Thanks very much for the replies. AFAIK, the Billion BiPAC 8800AXL does support IPv6, so I guess everything will move over/switch when appropriate.

The ways ISPs use IPv6 differ, as it's a very different animal from IPv4. Your router will almost certainly need configuring, to a greater or lesser degree.

The Billion 8800NL R2 certainly does.

So, do we now expect everyone to have to start configuring their devices? I'm slightly confused as to how this all works.

You shouldn't need to if you take an ISP-supplied one, as they are generally pre-configured.

No different in principle from how it is with IPv4 really, but more complex. No doubt ISPs that expect you to use other kit will supply the necessary info just as they do now. And don't, if they wish to discourage you from doing so.

Fair enough. What is there to configure on a Billion BiPAC 8800AXL? IPv6 seems like a waste of time to me, why can't we just stick with IPv4!

Edited by deleted (Wed 05-Apr-17 19:13:01)

In reply to a post by WilliamGrimsley:

Fair enough. What is there to configure on a Billion BiPAC 8800AXL? IPv6 seems like a waste of time to me, why can't we just stick with IPv4!

The main reason is that the IPv4 address pool is exhausted - the registries have handed out all the address blocks, so once an entity has used up all its allocated addresses, it either has to go to the marketplace to try to secure transfer of address space from others (which is only a short term solution) or deploy NAT and hand out private address space to customers (so-called Carrier Grade NAT). Some IPv4 protocols that work well through a local NAT (in your router) do not cope with two layers of NAT - router NAT on top of CG NAT.

CG NAT has been the norm on mobile networks for years, because there isn't the address space to hand out public IP addresses to all the mobile devices. We are beginning to see CG NAT on consumer broadband - Plusnet have run a trial, and Hyperoptic are now forcing many of their customers onto CG NAT unless they pay an extra monthly charge for a single static IPv4 address (presumably because, as a relatively late entrant, Hyperoptic have a relatively small IPv4 address pool).

There are other advantages to IPv6 than a larger addressing space. The larger address space means NAT isn't normally used, making firewall configuration more straightforward as you don't have to worry about internal <--> external address mapping (there are some IPv6 usage cases, such as failover, that use NPt - mapping a prefix to another prefix). Inefficient broadcasts have been done away with; IPv6's multicast handles all the situations that used broadcasts on IPv4. Routing tables are typically smaller than with IPv4 because the address space isn't sliced up into so many small blocks, making life easier for routing protocols and routers. Most devices can auto-configure on an IPv6 network if allowed to do so.

There is a degree of implementation pain with IPv6, especially when it comes to deploying it to home routers, but things have got better over time, in no small part thanks to work focusing on standards like TR-187. The biggest problem at the moment is perceived lack of consumer demand; IPv6 deployment and especially technical support has a non-zero costs for ISPs, who have so far received relatively few requests for it.

However, the move to IPv6 is eventually unstoppable. IPv4 will become increasingly constraining as the Internet continues to grow, and once IPv6 service provision is mature, the way will be open to eliminate cost and complexity by starting the process of decommissioning IPv4 on the Internet.

If you want to ignore IPv6 for now, do so - you really aren't missing out on much. However, you cannot ignore IPv6 forever.

If you want to ignore IPv6 for now, do so - you really aren't missing out on much. However, you cannot ignore IPv6 forever.

Indeed. The industry itself still isn't properly geared up for IPv6. Until they are then home use of it is largely academic rather than life changing.

In reply to a post by David_W:

There is a degree of implementation pain with IPv6, especially when it comes to deploying it to home routers, but things have got better over time, in no small part thanks to work focusing on standards like TR-187.

But no thanks to companies like Netgear who do not support BT's IPv6 implementation: https://community.netgear.com/t5/DSL-Modems-Routers/...

Sky utilise prefix delegation too, making IPv6-enabled Netgear routers largely useless for IPv6 usage in the UK.

In reply to a post by Oliver341:

But no thanks to companies like Netgear who do not support BT's IPv6 implementation: https://community.netgear.com/t5/DSL-Modems-Routers/...

Sky utilise prefix delegation too, making IPv6-enabled Netgear routers largely useless for IPv6 usage in the UK.

Netgear's response is pathetic. DHCP-PD is the TR-187 standard way of handing out prefixes for local networks and support of DHCP-PD is mandatory under TR-187 (pretty much the de-facto standard for IPv6 over PPP) - see TR-187 section 6.1 paragraph R-10.

Unfortunately the firmware quality even of premium segment consumer routers is often awful. Things would be a lot better if routers were based on well tested open-source distributions such as OpenWRT and pfSense (the latter still being immature on ARM and possibly being too heavyweight for a typical home router, though it is excellent - I use it here). However, that wouldn't give the marketing advantages from brand and product differentiation that the big players want - they want a router they can promote as better than anything else on the market even though, at their heart, many of these devices are embedded Linux machines.

Most consumers will never notice if the IPv6 functionality is broken as few will miss IPv6 at the moment. By the time they notice, the chances are that their router will have stopped receiving any firmware updates. The support lifecycle of consumer products is depressingly low, not least because vendors selling to the retail market care very little once they have your money. Devices sold on long term contracts to ISPs stand a better chance of being fixed, but this is not always the case - especially with the relatively limited clout of a small tech-savvy ISP. AAISP seem to have struggled to get ZyXEL to fix IPv6 related bugs in the ZyXEL routers AAISP sell.

Getting IPv6 support right on a consumer router is not easy, especially because of the multiplicity of IPv6 presentations for IPoE (as used by Sky and cable providers). I did quite a bit of IPv6 debugging and improvement on pfSense on the late 2.2 and early 2.3 versions. Other third parties have carried on this good work, not least by fixing bugs and improving the features of the DHCPv6 and DHCP-PD client daemon used by pfSense. The pfSense core team have continued to support and improve IPv6 functionality.

Personally, I prefer a multi-box solution, using a separate VDSL2 bridge (Huawei HG612), router (a rack mount PC running pfSense), switch (ZyXEL GS1920-48HP) and Wi-Fi (HP MSM460). This allows me to select the components most suited to my environment and to choose a router I can debug and fix myself. I've got a decade of experience with FreeBSD and used to be a software engineer working for a networking company. This is not a typical home network, but we are not typical home users - our setup here is more typical of a small business and is used for business purposes.

In reply to a post by David_W:

Personally, I prefer a multi-box solution, using a separate VDSL2 bridge (Huawei HG612), ....

Now a very under-performing modem component, and most of those available getting long in the tooth for a cheap commodity item.