by
ewhac » Fri Aug 28, 2020 4:10 pm
Thank you for finally directly addressing this question. I suspected it was related to support costs, but some of the conspiracy theories I was forming in my head were growing downright outlandish. I hope none of these additional costs come from static IP subscribers badly managing their machines.
dane wrote:First and foremost, supporting static IPs adds significant network complexity due to inflexible numbering configuration in the network. In other words, the need to allocate IPs statically into specific POPs and chassis is a significant challenge as the network changes and grows. On the Fusion service, this has proven very burdensome, making changes in the way we serve customers extra complex and error-prone.
Would it help if you were able to change static IP allocations from time to time? I very much want to keep static IP, but I don't need a
specific IP address. So if an email were to show up reading, "Hey, we're juggling our network around to support new subscribers, so your static IP block is changing from X to Y. This change will take place on $(DATE)," I'd be okay with that.
Finally, static IPs are no longer as critical as they were in the past. Cloud-based services and IoT remote management no longer requires a static IP in the way that older devices used to.
I think you're going to find various degrees of disagreement on this point. While this is true to some extent, there remains the issue of control. I'm sure you've heard the aphorism, "The Cloud is just a fancy term for, 'Someone else's computer.'" Control of one's own presence on the network has slowly been pulled away from the end-user and increasingly centralized, leaving us with dumb terminals designed solely to consume, and little opportunity to create and produce. Yeah, I could setup a VPS and host a mail and Web and Mastodon server there, but that machine is not mine, running software I don't control. Yeah, I could post a video to YouTube, but it can be pulled down at any time for any reason, including no reason.
That having been said: Have you considered static IP via VPN? ("What?") Basically, set up a VPN concentrator whose internet-facing IPs are subscribers' static IP allocations. The subscriber then sets up a VPN tunnel from their box to the concentrator. Poof! Their box looks like it's sitting on a static IP address (because to all outward appearances, it is), and you get to centralize static IP management on a handful of VPN machines in a manner most convenient for you and avoid all the annoying one-off POP routing. (This has the added benefit of, if someone starts attacking my box, I can tear down the VPN link, and still be able to talk over the net via the normal DHCP connection.)
Something to think about, maybe?
Again, thank you
very much for concretely addressing this question.