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
BTW, @Calin how is IPv6 going? I want to make a tutorial on how to setup a 6to4 gateway using one of the cheap nat boxes even with OVZ and your is the only one I have atm. I need IPv6 working for making sure it works :P
Currently I am testing http://www.litech.org/tayga/README-0.9.2 + nftables rule to masquerade IPv4 dynamic pool, works fine so far.
@Maounique didn't you say that a box with Calin will grant you IPv6 connectivity out of the box? Asking because maybe going with such solution might leave you both without IPv4 and IPv6. Scuze.
Hello , @Maounique 2a0f:9400:7e11::/48 it's allocate on our ISP ASN , I m waiting for Orange to route...
Regards
Maybe a question of perspective. The perspective of the vast majority of people is not "is it a well designed protocol?" but rather "can we use it? Is it easy and simple to use?" and "how much do I have to relearn?".
As for "new": it's (perceived as) "new" as in "it came after IP4".
Be that as it may, IPv6 will not take over and rule the internet as long as there are people whose experience is like that of Matt Duggan - and that will continue due to intrinsic problems.
Finally, we here, especially the techies, can continue to discuss the way we do - to convince and win over normal users you need to address their perspective.
What I said is that, if you can get a cheap box with both IPv6 and IPv4 (even an OVZ with NATed IPv4) you could make an IPv4 gateway out of it for your IPv6-only mesh. Currently @Calin doesn't have IPv6 working so it does not qualify, but if he were to have it, I would have tested a few setups and make a tutorial for the one I find the simplest and easier for newbies in the idea that, if it works for OVZ with NAT-ed IPv4, then it should be working in any other common type of dual-stack setups.
I am looking for a different thing, probably would go with a proxy of some kind, it would have the added benefit of caching, filtering ads and whatnot, people might learn a lot in the process.
I am thinking of forwarding ANY request that fails to the proxy which would take it up and then look up through the IPv4 network anew, like the app in need of a connection would be running locally. ATM I have some ideas but not much time, once @Calin implements IPv6 I would be starting rethinking it. It should also be scalable so an ISP, instead of CGNAT-ing which would make it impossible to host anything anyway, would implement a local solution to go out over the IPv4 when needed, regardless of the protocol by re-writing the address transparently and proxying the results back to the user in case their query failed over IPv6 for any reason so a kind of fail-over is implemented. It is very ambitious, I know, but if it is not impossible, then I don't think I should be thinking of it :P
I am not saying your approach is flawed in any way, just that a proxy would be more agnostic about the underlying setup and would offer the opportunity for other services and extended learning. If that is too complicated (as I fear it to be, tbh) then I would fall back to a simpler albeit less versatile solution.
@rcy026 to be fair, the reason you don’t know why it’s so hard for Github to roll out IPv6 is because you’ve got zero insights
If you had, you’d likely know
Well, true, that is why we are both curious about why they didn't manage yet, none of us had to manage such vast and complex setups, in theory they should be able to find a solution in any situation, even ugly ones that work are better than none, the fact that they couldn't in a decade is puzzling and makes me think not enough importance has been allocated to this. It is still working, very few people left over this, we can still plod along and say we are working on it.
I'll bite - what's the reason they haven't rolled it out (despite having said they're "working on it" for literally years)?
I mean, large scale websites can manage it (hello, Farcebook). Git itself works fine on it (hello, every other git-based solution). What special sauce does Github contain that makes it somehow incompatible?
unless elon invents IPv69
To be fair I think that IPv4 holdouts do not state that anything is incompatible (barring apps written specifically single-stack), instead they claim that rolling out IPv6 is too costly and is not justified by the tiny percentage of IPv6 users and that using various hacks in order to extend the IPv4 usage and the IPv4 ranges renting or buying is a cheaper solution (regarding money, time and maintenance) at least for the time being.
In essence, they say that, since github didn't roll out IPv6 completely yet, then it means they did a cost analysis and decided it ain't worth it for now, therefore them holding out on the legacy internet is justified as well.
I don’t know, nor does anyone here
Arguing is just silly when nobody knows the facts, it’s like two blind guys fighting over who’s got the biggest dick
@Maounique ”ipv4 holdouts” lmao
I even thanked some guy for making his website avaliable on IPv6 in this exact thread
You’re obviously feeling some kind of human connection and sense of belonging through your v4 vs v6 fighting but I’m really not that interested in participating, not on any of the ”sides” in this ”battle” of keycaps
I don't mean you or anyone in particular, I was referring to the way of thinking which generates such strong opinions in some people.
This situation resembles very much the alt-right thinking on fossil fuels. Conservative, passeistic, longing for a "great" past and trying to restore it to "it's former glory" while the rest of the world moved on long ago and there is no way back. The attempt to hold back progress will likely fail and will bring a lot of pain to everyone.
He asked me a question and you were ”explaining” how the ”holdouts” reasoned lol
But sure, all good
I was actually trying to explain to him that would be a strawman argument, because the holdouts are not claiming that, none here, anyway, so it would be pointless to ask that question to anyone, holdout or not. Not even @jsg claims that things are incompatible with IPv6, just that it is not worth (or needed, for that matter) implementing it yet and that is a completely different argument. We should be questioning that, not things we imagine they might be thinking of.
Ahh, so you have no actual idea if it's hard for them technically, or if management just decided they didn't want to because someone like jsg told them IPv6 sucks.
Because your "you don't know it's so hard, you've got zero insights" kind of made it sound like you knew there was a good reason.
Just like with every Microsoft Windows release? And yet we survive.
Well, if the pro-IPv6 crowd here doesn't trust me, how about the network guru of rackspace?
He said that not a single customer of theirs has requested IPv6 (at least as of about 1 year ago).
Plus he said (my summary) that IPv6 is basically a major pita, one major reason being that IPv6 willy nilly changed quite a bit in the "mechanics" beyond the addresses/address space. Another major criticism by him is that IPv6 extending addresses absurldy/insane to 128-bit addresses.
Funny side note: I seem to actually be the more generous critic (suggesting 64-bits). That man says that extending IP addresses to 40 bits or 48 bits (ca. 65000 times todays IP4 address space) would be plenty enough.
And he paints a nice a funny metaphor, saying that according to scientists there are about 6 billion earth like planets in the milky way and that if each of those had as many grains of sand as Earth has, IPv6 would still provide a couple of billion IPs for each grain of sand in the whole milky way.
(Now I'm waiting for a certain particularly dumb but loud a##hole to state that rackspace's network guru is a clueless wannabe expert with no education ....)
I have no idea, and neither does anyone here, even though some of the V6 fighters are implying that it’s just a switch that nobody at Github bothered to switch, but they don’t know that, it’s just their view disguised as an argument
To trust you what? Are you some sort of expert on this matter? All what you gave us are opinions.
Besides that you mostly argue with the people who know about this topic way more than you.
IF your comparison even made sense it would trigger the response "and yet IPv6 - after over 20 years - still fails to replace IP4".
Btw, about the only major change since more than 15 years in Windows I noted was when they changed the home page (to "tiles" or somesuch). But I admit that I very rarely use Windows and hence may not have noticed other significant changes.
(I just saw your new comment and found it unworthy of a response).
Perhaps, but the technology progressed so much since IPv6 was "invented" and we can now afford huge routing tables (and have to, because everyone and their dog can now afford to have own allocation).
The routing tables size issue would have appeared anyway, it is growing with IPv4 as well and has little relevance regarding the size of the allocation, /29s IPv4 or /36 IPv6 routing, the determining factor is the lookup time and the number of entries and that is stack-agnostic, the only difference is in the actual bit size taken in the memory, but memory is much bigger and cheaper today than 20 years ago. Even the number of entries and looking up those have been improved with faster and lower powered processors and memory even cheap routers of today could have.
Nuh, you (and obviously myself) just are IP4 zealots - just like the Github people. But that's no problem according to quite a few "experts" here because Github isn't that important - except for the fact that lots of vital software from compilers, network stuff, OSs, and other vital stuff wouldn't be easily obtainable.
It's quite obvious that the Github people are just clueless and evil, really, really evil. Cause, you know, IPv6 "just works fine"(TM) and will replace IP4 "very soon"(TM) since over 20 years.
O, ffs ...
Apple requires apps to support IPv6-only networks since many years. https://developer.apple.com/news/?id=05042016a
Great, stuff like this can actually make a difference
Imagine if Google started announcing SEO benefits to IPv6 reachability
If I haven't dreamt about it, I think there are some. I might have read some place but I am actually hating SEO stuff so not in the loop at all.
One more reason to avoid apple. I occasionally worked with one of their products but never owned one. Thanks for confirming that my decision was good one.
Works fine, yes, will it cause IPv4 extinction very soon, no, but that is because of political, not practical reasons. Many people would continue to subsidize the IPv4 brokers by buying or renting and those would have a lot of money to push back against IPv6. As I said, a lot of similarities with the fossil fuels saga.