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
Ok, let's say every single person in the "future world" would get 1,000 /64 IPv6 subnets. In that case, the population would have to be 18446744073709551 people. Given the earth's total surface area (196.6m sq. mi), that means there would have to be over 93,685,851 people per square mile.
We might not know the future ... but there's definitely a limit to how many people this Earth can hold. And that limit is significantly smaller than the number of IPv6 addresses out there.
You're clinging on to the age-old argument that we're going to "run out" of IPv6 if we're throwing it around so freely. But if you just took a second to do some basic calculations, you'd see that the exponential scale of IPv6 vs. IPv4 makes that pretty much impossible.
Honestly ... you said it yourself in your original post: that you don't work with IPv6. IPv6 wasn't designed to operate the same way as IPv4. You shouldn't be throwing around claims when you're not even familiar with the subject. There are plenty of technical reasons as to why people need such large subnets (again, such as SLAAC).
Nope. I'm clinging to reason, reason as in "do not waste IPs. Wasting IPs based on "totally reasonable premises" already led us into trouble.
And I'm clinging to "Why the hell would everyone need 4 billion times 4 billion IP addresses as a "reasonable minimum"? Why not handing IPs out generously - but not insanely? Say 16 IPs per person ... and if they really need more and can explain that sensibly and realistically they can get many more as in "as many as todays total IP4 based address space".
Your premise can not be proven true but a very similar one already was proven wrong.
The reason is that IPv6 was designed with these large subnets in mind. Again, bringing up SLAAC, which is designed to work with a /64 ... no smaller.
IPv6 was designed to be an upgrade for IPv4 ... not a one-to-one replacement. For that, some "sacrifices" would have to be made ... such as a few wasted addresses here and there. However, this was accounted for when IPv6 was designed.
If they just wanted more address space with the same numbering scheme ... why not just add another octet to IPv4? Why did they have to quadruple the number of bits representing an IPv6 address? Because IPv6 isn't a one-to-one replacement for IPv4.
Don't get me wrong, I honestly don't mean to attack you but I'm under the impression that you preach IPv6 but didn't fully understand it. Let me explain:
SLAAC isn't a good thing but a PITA. To provide just one prominent example: DNS, especially recursors. And btw. there were basically wars (e.g. between Google and Microsoft) about that (with both acting aggressive and idiotic).
Plus, in reality most people using IPv6 use DHCPv6 and not SLAAC. The reason should be evident: SLAAC basically introduces a funny random generator.
From my point of view SLAAC is just yet another reason to be opposed to IPv6.
As for more addresses, I submit that had they proposed IP5a with 64 bit addresses and some sensible extensions to all the relevant protocols instead of pulling the IPv6 "we do everything differently" monstrosity a major part of the internet would be upgraded to IP5a by now. Yes, that would only offer 4 billion times 4 billion IP addresses but that would be plenty enough for quite some decades to come ... (and by then the world will have changed so considerably that it probably doesn't make sense anyway to try to predict what comes after IP5a).
When you consider IPv6 as a mathematical formula, there is no waste of resources.
To put it simply, IPv6 with 128 Bit has more than 9.22 long scale trillion (10^18) /64 subnets which means there are more than 140.73 short scale trillion (10^12) /48 subnets. Usually, a person or a company won't need more than one /48 subnet. The fact, that the most clients won't need more space compensates the cases where persons or companies needs multiple /48s.
But even if we have a loss of 75 % due to waste we have more than 35.18 short scale trillion /48 subnets left. That is – remember, we're speaking here about /48 and not /64 subnets – exactly 8,192 times the theoretical space IPv4 can offer with 32 Bit.
It already needed decades before IPv4 became a scarce resource. Even if we would use as many IPv6 /48 subnets every month as IPv4 single IP addresses has, it would take us 682 years to exhaust the IPv6 space. That said, the grandchildren of our grandchildren are absolutely not going to exhaust IPv6.
@Kaulkwappe
One big problem though is that none of those approaches - incl. IPv6! - really are free. All of them eat processor cycles and memory (which again must also be read/written) ... and all of that requires completely new architectures and/or processors ... which, oh wait, might be yet another reason why IPv6 did NOT take off as planned multiple times. Those "small details" actually bring along billions and billions of cost.
So, how about looking at it from the other side and asking "can we solve IP scarcity with available and cheap resources (like 64-bit processors/ASICs/FPGAs)"?
I'm glad to inform you that, yes, we could, hurray! Plus we could stay with most of the protocols and only need minor changes. The price to pay? Every person may have only a couple of THOUSAND IPs, unless they can demonstrate sound reasons for needing more (e.g. ISP, hoster, ...).
Sounds damn good enough to me.
We don't need IPv7, so we don't need to think about that.
Might be, but the main reason for the low deployment is or was that IPv6 wasn't needed in many cases. There is (still) no technically reason to use IPv6 if you run a stable system with IPv4 and you have some IPs left.
We are. And we have seen that it took very long time to get IPv4 exhausted although so many IPv4 addresses were wasted and aren't still used for anything.
You are right. But exactly the same is true for IPv6. We do not need 128 bit IP addresses.