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
I can get hold of 5 /64
And 5 /48
From what I remember we are going to run out of v6 within 5 years if we keep on going like this
I love you so much. Please marry me.
To save yourself the embarrassment, no sir, you're unfortunately VERY mistaken.
It's better to not post at all than post just to 'increase participation.'
No.
> And 5 /48
That means there's only 281,474,976,710,650 /48 blocks left :O
We're doomed
only 281,474,976,710,650 /48 left, must start panic buying
We will get the the run out point as we have Ipv6 allocated at a click of a mouse,
If normal people get into Ipv6 we are doomed simple as that.
Nope, (someone else) correct me if I'm wrong.. but aren't consumers supposed to be designated a /64?
If they allocated /48s to everyone in the world (7.16 billion) each person would get 39,300
I think people underestimate how many IPv6 addresses there actually are
Not enough, I need at least 100,000 /48s
Seriously though, what is the use of a /64 for an end user? Also how would all those IP's fit into the networking page of solusvm!?
Also isn't there issues routing a whole /64 to an openvz VPS? ( @Magiobiwan )
Sell them to @mtwiscool
But somehow, someway, I just know there will be a shortage in 20 years.
Happens every time someone says "that's more than enough for any possible use".
How many more gadgets do you think people will use / get. It can't be anywhere near as many as ipv6 addresses and I don't see a reason why people need multiple ip's in their thousands per person, since you get so many ports per ip
You're looking at it the wrong way :-)
Every user is supposed to get a /64 and give a subnet, e.g. /96, /128, whatever you like to his devices in an effort to get rid of NAT.
The protocol isn't just a longer address, it's a complete redesign of addressing...
/64 is the minimum for auto assignment/SLAAC autoconfigured VPN on IPv6. https://en.wikipedia.org/wiki/SLAAC#Stateless_address_autoconfiguration_.28SLAAC.29 the EUI-64 mechanism for stateless autoconfiguration of IPv6 addresses requires a subnet to have 64 bits. This means most, if not all subnets (except point-to-point links), will have a size of /64 in the future.
Pretty much the only way to run of IPv6 before it becomes obsolete is human stupidity. That can't be fixed with any number of bits. Inevitably, at some point, someone will figure out a use that IPv6 was never intended for - and that will be the threat.
And really, 640KB of RAM ought to be enough for anyone...
I'd put a whole lot of money on the fact that eventually ipv6 will be exhausted. I will be through
proliferation of IP addressing. Only a dozen or so things in my house have an IP and they're all computers, mobile devices, or network devices. Someday every toaster will have an IP address.
mismanagement of the address space. When there is a "virtually unlimited" amount of everything, humans will use it up. There were once 5 billion passenger pigeons in the world...today, zero. People will hand out big blocks of addresses and the people will hoard them (another human trait).
In reality, ipv4 should be sufficient. I mean, who really needs more than 10.x, 172.16.x, and 192.168.x + one public IP? But human nature...
IPv6 allocations tend to be based more on efficiency and network logic than actual need. Also keep in mind, once that /64 hits you -- you are free to sub-divide it as you wish.
Is ipv8 next?
Here's the problem in a real-world example. In my own house, I have more than 200 individual physical devices that should have their own IP addresses. This is fine when using a 10.x range, but up until IPv6 I only had one public IP address, as dictated by my Internet Service Provider.
On these 200+ physical devices, are numerous virtual devices or instances of port-centric applications that must run on separate addresses, or it adds a usability deficiency as the end-user has to enter a port in their application as well.
That being said, I have a 10Gbps drop with 1Gbps commit (95th percentile); not something many people can get to their home.
@linuxthefish Well, it can make container booting a heck of a lot slower, as vzctl has to assign the IPs one by one. If each VPS has a /64 AVAILABLE to it, but doesn't actually have all those IPs ASSIGNED, it's fine. If the VPS has all 2^64 IPs assigned (18,446,744,073,709,551,616), then you get into problems. Say it takes 1 second to assign each IPv6 address (it takes less, but this is a simplification). Then each container would take a LONG LONG LONG LONG LONG time to boot. Also, I haven't done the math, but I highly suspect that the buffers for ALL those IPs would take a shit-ton of RAM.
They must be a more optimized way for ip's to be added like ram based compression?
wait, what are you putting(or doing) at your home?
200 network connected sex dolls.
They extrapolate my preferences based on my network shares. And I like to call them "Artificial Anatomy" instead of that derogatory term.
Thanks to the OP for bringing this to my attention. I've been able to give my proxmox containers their own IPv6 address using this tutorial:
http://mmaton.com/2014/02/ipv6-only-proxmox-ct/
What is your point? In IPv6 the entire fc00::/7 range can be used as you desire.
For someone who has 200 devices and a 10Gigabit line you seem to now know a lot about networking, something smells about this story...
Why would you do that? There's a reason CIDR exists.
If OpenVZ doesn't support CIDR, then that isn't IPv6 being poor designed or anything, but OpenVZ not working as it should...
What "math" is there to do? You have the number right there in your post. Multiply by 16 bytes (per IP), and you get the amount of RAM in bytes that's needed to simply store all these IP without anything else. Hint: perhaps even all servers in your DC don't have that much. In other words, it's an amazing idiocy to even half-seriously examine the possibility of having the whole /64 worth of IPs assigned one by one. There are now means to assign the subnet wholesale to an interface (AnyIP), but that has little relevance to the current discussion or SolusVM/OpenVZ specifically.