IPv6 discussions

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
Somewhat a hack -- working around for LAN/wireless clients behind Double NAT, since Singtel ZTE F8648P XGS-PON ONR does not seem to be able to delegate /64 IPv6 to sub-routers. Tests show that the following work-arounds work well under both Windows and Linux (should work under macOS and BSDs as well).

1) one of the sub-router uses Singtel 6rd (only one instance of 6rd allowed).
2) the other sub-routers running OpenWRT can use NAT66 (not supported by pfSense and Asus).
https://openwrt.org/docs/guide-user/network/ipv6/ipv6.nat6

In this case, I use Singtel 6rd on RT-AX86U; Using NAT66 under OpenWRT. No solution for pfSense yet.

Singtel ONR (native IPv6) -- LAN clients get IPv6 using DHCPv6
Singtel ONR (native IPv6) -- RT-AX86U in router mode with Singtel 6rd -- LAN/wireless clients gets IPv6 access
Singtel ONR (native IPv6) -- OpenWRT virtual router (Double NAT) with NAT66 --LAN/wireless clients get IPv6 access
 
Last edited:

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
Interesting viewpoint from chief scientist of APNIC. I think APNIC is perceived as pushing for IPv6. But the chief scientist seems to say IPv6 is not necessary either since NAT just works fine for many network operators and nations. But for nations like India and Malaysia and China, IPv6 is probably the way to go.

https://www.theregister.com/2024/09/16/240_4_ipv4_block_unnecessary/

Mon 16 Sep 2024 // 06:32 UTC
The 240/4 block of IPv4 addresses – the six percent of the available IPv4 space that is currently not available for public use – should be left alone rather than being added to the pool of available internet resources, according to Geoff Huston, chief scientist of the Asia Pacific Network Internet Center.

...

"Two thirds of the internet is not on IPv6 and is thriving on network address translation (NAT)," he argued.
...

Huston also expressed concern that freeing the block for public use would mean myriad new routes become available – meaning roughly 7,000 network operators around the world would get the job of assessing their safety. Again, he feels there is little incentive to do so, given the internet and IPv4 number space as currently constituted work well, and that IPv6 is a more-than-adequate replacement for those who can't rely on IPv4. Huston pointed to India's widespread adoption of IPv6 – after the nation missed out on an IPv4 allocation commensurate to its population – as an example of a scenario in which NAT just can't do the job, and IPv6 is therefore necessary.

Huston therefore argues that for many network operators and nations, moving from IPv4 to IPv6 is also not necessary. Extensive use of IPv4 NAT, he argues, will allow connection of as many devices as is possible if each were given a unique IPv6 address.
 
Last edited:

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
So if we read the above comments, maybe it is not bad for Singapore's lagging behind in IPv6 adoption. IPv4 and NAT just work for most of us, for the ISPs and for Singapore as a nation. That is probably why IMDA does not really push for IPv6.
 

firesong

Supremacy Member
Deluxe Member
Joined
Jan 17, 2001
Messages
8,629
Reaction score
4,638
So if we read the above comments, maybe it is not bad for Singapore's lagging behind in IPv6 adoption. IPv4 and NAT just work for most of us, for the ISPs and for Singapore as a nation. That is probably why IMDA does not really push for IPv6.
I don't think The Register is quoting him in full. Might be good to obtain the full transcript to get a better sense of what was actually said and in what context.

The fact that he's pushing back against releasing the 240/4 block is yet another clear sign that native IPv4 addresses have run out. This is the core reason behind why people are asking for release of that block of addresses.

Just using a green lens, the NAT performance penalty is real, and is something worth noting here - we have a lot of overpowered internetworking hardware that are wasted energy because of the poor network performance due to NAT. Added up globally, that's a lot of wasted energy. In terms of inferior network performance, a lot of resources are expended trying to maintain status quo from sections and people who refuse to move because they are trying to protect their setups (and rice bowls), when perhaps it makes more sense to evolve and bite the bullet properly.

At current expansion rates, one can make the case that IPv6 will also run out of addresses hypothetically within the next decade. Even having 2^128 or 340 undecillion addresses may not be enough for a global network if we truly look at the smart home, with every lightbulb, door/gate lock and other devices each being assigned one. Especially once you remove the reserved blocks. Trying to hold on to IPv4 with only 4 billion addresses, with a bulk of it reserved and not for public use, is quite impossible.

I also dislike the argument that IPv6 is "only for poor countries" or "only for large countries", cos it makes the case that rich countries overpurchased IPv4 addresses and implies that we cut them out. That's just feeding public ignorance but there's an element of truth to it.
 

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
I don't think The Register is quoting him in full. Might be good to obtain the full transcript to get a better sense of what was actually said and in what context.

Indeed I also can not find the portion of The Register report which mentions "Huston therefore argues that for many network operators and nations, moving from IPv4 to IPv6 is also not necessary. Extensive use of IPv4 NAT, he argues, will allow connection of as many devices as is possible if each were given a unique IPv6 address."

The blog post of Geoff Huston is here.
https://blog.apnic.net/2024/09/10/looking-for-240-4-addresses/

Or here.
https://www.potaroo.net/ispcol/2024-09/2404.html

Both posts have no mentioning of NAT being the solution.
 

ddgp1105

Member
Joined
Oct 2, 2007
Messages
125
Reaction score
27
I didn’t request IPv6, but today I found an IPv6 address in my router.
Singtel started IPv6 liao?
rqnLZzL.jpeg


test the ipv6,i think it's work

qwnTInM.jpeg


H4WYnba.jpeg
 

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
I didn’t request IPv6, but today I found an IPv6 address in my router.
Singtel started IPv6 liao?
rqnLZzL.jpeg


test the ipv6,i think it's work

qwnTInM.jpeg


H4WYnba.jpeg

Nice.

Singtel started new native IPv6 deployment since Nov 2023 (/56 IPv6). I accidently found that with a new default OpenWRT installation.

But in reality, only ONT users will get the benefits easily when their areas got the deployment. For Singtel ONR users, you have to requst through Singtel and most ONR users may not know and may not care anyway.

So in the end, the IPv6 adoption rate among Singtel Fibre Internet users is still low.

For Singtel mobile users, 5G SA does give public routable IPv6 address.
Singtel 5G NSA --> sometimes have IPv6 sometimes not.
Singtel 4G --> no IPv6.
 

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
Previously I have tested NAT66 function to be working under OpenWRT, but not pfSense/OPNsense or Asus.

Recently ZTE has updated the FW version to V1.0.0.4B8.8000 which supports NAT66 function.

1) Tested NAT66 function of ZTE BE7200 Pro+ under router mode and it works fine.

Singtel ZTE F8648P ONR --> 2.5G/10G switch -- ZTE BE7200 Pro+ in router mode (Double NAT)

ZTE BE7200 Pro+ get /128 IPv6 address from the Singtel ONR and then wireless client gets non public routable IPv6 addresess, but then IPv6 test whos that wireless client can pass IPv6 test and the address shown is the /128 IPv6 address.

2) One more test with Double NAT 66 and it works as well.

Singtel ZTE F8648P ONR --> 2.5G/10G switch -- OpenWRT router (Double NAT with NAT66) -- ZTE BE7200 Pro+ in router mode (Trippe NAT, with NAT66)
 

joeltng

Member
Joined
Dec 29, 2010
Messages
369
Reaction score
25
Wanted to ask if anyone here on Starhub having issues with IPV6? Only realised this when my parents began complaining about some apps not loading. went to https://test-ipv6.com/ and indeed ipv6 doesn;t seem to be working. tried pinging google ipv6 dns and cloud flare one all cannot be pinged. Wierd thing is cloud flare speed test works and can detect ipv6 :/
 

Mach3.2

Great Supremacy Member
Joined
Apr 8, 2011
Messages
72,405
Reaction score
2,465
Wanted to ask if anyone here on Starhub having issues with IPV6? Only realised this when my parents began complaining about some apps not loading. went to https://test-ipv6.com/ and indeed ipv6 doesn;t seem to be working. tried pinging google ipv6 dns and cloud flare one all cannot be pinged. Wierd thing is cloud flare speed test works and can detect ipv6 :/
yes, not just you.

https://forums.hardwarezone.com.sg/...cp6-gateway-not-working-with-starhub.7066471/
 

sacredrays

Arch-Supremacy Member
Joined
Feb 1, 2009
Messages
12,794
Reaction score
1,932
singtel fibre ipv6 is select dual stack or 6rd better? first option better right?

lol, wtf i kena logged out of forum when i go enable ipv6
 

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
singtel fibre ipv6 is select dual stack or 6rd better? first option better right?

lol, wtf i kena logged out of forum when i go enable ipv6

Dual Stack better, if it is working for you.

As per the report, ONR users need to request through Singtel to get native IPv6. I got it working with my ZTE F8648P ONR (Singtel 5Gbps plan).
 

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
Last edited:

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
Some IPv4 vs IPv6 ping results for reference from my Singtel 5Gbps connection (wired Linux client to the Singtel ZTE F8648P ONR) in the following thread (page 127 and 128).
https://forums.hardwarezone.com.sg/threads/whizcomms-new-fibre-broadband-provider.5494964/page-127

One interesting result is that IPv6 connection to Cloudflare Family DNS is better than IPv4 at this paticular time.
Bash:
mcuee@miniroute10g:~$ ping -c 4 family.cloudflare-dns.com
PING family.cloudflare-dns.com (2606:4700:4700::1003) 56 data bytes
64 bytes from family.cloudflare-dns.com (2606:4700:4700::1003): icmp_seq=1 ttl=50 time=2.72 ms
64 bytes from family.cloudflare-dns.com (2606:4700:4700::1003): icmp_seq=2 ttl=50 time=3.02 ms
64 bytes from family.cloudflare-dns.com (2606:4700:4700::1003): icmp_seq=3 ttl=50 time=2.81 ms
64 bytes from family.cloudflare-dns.com (2606:4700:4700::1003): icmp_seq=4 ttl=50 time=2.62 ms

--- family.cloudflare-dns.com ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3003ms
rtt min/avg/max/mdev = 2.620/2.792/3.024/0.149 ms

mcuee@miniroute10g:~$ ping -4 -c 4 family.cloudflare-dns.com
PING family.cloudflare-dns.com (1.0.0.3) 56(84) bytes of data.
64 bytes from family.cloudflare-dns.com (1.0.0.3): icmp_seq=1 ttl=49 time=34.9 ms
64 bytes from family.cloudflare-dns.com (1.0.0.3): icmp_seq=2 ttl=49 time=35.0 ms
64 bytes from family.cloudflare-dns.com (1.0.0.3): icmp_seq=3 ttl=49 time=35.1 ms
64 bytes from family.cloudflare-dns.com (1.0.0.3): icmp_seq=4 ttl=49 time=35.3 ms

--- family.cloudflare-dns.com ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3004ms
rtt min/avg/max/mdev = 34.859/35.074/35.316/0.165 ms
 

joeltng

Member
Joined
Dec 29, 2010
Messages
369
Reaction score
25
Trying to see whats up with starhub IPV6 for affected people, wanted to ask you guys as ipV6 is really new/foreign to me.

My WAN IPv6 Address shows as 2406:3003:1000::xxxx so I would expect the gateway IP WAN to be 2406:3003:1000:xxxx::1/64 correct? Cause currently its 2406:3003:2000:xxxx::1/64 and conversely all my client devices also start off with 2406:3003:2000:xxxx::xxxx. if so it might be some kind of misconfiguration going on.
 

liangtam

High Supremacy Member
Joined
Aug 20, 2002
Messages
38,841
Reaction score
181
Properly they should and will assign you 2 block of IPv6 address block. Your device will act as the gateway on the LAN side and an actual proper v6 router doing routing.
You need to discard the v4 NAT/PAT behaviour/concept that is prevalent in the old setup
 

xiaofan

High Supremacy Member
Joined
Sep 16, 2018
Messages
31,562
Reaction score
8,956
Just FYI.

My Singtel ZTE F8648P ONR with Singtel native IPv6 enabled.

WAN IPv4 address: 115.66.77.x
WAN IPv4 Gateway: 115.66.79.254 (Singtel side)
LAN IPv4 address of the ONR: 192.168.1.254 (this is the gateway for LAN clients).

WAN IPv6 address: 2400:d802:d10::xxxx
LAN IPv6 addresses: 2400:d802:dd3:xxxx::/64

BTW, the gateway for the IPv6 LAN clients is not 2400:d802:dd3:xxxx::1 in my case. It is 2400:d802:dd3:xxxx:xxxx:xxxx:xxxx.xxxx.

The next hop goes to Singtel side 2400:d801:4001:611:: (from mtr).
 
Last edited:

astones153

Member
Joined
Jun 8, 2021
Messages
159
Reaction score
82
Trying to see whats up with starhub IPV6 for affected people, wanted to ask you guys as ipV6 is really new/foreign to me.

My WAN IPv6 Address shows as 2406:3003:1000::xxxx so I would expect the gateway IP WAN to be 2406:3003:1000:xxxx::1/64 correct? Cause currently its 2406:3003:2000:xxxx::1/64 and conversely all my client devices also start off with 2406:3003:2000:xxxx::xxxx. if so it might be some kind of misconfiguration going on.
Normally, Starhub's IPv6 gateway should be on an fe80 address, they use ULAs for some reason. Your WAN and LAN are behaving perfectly normal, it's on Starhub's end that their gateway is not properly advertising itself to end users.
 
Important Forum Advisory Note
This forum is moderated by volunteer moderators who will react only to members' feedback on posts. Moderators are not employees or representatives of HWZ. Forum members and moderators are responsible for their own posts.

Please refer to our Community Guidelines and Standards, Terms of Service and Member T&Cs for more information.
Top