Subj : RIPE to run out of IPv4 this week? To : Andre Grueneberg From : Michiel van der Vlist Date : Thu Aug 09 2012 23:09:31 Hello Andre, On Sunday August 05 2012 23:37, you wrote to me: MvdV>> Then came TCP/IP. Even Novell switched to TCP/IP. And where is MvdV>> IPX now? Now not even two decades later, IPX is gone for all MvdV>> intents and purposes. IPX/SPX and TCP/IP can run "dual stack" on MvdV>> the same physical network. But once you have TCP/IP running on all MvdV>> your nodes, there is no added value in keeping IPX alive. AG> Good example ... Or maybe not. ;-) AG> IPX/SPX is still in all those OSs. It's in Windows, To my surprise, I see that it is still present in XP. I had to look for it, but indeed it is still there. So maybe it was not susch a good example to make my point. What about others? Banyan VINES is stll in W98, but it is gone from XP. What about X25? Appletalk? IBM token ring? ARCnet? AG> it's in Linux ... you can still enable and use it. I suppose so. But only on the LAN. So what good would it do? AG> You and me, we might not know anyone out there actually using it ... AG> but there definitely are people with some really ancient legacy. Outside the museum? AG> It's like at my work ... for some reasons we still have some AG> Mainframes running MVS around. You do not usually realize these AG> dinosaurs are around, but they definitely are. What are they used for? Are they actually doing something useful or did nobody bother to switch them off? MvdV>> Something similar could happen with IPv6 and IPv4. AG> That's the point ... it'll be there for a very long time to go. Well, we will see. You may be right in that it may be a long time before the last IPv4 capable system will be switched off. I expect to live to see it gone for all intents and purposes and to see it only in a museum. AG>>> For a long time, (minimal) IPv4 support will be needed for AG>>> several transition methods. MvdV>> why? Once "everyone" has IPv6, what is the added value of MvdV>> dragging IPv4 along? AG> When do you believe this "everyone" will be? I write it in quotes to indicate that it should not be taken too literally as in "no exceptions". More like "everyone" has a cell phone. (My 03 year old mother doesn't) AG> When everyone has disposed off his 10 years old IPv4-only printer? It will take as long as it took the printers with Centronic interfaces to fade out. Yeah, there maybe someone out there who still has an old ARS33 in use. I dumped mine 10 years ago as a collectors item. AG> When this good "old" embedded thermometer with IPv4 will stop working? My T65 (http://nl.wikipedia.org/wiki/T65) is still in working order. But I can't use it to make outgoing calls any more. Pulse dialling is no longer supported by what ever I can hook it ip to. AG> There's still people out there developing IPv4 only. in COBOL? ;-) AG> These will be mostly internal applications/services, but there's AG> little value in replacing it ... software is cheap. Hmmm... AG>>> Just see Tore Anderson's presentation at RIPE64 ... I believe AG>>> he's got a good point. AG>>> https://ripe64.ripe.net/presentations/67-20120417-RIPE64-The_Cas AG>>> e_for_ IPv6_Only _Data_Centres.pdf MvdV>> I'll have a look at it one of these days. I still fail to see MvdV>> what the added value of maintaining IPv4 is when "everyone" has MvdV>> native IPv6. I read it. Yes, I agree, Tore has a point. What he proposes may be a good solution. For datacentres. AG> My point being: This day is far-far away. And for most people Internet AG> means business. Losing 1% of customers can make a huge difference ... AG> at least that was the excuse for slow IPv6 adoption. I think that was indeed just an excuse. Losing 1% of the customers may or may not be a big deal. It all depends what type of customers. The tail end of the curve is seldom profitable. Cheers, Michiel --- GoldED+/W32-MINGW 1.1.5-b20110320 * Origin: 2001:470:1f15:1117::1 (2:280/5555) .