IPv6. (Or IPX, or DECnet, or…) Some bridging implementations allow assigning an IP for management purposes, including RouterOS's, but having an IPv4 address on an interface doesn't stop it from passing IPv6. Furthermore, RouterOS allows you to have multiple IPs on an interface, including that for a bridge.
STP operates at layer 2, below the IP level. Adding IPv6 to your network will not affect it.
Management networks can remain IPv4-only, as I pointed out above. All I'm suggesting is that you allow IPv6
as well, not that you re-IP the world.
The only thing on your list that touches IP is routing, and I covered that above: you go dual-stack, keeping your existing IPv4 routing rules (plus firewalling, etc.) and simply
add IPv6 rules.
If the problem was as dire as you suggest, how did the world get to
the current level of adoption in the first place?
most CPE have a PPPOE client in the firmware
If by "CPE" you refer to that which your customers have, presumably provided by you, then why are you bringing them into it? The thing that started this thread is that one of your WISP's internal service VMs was trying to use IPv6. Allowing that doesn't change what your customers do except insofar as fixes made to allow it probably allow your customers to switch over to IPv6 as and when they're able.
If you mean some bit of CPE provided by your own upstream ISP to you, we can't be talking about a huge capital expenditure to replace it if it truly is as bad as you suggest.
In both cases, even if we accept your claim that the CPE has a hard-coded PPPoE client that can't be upgraded to support IPv6, that
still doesn't affect whether you can support IPv6. If you actually dig into the dual-stack links, you will find that there are plentiful options for getting around that, which is one of the answers behind my question about Google's IPv6 adoption data.
Alvarion
A WiFi company that went out of business in 2013, when 802.11n was still the hotness. It's probably time to upgrade anyway.
Also, WiFi — being a transport medium — is layer 2, below both IPv6 and IPv4. The radios doubtless have management IPs, but that can remain IPv4.
The only way this claim about them being old and out of support matters is if they have routing rules of their own and refuse to pass IPv6. That sounds like a consumer-grade all-in-one design, not a WISP backbone, where the paths are more or less fixed, to get traffic back and forth to the local center of operations.
If your Alvarion WiFi radios are merely data transports, then we're back to the "it doesn't matter" case.
Ceragon
…is still operating
according to Wikipedia. Its web site is still working, and they claim to be providing wireless networking equipment still.
Yes, they got sold. So what? As long as the equipment is still getting updates, what does it matter?
Also, being a wireless networking company, I'd bet we're back into "dumb transport" territory anyway, where IPv4 vs IPv6 doesn't matter.
Morola Canopy
The
Motorola Canopy system does at least appear to be out of production, but I found a manual, and it is as I claim above. Quoting it: "Switched Layer Transport with support for all common Ethernet protocols including IPV6, NetBIOS, DHCP, IPX, etc."
IT DOESN'T CARE.
Probably someone of your ilk might be able to research and craft ipv6 to pass thru them…
What irritates me is that as a WISP operator,
you are presumed to be of my "ilk." If you truly are incapable of doing this, I'm tempted to ask what ISP you run so I can be sure to
not ever use it. Not only do I use IPv6 services, today, I have to wonder what else you're dragging your feet on.
I mean, even frickin'
Comcast has got their ducks in a row on IPv6 these days! You're behind
them!
commercially there is no point
Customer satisfaction and retention are pointless?
unpredictable deployment costs and risks
First, there are experts who
can predict this. You may be able to hire one, if only as a consultant for a time.
Second, if you must do this yourself, with existing staff, that's what testing labs are for.
Third, being nearly three decades into this project, a lot of this is written down, recorded in video form, and available for in-person training. It's not like you're blazing new ground here.
monumental research and lab testing
And now we get down to the nub: motivated reasoning to avoid doing work.
I just installed a IPv6 firewall that drops everything and problem gone.
And now you have
a new problem.