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.
IPv6 Hall of Shame
This discussion has been closed.
Comments
Colons should be renamed "double nips".
It's because they only deduct IPv4 address cost as they sell it to someone else anyway. But the routers, equipment and manpower to maintain IPv6 they can't just sell to someone else so it's baked into the cost if you use it or not.
Anyway, I have found the PulsedMedia post with some figures and from the numbers there it seems the AI estimate is quite good.
The active routes need to fit in what some call the FIB and with dual stack you will need larger FIB or you'll need a router upgrade/replacement sooner. From the PulsedMedia projection for example 22 years for IPv4 and 5 years for IPv6 till the FIB space is exhausted for the 6.5M routes. But it seems the hardware is the "cheap portion".
From my perspective IPv6 is offered for free and I don't use it, it's simply double the work for everything and you can't run on IPv6 only. Contrary to popular belief IPv6 is not either more or less secure than IPv4(even the NIST publication states that).
APNIC has tested it at end of last year; run IPv6 only service and only a third of users can reach it. Assuming current trends IPv6 transition will be at end of 2045, but that's just speculation. USA for example sees no IPv6 growth since 2019. And they make a good point saying if the internet will run till 2045 without end to end connectvity what's the point of IPv6 and why would someone feel the need to restore end to end connectivity at some point in the future.
yo yo @DrNutella
What kind of numbers we need here
I can say this, since its a yearly biz, I aim to keep 1 Years reserves in the accounts at all times, its got about 6 months in right now. So still need to adjust the numbers, add more products etc then it will be in a very good spot. But we are 2000% over last year.
Join DediRock & Live The Dream™
Gayporn has been mentioned... What we discussing again?
A provider noticed that storage VPSs are often requested. Made them think about how and how much customers use those, and whether to pry on their customer's stash, perhaps enjoy themselves. An LLM was consulted to get a rough estimate, and this is what it responded: "The average storage VPS utilization can typically range from about 0% to 100% spicy, depending on the specific orientation, how many per day, and tastes of the customer." Perhaps @CharityHost_org can be of use to give some specific numbers.
You may be right, hitting the wall is a lot harder than hitting the road and yet so many people choose to hit the wall.
So keep trying to scan ipv6, I wish you luck.
Yes, IPv6 you get discount because it's more or less useless for most of the population.
As for end user pricing, not certain on VPS services how much it increases. Serverhunter and similar services would have the necessary data to do the comparison tho.
On the provider side tho, the cost increase is vastly more than 10-20% on the networking side for most part, for what only 1-3% use, and what is required only by 0.1% of the userbase. Maths do not math.
It's cheap or borderline free to add IPv6 tunnel tho, but even that is not free for seedboxes because it f***s up so much software, and some almost force ipv6 use if enabled, causing bad routes and shenanigans like that. We favor robustness, stability and reliability above all else, so it's simply too expensive to add and micromanage.
There are a plethora of ways IPv6 adds friction.
I would understand these complaints 20 year ago but not today when certain countries have half of the traffic on IPv6, mobile phone operators already support IPv6, dual stack on Linux works well for quite some time and commonly used software can handle IPv6 traffic as well. Even the most stubborn webhosts now support IPv6 while the big guys are automatically deploying dual stack servers without issues.
And everything has IPv6 to IPv4. CGNAT etc. is common too. Your argument falls flat on it's face. Further you simply ignore any argument against, spoken like a true religios zealot.
The friction exists, it is real, and people who have to actually maintain stuff people like you use see it clearly. The friction is so high that some larger scale carriers even recommend their customers to disable IPv6 always unless they have a use case which cannot be accomplished with IPv4 (so default disable for everyone).
We are talking about a company building "backbone" networks, carrier grade stuff, ISPs being bulk of their customer base.
Dafuq are you talking about
Thanks, I always knew they were one of the shadiest providers and this is just more evidence. File them under "lazy bums who hate doing their jobs."
Why do you as a server host have 6.5M routes? And why do you think router manufacturers didn't expand their FIB to match the . And why are you assuming IPv6 will have more routes than IPv4 instead of growing until it matches IPv4 and then stopping?
HAHAHA you don't like end to end connectivity? You know that Pulsed Media is a seedbox provider meaning they rely on end to end connectivity for them to exist at all?
If you want to recreate AOL and Ma Bell then go ahead and destroy end to end connectivity. Reading this comment will cost you $1 in such a world.
Did you know your cellphone is IPV6 only, and uses a gateway to access IPV4 servers, and that's only because so many server providers are too lazy to update? Your internet literally goes slower and your customers load your site slower when you don't use IPV6.
Google sees that 50% of users access them over IPv6.
https://www.google.com/intl/en/ipv6/statistics.html
It took a long time to reach the first 30%, but the following 20% were added in just the last five years.
ipv6 threads making let great again, appreciate the entertainment
Sadly a moderator deleted my parody of your thread, "VPS Hall of Shame", about providers who waste money providing a VPS in your VPS.
Any chance to weave it into this one?
I didn't say that. I said there are different methods of network reconnaissance and scanning is not the only one. Also, it's not a good idea to rely on security through obscurity. It's explained in the NIST publication.
How about a quote from Chief Scientist at APNIC, it's a statement from last year: "There was no real lasting benefit in trying to leap across to just another 1980s address-based architecture (with only a few annoying stupid differences, apart from longer addresses!)." If you don't understand it, then maybe consider you just need to give it some more thought.
End users always pay for it. Routers have a special kind of memory that can make decisions in a single memory cycle because anything else is too slow and would not be able to keep up with the packet arrival rate. When you buy a router it's the size of that memory that you actually buy as that's what matters.
As far as I know today the IPv6 routing table has 4 times less entries than IPv4 table. But every entry takes 4 time more space in that expensive memory. Supporting dual stack doubles the price as you will need to replace the router twice as fast. And when IPv6 table grows to the size of IPv4 table it will take 4 times more memory and you'll need a new router more often.
What you commented does not come from PulsedMedia but from APNIC Cheif Scientist, so here's another quote from him on the matter: "It's a 1980's concept whose time has come and gone. ... We'll keep on running this for a while as long as it's not expensive to run, but I don't think it's got any future. End-point addressing is just not there any more."
Destroy? Pehaps you didn't hear of large IPv6 peering disputes, so there's nothing to destroy. It's been gone for decades and it still doesn't cost me $1 to read this comment.
ipv6 provides security through hardness, make things so hard for the attacker that his chances of success dwindle to close to zero, that is the way real security operates.
It's based on math not obscurity, there's nothing obscure in the way ipv6 operates.
You understood me wrong. I don't care if someone scans my IP space or uses some other method to achieve the same. I can't rely on the assumption that it's impossible for someone to find what they are looking for, that would be security through obscurity.
IPv6 was not designed to be more secure, faster or anything. The main point was to mitigate the future risk of IPv4 pool depletion. That's not a problem any more and has been solved in different ways, that's why IPv4 is still here.
I used to be a huge advocate for IPv6 until I started running into routing issues, turns out if your provider isn't peered with both Cogent and HE on IPv6, then your server is inaccessible to a large chunk of the internet. So I think IPv6 won't be widely adopted until Cogent and HE resolve their peering dispute.
Also, since IPv4 has been around a lot longer than IPv6, routing and peering are better on IPv4 than IPv6 in general.
It's your ISP's job to peer with both of them, and if it doesn't, it's debatable whether it's actually sold you internet service.
Cogent does the same bullshit with IPv4 as well, but you don't know about it because every ISP learned this lesson already with v4 and for some reason didn't transfer it to v6.
Oh does your preferred provider only peer Yggdrasil? No IP at all?
The longest prefixes accepted in the default-free zone are 48 bits. Router makers know this. Some go up to 48, some 64 because not every network is the default-free zone.
That's just fake math
No wonder the internet in the APNIC region is utter shite. Didn't Twitch.TV ban the whole region because internet prices were kept high by Ma-Bell-style national monopolies? The same ones who hate end-to-end connectivity?
Perhaps you don't know what the world was like before end-to-end connectivity.
I had the same case, i just create a free HE tunnel to the near dispute free ipv6 endpoint of the network, it's work with 2 ipv6 interfaces, not ideal i agree but it's work fine for private networking over ipv6
Perhaps the quote is confusing for some. He's talking about IPv6 being another 1980s address-based architecture with no real lasting benefit to IPv4.
I didn't make the numbers up, this is a quote from 2023: "The absolute size of the IPv6 routing table is growing much faster than the IPv4 table. These two tables will require the same storage/lookup size in around 1 year from now, given that each IPv6 entry is 4 times the bit size of an IPv4 entry."
The guy is from Australia and is the one who was building AARNet.
But, what are you saying? That the internet in the region with largest IPv6 growth and small number of IPv4 addresses is "utter shite". India has the same problem, how is the internet there?
99.9% of people do not need it and they will not pay for it unless by force. Most people don't even know exactly what IP is or even if they do they don't care if IPv4 or IPv6 is used to connect them to the internet.
Count me in for an entry on the list. Fuck IPv6.
this discussion has convinced me, ipv6 is the devil.
I expect the number of SMTP servers which only accept inbound mail on IPv6 is pretty low, and the volume they get lower still
But the number of SMTP servers accepting inbound mail on IPv6 should already be reasonably high (and growing - Google accepts mail via IPv6, and most of Microsoft-hosted domains also accept mail via IPv6 now). And with IPv6 you are less likely to run into IPv4 reputation issues (because of bad subnet neighbours).
Yes, but for the same reason mailhosts will just block whole subnets instead of individual IPs as they know spammers have far more addresses to spam from.
I still don't get your complaint about address-based architecture. Do you exclusively use Yggdrasil and telephone calls? How are you reading this forum? AARnet is using exactly the same address-based architecture, and methinks he just doesn't want to upgrade all his routers from 1989.
Do you know what the telecoms world was like before end-to-end connectivity? It was very shit for the end users. And then it was good, when end-to-end connectivity was introduced. And then it was shit, when end-to-end connectivity was taken away. You are here.
Just one little problem: without an IP4 a VPS is basically worthless if visitor's/user's don't have ipv6 available or don't use it.
Also, @maxxxxx just asked an innocent question and didn't attack ipv6; while I personally don't care for the included cost of ipv6 in my VPSs quite many ipv6 fans do care or even complain about the included cost of IP4 (if included), so what's the problem with OP's question?
Any tangible and credible evidence for ipv6 being allegedly safer than IP4?
We are talking about dedicated public IPv4, not IPv4 access in general.
Major cloud providers have long offered virtual servers with internal IPv4, which can be mapped with optional dedicated public IPv4 that costs extra.
A virtual server without dedicated public IPv4 can still be accessed through cloud provider APIs or reached through access gateways.
These are frequently used to deploy database servers and backend compute instances that are never accessed by the public audience, or web frontends that are published through the cloud provider's CDN service.
That's all nice, BUT: The vast majority of VPS users want and need a "dedicated" IP4 (I guess by dedicated you mean static and publicly visible/routable)
"to deploy database servers and backend compute instances that are never accessed by the public audience" - for those the IP4 addresses are only a fraction of the cost because there are multiple large to very large IP4 ranges for that.
TL;DR without an IP4 a VPS is basically worthless.