To preface: I'm not making arguments against IPv6; rather against the lack of sound reasoning being the driving force behind it. I like the next shiny obstacle as much as the next engineer, but have had too much PHB experience to allow it to distract me from making a valid case.
>> This question comes back up every few months, and every time I wonder: >> what is the justification case for IPv6? Aside from those home > We're running out of addresses, and we better start deploying two > years ago. Unless you want to start living with NAT at ISP level, > which would suck. This has been happening for years; some ISPs are selling it as 'enhanced security' connections, others are just doing it silently. For >90% of the population, ISP NAT is 'good enough' and often better than what they have. Although distasteful, I also believe the "pay for a public IP" scenario is awfully likely; they'll just roll it into the T's & C's of a "business-class" connection and treat it the same as a static allocation. >> hackers that are desperate for a full 128 bits of addressing to route >> the twelve devices on their network (never mind my public wifi network >> that eats an entire /17 with all its churn), where are the potential >> users? Who has put off rolling out pfSense or a similar platform > Everybody. Mobile device users for starters. I presume you mean mobile devices are potential users. Unfortunately, you have a theoretical disconnect - not only would (my number) less than 0.5% of mobile device users _need_ a publicly routable IP, the truth of the matter is that on most cellular connections I've worked with even though you're assigned a public IP (unless connecting via a Windows phone), you are allowed zero inbound connectivity and have to initiate everything from the mobile. How is that any different from NAT? I've been around the block a time or three in the mobile space, and although global addressing is attractive I just don't see that market as a driving factor. >> because it didn't implement IPv6? What about the fact that for the > You're talking about the past. There has been no address scarcity > in the past. I am most certainly speaking in the past tense, but allow me to alter it for your strawman: who won't roll out platform X tomorrow because it doesn't provide v6 services? Ever since IPv6 was ratified people have been moaning about address scarcity - why are 39 /8's still unallocated and many huge spaces are not even publicly routed?. You make the case earlier that we should have been deploying this two years ago, and now try to say I shouldn't talk about the past. Why the double standard? >> huge majority of users, the magical IPv6 land of ponies and sugar >> cakes will end at their border unless they tunnel it out to some > > Why can't I terminate a 6to4 tunnel in pfSense? So I can offer > my customers native IPv6 connectivity, which my hoster doesn't, yet? Same question - you want to provide it, but what justification is there? Are you losing or missing clients because you don't offer native v6? Why (if they are) are customers requesting it other than it's a shiny new foo? Surely you've done supporting cost and market analysis? If you could prove even one lost customer, that would be a viable case for directly funding adding a 6to4 tunnel to pfSense; two, and you'd likely be coming out ahead. >> 3rd-party provider? Yes, some ISPs are starting to offer v6 >> connectivity, but those are few and far between. > > I have a small business with a /24. In order for me to make money > I will soon have to order another /24. And then another. This is the normal course of business: you purchase a fixed amount of a consumable asset and when said asset is depleted you make the business decision to replenish your supply, go out of business, or pursue other venues. Where is the problem? If you have failed to keep up with the cost of that asset and plan for the expense of replenishing it, suddenly being gifted 72 quadrillion times more of the asset is only going to postpone your business' demise from poor planning. >> I'm not against IPv6, I just disagree with the periodic >> Slashdot-induced handwaving 'emergency'. We've been "on the cusp" of > > Slashdot-induced, huh. The query is posted on the same day a hand-waving article hits Slashdot's front page; the first response is you posting a link to said article. Make the connection? >> "an addressing crisis" for years, and the fact that someone has >> slapped a ruler on the current allocation trend and come up with a >> number of days under 1000 doesn't really cause me concern. Who can >> present a reasonable case for adoption before the current 2-3 year >> timeline? > > Do you realize how long hardware deployment takes? Right now > we're driving at a nearby brick wall with a floored pedal. Yes, yes I do. My first IT job was working on a wireless hardware team wherein we managed both the infrastructure and clients for ~4.5k international locations and >50 client devices per locale. We went through three complete hardware refreshes (two total technology changes) in the space of 3 years. Again, I am not unsympathetic to your cause and would gladly do the work if my employer paid me to do so, but unless there's a real business case, development will have to continue at hobby speed. RB --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
