Domestic IPv6

Interested to see that Zen are offering residential customers the option of joining a trial. Accepting that IPv6 will be required eventually, is there any benefit to an ordinary residential customer, to take part in the trial now? Any potential downside?

Reply to
News
Loading thread data ...

No downsides, certainly: Zen are offering IPv6 in addition to the existing IPv4 service, not as a replacement. If you join the trial and decide you don't like it, you can just turn off IPv6 on your end and everything will be back as it was before.

Benefits to an ordinary residential customer are minimal at the moment but will grow with time. Some websites may be marginally faster over IPv6 than they are over IPv4 (for example, Facebook is all-IPv6 internally and IPv4 users access it through a proxy). A very few services may be accessible over IPv6 and not over IPv4; the number of these will surely grow but all major services will remain accessible over IPv4 for the foreseeable future.

Benefits to somewhat-technical residential customers may be more significant. For example, I find it very useful to be able to use ssh between various computers at various sites I run without having to mess about with VPNs, port forwarding, and so on.

Stephen Early

Reply to
Stephen Early

I was looking at this yesterday. What I couldn't work out was if I went on IPv6 trial at home would the router in effect run as 2 systems, i.e. IPv4 Wan side would still remain intact as would all internal LAN addresses but they would also run a parallel IPv6 address so for example my VPN between home and work would still work via private IP addressing.

Work router isn't IPv6 compatible (though I have a spare that's the same as the home router which could be changed out)

Also... :) The admin side of our website is locked down to only the home and work IP address and unless the web host upgrades to IPv6 then I'd still need IPv4 addresses to remain present for server access.

And... :) With IPv6 does every single allocated IP address to every device in effect become a static IPv6 address?

Cheers Pete

Reply to
www.GymRatZ.co.uk

A&A have offered IPv6 as standard for > 10 years now.

Yes, you should get lots of fixed routable IP addresses without NAT. OTOH, you do need to make sure you firewall well, as you won't be getting the protection NAT gives you against external access. Also, the IPv6 address space is to big and too holey to do viable scanning for systems, although you certainly can't assume that means no one will try.

Reply to
Andrew Gabriel

I've been using IPv6 for 4-5 years now via Entanet. It seems to "just work". My network is "dual stacked" and the Internet side of things tries IPv6 first, then v4.

We've been (and have in many places) running out of IPv4 addresses for a long time. I'm not sure that everyone will ever be fully IPv6, but as time goes on more and more of the Internet will.

So no reason to not go for the trial - if your router will support it.

Gordon

Reply to
Gordon Henderson

You can play Loops of Zen:

formatting link

IPv6 only.

Reply to
Ron Lowe

In message , at

11:41:55 on Thu, 19 Nov 2015, " snipped-for-privacy@gmail.com" remarked:

General rule: any pundit claiming there are undecillions of IPv6 addresses available to everyone has no idea what he's talking about.

Not only is every end user given at least a /64 (and it could easily be a /56 or /48), but 7/8 of the address space remains "reserved" - perhaps for whatever succeeds the Internet.

As a result, the number of end users it can accommodate might be "only" around 16,000 times that of IPv4, once you've excluded the latter's reserved space and /8-hogs.

16,000 may still be a big multiplier, but it's far from the almost-infinite that the undecillion brigade continue to promote.
Reply to
Roland Perry

They're trying to be more careful regarding address allocation this time round. With IPv4, addresses were initially given out like confetti and with no regard to geography. That has led to large router tables. Later on they tried to be rather more careful, but the v4 router table size (I haven't looked at it for a few years now) is still of the order of several hundred thousand entries.

That aspect is much stricter under v6.

Reply to
Tim Streater

I did wonder why they went to the fairly insane /64. Would have thought a /96 would have been perfectly good (that's a "Whole Internet of IPv4 addresses" for each and every company and person.)

Given that even mega corporations tend to be split over physical sites, a number of /96s would be perfect for them - one per site.

Reply to
Tim Watts

In message , at 16:19:19 on Fri, 20 Nov 2015, Tim Streater remarked:

Confetti suggests lots of small allocations, but actually it was the other way round with far too large blocks being allocated to organisations who didn't need that much.

IP addresses are not correlated with geography, but with networks. To some extent networks are geographic (for example Demon only had a presence in UK and NL) but the suggestion of a correlation tens to make toes curl in the RIR world.

Departing from Class A/B/C allocations did that, caused by the explosion of demand from 1995 onwards.

It keeps busting various limits in routers. Currently (20th Nov) it's

579999 or 315187 CIDR Aggregated.

v6 had a concept that you could only get sub-blocks from your upstream, which would help with aggregation. And with the large blocks the upstreams were given, they may well only have one to be routing. But this effect is more of a happy coincidence than a design goal.

And then there's all those people who don't have an upstream (IXPs were the first to bring this to the community's attention).

Reply to
Roland Perry

You need at least 48 bits in the local address for stateless autoconfiguration to work (ie a machine can figure out its address by listening for broadcasts of the prefix and then appending its MAC address, with no need for a DHCP-style allocation). I suppose you then need another bit to say 'the address I randomly picked isn't based on a real MAC address' to avoid clashes, at which point the 49 bits get rounded up to 64.

That's basically how it works, but with /64 instead of /96.

Note that it's a lot easier to change the address allocation policy down the track than change the packet format. So if 2^64 network addresses turns out to be too few we can move the /64 boundary and come up with new configuration protocols (and ISP-level routing setups) but we don't have to change every router that the packet passes through as we do with the IPv4 to IPv6 transition. The Internet routers along the way don't care whether your local network is a /64 or a /56 - they just ship along the packets based on the routing tables they have.

Theo

Reply to
Theo Markettos

Thank you for that - very informative!

Reply to
Tim Watts

But all the relevant routers have been v4/v6 dual-capable for years. At least 10.

Reply to
Tim Streater

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.