New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Common but wrong metapher. One could see back then that some day more memory could- and probably would - be needed. It's just that bill "I'll jab you" gates pleased to not think that far.
I fully agree. But: there is a natural limit. There can't live an unlimited amount of people on this planet and those that can and do or will live here will not need more than say a couple of thousand IP addresses per person. Short, one can see that 64-bit IP addresses actually would be "hugely overspec'd".
Plus, btw that whole "people might need 1000/1 mio/1 bln IP addresses per person" is nonsensical for (at least) two reasons, (a) if you want to play the 'unlimited' game IPv6 addresses might be too small an address space too, and (b) we could do NAT even with yesteryear's hardware without significant pain, and we will certainly be able to do it tomorrow without significant pain given how hardware performance increases.
Actually, it is, if the rewrite would achieve the holy grail of working either way. I mean, your device can support only IPv4, the system will fall back to ipv4, if it supports IPv6 or whatever new system, it will prefer that.
It is not the rewrite which means it is not backwards compatible, if the rewrite supports the new system AS WELL AS the old one when it is not compatible with the new one, then it is backwards compatible. Say, automatic tunnel IPv4 over IPv6, every provider will be native IPv6 and if your device will not be, the router will detect that and create a tunnel to another IPv4 at the provider's end for your antique.
Say we would have a system called IPv4 2.0 where we double the bits of the mask, if your device works only with the old system, it will be detected and presented with a compatible interface in a reserved space like 000.000.000.000.123.123.123.123.
That would have solved the problem with minimal re-writing.
I wasn't doing the engineering, just making an observation.
Or 2^128 addresses, right? I mean, why aren't we using 2^4096? Such short-sightedness...
Sure, right now I can't think of a single reason why you'd need more than a thousand addresses each but isn't the point of future proofing to try and cover off use cases you can't think of?
I mean nobody is bitching about how there are >200 trillion MAC addresses; people are only butt hurt over IPv6 because it's a change from IPv4 and change scares them. Are you seriously arguing that you wouldn't be whining if only IPv6 addresses were a bit shorter? Pfft.
As for NAT... I don't know many network guys who view NAT as anything but a horrific bodge that should die a painful fiery death.
If you ask most eyeball ISPs that have implemented v6, most(60-80%) of their traffic is IPv6. This is because most content networks that makes up most of the traffic have implemented it as well. So it already is "relevant". CG-NAT is an ugly solution and makes the experience of using your service worse, so it is in your interest to turn on v6 as well.
No. Actually future proofing is about covering/being prepared for cases that are beyond current practical needs but within the range of the realistic.
That is what b. gates ignored back then and what the designers of IP4 in some way ignore.
Probably many, yes. But "change" isn't my point, at least not sensible and actually needed change.
"whining"? Sorry that's not a level for discussions with me. Try such lines with your girl friend/wife.
And I know many network guys, and I'm talking about hardcore professionals, who know and value that NAT saved us and did allow for many years of the internets survival.
Anyway, let us stay away from that level and limit ourselves to technical and relevant arguments.
"ugly"?
"[user/customer] experience" That's where I usually stop reading. But then, I'm not a "UX engineer" ...
A cynic would suggest you're using that as an excuse not to answer the question, but...
"oh wise and wonderful jsg, if you don't object to change is your only objection that IPv6 addresses have too many bits? Would you quietly accept them if they were 80bits long? 64bits? 42bits?"
I mean seriously, what's the optimal bit size, and why is it so wrong to err on the side of "ahh screw it, double the bits and we definitely won't run out any time soon"?
As for NAT; yes, it got around a lot of problems, but that doesn't make it a good solution. And at least with all those IPv6 addresses we might avoid CG-NAT-v6
"[user/customer] experience" That's where I usually stop reading. But then, I'm not a "UX engineer" ...
I think you misunderstood me. IPv4=worse network performance(throughput, jitter etc.) for your customers, not to mention broken P2P and E2E connectivity for your users. PS: https://tools.ietf.org/id/draft-donley-nat444-impacts-02.html#rfc.section.3.2
Not sure why you think that's part of UX?!
Kindly cut the BS
I still wouldn't like IPv6 but I'd be much more ready to accept it if it had a 64-bit address space.
as large an address space as possible within reason
Because it doesn't come for free. It costs memory, computation, etc. - and often exponentially rather than linear - plus the longer the addresses the harder to remember and handle for humans.
Plus, we "definitely won't run out any time soon" with 64-bit addresses too.
Yes, I probably misunderstood you.
But: one can be picky and find some disadvantages in anything and everything. With IPv6 one doesn't even need to search; It blinks neon-violett right into our faces.
That said, I do agree that IP4 wont carry us far into the future. Yes, we do need a successor - preferably one not designed in a mental asylum but by real engineers who only change what's really needed and sensibly so.
And yes, I'm all for being generous and going for a huge reserve. That's why I think that we shouldn't go for 48 bits (which would serve us well for decades) but rather for 64 bits, which just so happens to also be the max size of currently widely available technology and processors.
I think the fact we keep having this argument suggests it's not blinking as brightly as you think.
Ok, so it uses a few bits more storage (but seriously, against the ever growing tide of data an extra 64 bits on IP addresses is ... of questionable concern) and it's trickier to memorise an individual address (again, seriously nobody normal memorises addresses, that's what DNS is for) but... blinking neon-violet disadvantages?
If I buy a VPS I get /64 easily (hopefully provider friendly), if I do IP allocation from LIR will get /48 easily, if I still keep IPv4 compact allocation policy then I should go crazy, so I think it's just a new thing, let's accept it and don't invoke the IPv4 issue.
An extra 64-bit of address translates to pumping up many pieces (e.g. tables) to the square! Plus it puts us into a situation very similar to doing 64-bit computations on a 32-bit processor. Can be done but is considerably slower and wastes lots of energy - and again, what for? Even with 48-bits we'd be way beyond the population this planet could sustain.
As for "blinking", oh well we disagree. No problem, I can accept that our views are different.
I'm pretty sure that if computers 40 years ago could handle 32 bit IPv4 addresses, modern hardware can cope with 128.
As for wasting energy... I'd love to see the calculations that the extra 64 bits over what you'd find "acceptable" because of all the objections I've heard, the energy costs are one of the most speculative (to be exceedingly generous)
Gr8 b8 m8 i r8 4/8
No, I need not prove anything. YOU need to prove that 128-bit addresses are really needed and that 64-bit addresses aren't sufficient.
Good luck, I've yet to see a sensible argument for that that holds and shows that as many IP addresses as there are molecules in our whole solar system are somehow not sufficient.
Perhaps I'm ignorant but even after reading this thread, I don't see the issue with IPv6. Would smaller have worked? Sure. Is it a problem? Not really.
The problem is that the two are separate entities, you cant integrate ipv4 into ipv6, a different approach with minimal stack re-writing would have worked better and if we needed 256 times more IPs in the future that could have also been integrated the same way.
Opting for a totally different approach with real costs today for the needs that COULD appear100 years down the lane and which could have been solved again with minimal cost is the definition of overshoot.
So, you can just assert that 128 bit addresses generate an intolerable increase in energy usage - with no actual facts to back it up - but I'm the one that has to somehow prove that excessive future-proofing without any real cost isn't a bad thing?
I really don't care enough to bother. You enjoy your NAT.
Yes, the one who submits that an address space larger than the amount of molecules in our solar system allegedly isn't sufficient has the burden of proof, especially when considering that 128-bits is beyond the word size of currently available processors.
Well, you obviously cared enough to not accept my let's agree to disagree offer.
Well you're so comically obsessed with your "molecules in the solar system" figure that you're incapable of making any rational argument about why that matters. The fact that IP addresses are double the size of (current) processor words is almost entirely meaningless.
It's kinda hard to ignore such willful misinformation, but I'll try.
Well, I try to match your "4 billion times 4 billion IP addresses isn't sufficient, we need yet 4 billion times 4 billion more!!!"
... says the guy who, after being asked twice, did not offer a rational explanation for his "we absolutely neeeeed 128 bit addresses!!!" demand.
Because ahnlak says so? Sure ...
Wow, now you feel to even know my intentions?
Try again when you actually have arguments beyond "No! Because ahnlak says so!" and "you are evil!".
https://dmca.fileditch.com/ipv6.mp4
Jokes aside, great thread as usual. I see little point in IPv6, it's complicated and you have to run two different protocols, IPv4 and IPv6 with seperate firewalls and so on and on..
There is no real shortage of IPv4's. We still got Class E with 300M IPs literally "reserved for future use", and as soon as the price of IPv4 goes up then there's a lot of new seller
It's a heck of a lot easier to just have IPv4. Everything can use IPv4, but not all can use IPv6
Being lazy is not an excuse.
As I already told you Class E is something that many vendors/software mark as invalid/bogon so they'd need updates. And in any case it would only ease the shortage for a while but it is still not enough.
You are whining about some really simple stuff, man. Do you have 7 digit phone numbers or 10 digit? Did you rage on the phone company for going to 10 digit phone dialing? Fuck, is your password 1234567?
That's simply because you're not a network engineer and can only think of your super simple home situation, not large campus's, camp grounds, stadiums, global corporate offices, etc. You routinely say shit you don't understand that actual problems people run into.
No, it's like "go outside and play in wide open outdoors instead of playing in the basement all day". Nobody says you need to play in a different state, you can play at the local park or fly to Tokyo. The point isn't to limit the play area just because you can chill in the basement all day.
Again, it's accepted. You can't make a network device without IPv6 support these days. Or 5 years ago.
If that's all you've got I've clearly touched a nerve, so I'll just assume I won and we can leave it at that, ok?
Can you explain why you keep mentioning Bill Gates as if he wrote the RFC's? It just makes you look crazy without any context.
You're hardcore whining. Gaslighting won't work.
Bull, fucking, shit. If not outright lie, ask them to come here and explain their reasoning.
But you refuse to look up the problems with IPv4 and refuse to acknowledge actual problems. You can't discuss on a technical level, only emotional.
yoursunny summer host Inc has a secret deal with Starfleet Headquarters that we will start providing experimental IPv4 service out of Class E addresses someday.
All of Class E ranges are mine.
This wasn't a problem in the 90's and it sure as shit isn't a problem in 2021. Why do you keep making shit up? Do you even know how the headers would be processed? If so, you wouldn't be crying 128 bit all the time. The fact you can't point to a single company saying the same thing in 30 years should tell you you're wrong.
Jsg: Someone should tell pi to fuck off with all the digits when 3.14 works for most.
IPv6 is already the priority and falls back to IPv4.
I can't make sense of what you're saying. You can't just make up your own definition of backward compatibility that includes making changes. That's bonkers.
If IPv4 needs a rewrite to work with its replacement, the replacement isn't backward compatible.
If IPv4, without changes, can't talk to the new IP address scheme, then it's no different than replacing with another version that can be rewritten from ground up instead of bloating into a bigger IPv4 mess. So all the drawbacks and no benefits.
Perhaps you're not aware that many devices have IPv4 are implemented in silicon, not software with upgrades. Seriously, this is just silly, really.
You can buy a PCIe 3.0 x16 card and put it into a PCIe 2.0 slot, but it'll only ever work as good as a PCIe 2.0 slot. The PCIe 2.0 will never have the bandwidth of a 3.0 slot. At some point, you need to replace the mobo and use PCIe 3.0 or be stuck in the past kludging shit into PCIe 2.0.