[yakcyll]

One thing that struck me was the sizes of blocks assigned. I get a dynamic /64 prefix from my ISP at home, which would be large even if it were assigned to my work office; why is the maximum prefix length /48 for a single site and /32 for 3k sites? Aside from the obvious argument of wastefulness, aren't we just priming the same issue we have with IPv4 now to occur thirty, forty years down the line?

[jeroenhd]

ISPs handing out /64 are quite stingy, /56 is the recommended range to hand out to clients. It used to be /48 but people got worried about address range exhaustion so they changed it.

The reason for these larger blocks isn't that you need several hundred billion IP addresses per se, but that IPv6 can't create subnets (without terrible tooling issues) smaller than /64. In a way, getting a /64 from your ISP is like being forced to use a router that's stuck in the 192.168.0.x space for DHCP. A /56 will give you 255 subnets, a /48 will give you 65k in total. More than enough I'd say.

A /32 will give you as many subnets are there are IPv4 addresses out there today, I don't see why you'd need that. It's nice of them to offer it (for a significant price, of course) but I don't think businesses really need address space that huge.

IPv6 has a ridiculous amount of address space, we may as well use it.

[lokedhs]

My ISP gives me a static /48. I'm currently using 5 subnets, so I could probably have survived with a /56. I'm glad I have a full /48 though.

I switched from my previous ISP because they only have me a /64. It was quite honestly useless for me, since I couldn't even split it into two subnets. They did it because they clearly had old network equipment and were using 6rd to provide IPv6.

[GoblinSlayer]

I believe all ip4 lans I saw survived with just one subnet. What's the need for more?

[philjohn]

I'm with Zen in the UK and they give out /48's which is nice, and makes subnet management much easier.

[aaronax]

/64 is not large--it is only enough to run one network / broadcast domain. It is fine for 90%+ of homes, unless they want to do anything like run a separate guest or IOT network. Hopefully they would be able to obtain larger prefix delegations by simply requesting one with their router.

A /48 for a site allows a decent number of subdivisions along the easily human-readable nibble (16 bit) boundaries. Four characters each can be 0 through f.

A very small portion of addresses have been allocated so far. "According to the IPv6 Global Unicast Address Assignments list from IANA (last updated in Nov 2019), there have been 33 allocations made to the five Regional Internet Registries in total so far. This is equivalent to about 7,396,864 IPv6 /32 subnets which is approximately 0.172% of the total available IPv6 space." https://www.cidr.eu/en/ipv6

[throw0101a]

> A /48 for a site allows a decent number of subdivisions along the easily human-readable nibble (16 bit) boundaries. Four characters each can be 0 through f.

To put it in IPv4 terms:

* an IPv6 /64 subnet is equivalent to 'typical' IPv4 /24 (though you can fit much more than ~250 hosts in it)

* if assigned a /48, this gives you 16 bits to play with

* if you start with a typical IPv4 /24, and would be assigned 16 bits to use, that would bring you up to a /8

So the 'bog standard' IPv6 /48 is the equivalent of an entire IPv4 Class A address.

Some folks who have Class As assigned to them: AT&T, Apple, Cogent, Comcast, multiple assignments to US military.

* https://en.wikipedia.org/wiki/List_of_assigned_/8_IPv4_addre...

And none of those IPv6 addresses have to be NATed to be accessible to the Internet if you wish to provide public services: just change the config of your firewall from default-deny to allowing whatever portions of the network you wish to host service in.

[hampereddustbin]

Doesen't /64 mean that you can't create additional subnets within your ISP given range? I thought /56 was the smallest allocation an ISP could make for a residential allocation.

I think it's great that the smallest subnet size is designed to be as large as to never run out of addresses in any conceivable application, no more wasting precious time manually assigning addresses and thinking about subnet economics

[aaronax]

> Doesen't /64 mean that you can't create additional subnets within your ISP given range?

Effectively yes.

> I thought /56 was the smallest allocation an ISP could make for a residential allocation.

They can and often do make /64 allocations. There is an RFC (I think, might just be RIPE guidance or something) that recommends that ISPs issue larger to each customer. Many don't (as it is just a recommendation). Ideally they would allow a customer's router to request a larger allocation like /60 or /56 via a prefix delegation message.

[jeroenhd]

/56 is the smallest they're supposed to allocate for customers, but I've read stories about ISPs providing people with /128s on their CPEs...

A /64 can actually cause problems if you're chaining routers together. In IPv4 that'd give you double NAT which is obviously terrible and not recommended, but in IPv6 that's a fine use case that shouldn't cause any trouble as long as you have the ability to create sufficient subnets. With a /64, you're stuck doing weird stuff with DHCPv6 to get the subnets to work regardless.

[aidenn0]

You don't need double-NAT with ipv4 to chain routers together, just route between different private-allocation subnets. I was running a setup like that to get ethernet access to a part of the house via wifi. Changed it for ipv6 though because my isp only gives me a /64

[tsimionescu]

In previous threads, there were even people saying their ISPs provide them a single IPv6 address, so essentially a /128.

[staringback]

They may be misunderstanding how their ISP provides IPv6 address space. For example my Comcast connection assigns a single /128 to the router, then a bunch of /64s to each subnet I have set up.

[sigio]

Just see it as a 64-bit address-space, as one /64 will be for 1 home/location/lan.

Then we still have 32^32 times the amount of address-space we have in v4.

[VLM]

/56 aren't as generous as you'd think as companies "often" use VLANs and people selling network gear have always pushed for microsegmentation (rather than having a VLAN for the entire 3rd floor which would technically fit, have a distributed VLAN just for the three accounting people and their eight servers).

In theory you could have 10 bits just of VLANs without doing microsegmentation and strange virtualization games which everyone is encouraged to do, so smaller than a /54 for a corporate ISP account seems very questionable; may as well round each site to /48.

There is a high human labor cost to customization even with computer assistance for IPAM. Life is faster, simpler, and more reliable if "every generic ISP connection gets a /48"

At some point, for "IoT" and "security" reasons the concept of one broadcast domain per residential home will go away, hopefully soon. I don't want my soon to be exploited smart TV to have any access to my "real" VLAN, for example. My "home" and "home-guest" wifi networks should be on separate VLANs on separate /64 address blocks.

[tialaramex]

> aren't we just priming the same issue we have with IPv4 now to occur thirty, forty years down the line?

Why? Do the arithmetic.

[ninkendo]

2^128 is an enormous number. 128 is not.

If you start assigning semantic meaning to the bits in an address (the trailing 64 are the devices, the leading 0-31 are the ISP customer, 32-63 are the subnet) then things really do start to exhaust if you have a use case where the lines blur (a multitenant datacenter for instance, where it's not clear who the "ISP customer" is and different tenants want their own subnet ranges, etc.)

There's a lot of IP's, but it's easy to paint yourself into a corner if you make the wrong assumptions about what bits should mean what.

[iso1631]

Certainly giving a /32 to anyone who wants one is a problem

Giving a /56 or a /48 isn't.

There are 281 trillion /48s

If you gave everyone who has ever lived and gave everyone a /48, you'd have 281 trillion left

[ninkendo]

> If you gave everyone who has ever lived and gave everyone a /48, you'd have 281 trillion left

What if I got a /48 and my customer wants a bunch of /48's? Better ask my upstream for a /40 so that I can give it to them. What if my provider already has a bunch of people that want /40's? They'll probably want a /32. What if their provider has a lot of customers that want /32's? (and so on.)

I'm not saying any of the above make any sense (you don't typically have that many layers of "providers"), but I've seen it happen in really huge corporations that have this kind of logic:

"It'd be nice to know where a packet comes from using only its address, let's give a /64 to each server, a /56 to each rack, a /48 to each aisle, a /40 to each server room", etc... but suddenly your needs change and your servers themselves need to have multiple /64's, or you start needing to add regions to the list of stuff packed into an address. Suddenly you're asking for a /32 in order to keep your inventory management simple, even though you're barely using any of the addresses.

I'm not saying the above is a good idea, just that if you do dumb stuff with your address space, you can end up running out of addresses even though you have many many orders of magnitude more addresses than you actually use.

[bauruine]

As a LIR you get a /32 by default, no questions asked. If you need more you need justification and from my experience with RIPE and PI space they can be a pain in the ass regarding that. AFAIK the biggest one they ever give a LIR is a /28 with the note that they never ever want to hear from you again. There are 268'435'456 /28 available. I don't think i will still life when we reach the only 200 Million /28 left mark.

[Arnavion]

Also all the current allocation standards are for 2000::/3. If it turns out we have some unforeseen issue with the current policy of handing out blocks we can come up with something else for the five other /3's that are completely unused.

[ninkendo]

Giving a /32 to anyone who wants one without any questions asked is probably a good example of why "aren't we making the same mistake as ipv4?" isn't an entirely meritless question.

I mean it's probably crazy to imagine a scenario where we have more than 4 billion "local internet registries" in the universe, but it's not so crazy to imagine a scenario where the ability to ask for a /32 maybe needs to get constrained a little more.

In the end, I think the idea of making all subnets /64's may have been a mistake. It's cool that you can put MAC addresses right in the address and thus don't need DHCP any more, but carving out room for 18,446,744,073,709,551,616 devices per broadcast domain is a bit crazy on the face of it. Plus you need all the privacy addresses to hide your MAC address from trackers anyway, which is way more complexity than we really needed. IMO DHCP isn't so bad, it's boring/predictable technology at this point, and I'd be perfectly comfortable with shifting the "provider vs customer" boundary many bits to the right in an address, so that a typical subnet is more like a /104 and ISP's give out /96's, etc. That would give us a lot more headroom than the system we have now where anyone can grab a /32.

[HyperSane]

The last 64 bits of an IPv6 address are the host bits. I think it is stupid also, it should have been 16 or at most 24 bits for the hosts.

[postultimate]

The last 64 bits are derived from the MAC of the interface, which is quite a good idea (we probably won't see MAC exhausion because it's only two-level, and it's supposedly guaranteed unique). The stupidity lies eleswhere.

(Seriously. Just how fucking stupid do you have to be to replace a protocol whose addresses are too short, with one whose addresses are too short.)

[HyperSane]

Except MAC addresses are 48 bit.

[throw0101a]

> I get a dynamic /64 prefix from my ISP at home, which would be large even if it were assigned to my work office; why is the maximum prefix length /48 for a single site and /32 for 3k sites?

The default subnet size for IPv6 is /64, and a single "site" is /48. There are 16 bits between those.

Comparing with IPv4, where a 'typical' subnet is /24, if you were given 16 of space to play with as you see fit, you'd be assigned a /8—i.e., an entire Class A. (Which is what most companies use now anyhow—i.e. 10/8—and then have to futz around with NAT.)

So a 'typical' IPv6 allocation is as many IP addresses as what some of the largest corporations have. Plus all of those addresses are available for use on the public Internet is you wish: just change your firewall from default-deny to allow certain segments.

> Aside from the obvious argument of wastefulness, aren't we just priming the same issue we have with IPv4 now to occur thirty, forty years down the line?

No. The numbers involved with IPv6 are literally astronomical:

* Stars in the Milky Way: 400 Billion

* Galaxies in the universe: 2 Trillion

So (4x10^11 )x(2x10^12 )=8x10^23 stars in the universe.

* Size of IPv6 address space: 3.4x10^38

Find the ratio between addresses and stars:

* 3.4x10^38 / 8x10^23

IPv6 offers about 430 trillion times more addresses than estimated stars in the universe. From Tom Coffee's presentation "An Enterprise IPv6 Address Planning Case-Study"

* https://www.youtube.com/watch?v=7Tnh4upTOC4

Another way of looking at it:

* math property: x^y = x^(a+b) = (x^a )x(x^b )

* IPv4 addresses are 32 bits (2^32 )

* 2^32 ~ 4.3 billion

* So the IPv4 Internet has ~4.3B devices on it

* IPv6 subnets are 64 bits, /64 (2^64 )

So, a IPv6 2^64 subnet is the same as (2^32 )x(2^32 ), which means (4.3B)x(IPv4 Internet). I.e., a single IPv6 subnet can hold the equivalent of four billion (IPv4) Internets.

A third way:

* On the surface of the Earth (land+water), there are 8.4 IPv4 addresses per km^2. Not counting the oceans, that would be 28 IPv4 addresses per km^2 land.

* IPv6 gives 10^17 addresses per mm^2 (yes, square millimeter).

In terms of volume, 10^8 IPv6 addresses per mm^3 throughout the Earth.

* Via: https://news.ycombinator.com/item?id=28326806#unv_28331245

[ikiris]

No, because math.