High Ping / Broken International Gateway Routing to AWS Sydney (Battlefield 6)

Forums › 2degrees (including Slingshot, Orcon, Flip, Stuff Fibre, MyRepublic, 2talk and Vocus) › High Ping / Broken International Gateway Routing to AWS Sydney (Battlefield 6)

1 | 2

geek3001

304 posts

Ultimate Geek
+1 received by user: 443

ID Verified

Subscriber

#3499765 3-Jun-2026 09:36

mysticalxo:

Packets: Sent = 33, Received = 33, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 2ms, Maximum = 129ms, Average = 37ms

Wifi only unfortunately, in the past I ran a cat7 cable but don't have that option at new apartment. I am running 6G band only via my network adapters settings which is a Archer TBE400E V1 with latest drivers. These issues have not been experienced before and are only recent within the past 3 days.

Typical download speeds were 900mbps on steam before this weekend, now they are 250approx.

Can you connect by Ethernet cable to the router, even if only temporarily, assuming it is elsewhere in your own apartment? In this case, if the problem goes away, then you need to stop using WiFi and move your computer closer to the router and connect to it by cable.

Or are you just connecting to shared WiFi infrastructure that is provided within the building and have no physical Ethernet cabling anywhere? In this case you may be stuck with what you have and be unable to remedy the performance issue.

mysticalxo

11 posts

Geek

#3499939 3-Jun-2026 16:33

Hi aspired, Geek, Others, I completely understand why you see those earlier Wi-Fi logs and point to that first. A 37ms average and 129ms spike to the local gateway definitely isn't ideal. Unfortunately, running a physical Ethernet cable isn't an option due to the physical layout of this apartment. However, I managed to capture two completely clean traceroutes where my local 6GHz Wi-Fi connection behaved perfectly. As you can see below, Hop 1 to my HB810 router was rock solid at a flat 2ms across the board. The cross-tests I've run over the last X amount of hours show that the Wi-Fi is a symptom rather than the cause, and strongly indicate a CG-NAT layer bottleneck:
The Mobile Hotspot Test: When I connect this exact same PC via Wi-Fi to a mobile hotspot on a different cellular network, the baseline ping to the AWS servers drops straight back down to a stable ~55ms. If my PC's Wi-Fi card or local environment was causing the 130ms bottleneck, the hotspot loop would lag exactly the same way. It doesn't.

The ~250 Mbps Steam Cap: My Steam downloads suddenly hitting a hard, flat ceiling at around 250 Mbps on a Hyperfibre 2 plan (I previously had Hyperfibre 4 and forgot I downgraded when I moved) happened at around the same time the ping spiked over the weekend. Local 6GHz Wi-Fi easily previously pushed 900+ Mbps over wifi, so this hard cap strongly points to per-user packet queue throttling further up the line or some other issue not relating to a wifi driven connection.

Because my Hop 1 latency remains completely flat at 2ms on these fresh tests while the internal core nodes are violently spiking up to 147ms on the exact same rows, it completely isolates the heavy jitter to the Vocus core network routing layer. The local router spikes seem to be the hardware actively struggling to process heavily queued or jumbled data coming off that specific CG-NAT pool. Here are the fresh logs demonstrating both the internal core network bottleneck and the international routing drop-offs: Traceroute to AWS New Zealand (3.102.141.118): -------------------------

Tracing route to ec2-3-102-141-118.ap-southeast-6.compute.amazonaws.com
over a maximum of 30 hops:

1 7 ms 2 ms 2 ms HB810
2 86 ms 3 ms 3 ms v1.cpcch-ric-bng1.tranzpeer.net
3 25 ms 24 ms 24 ms 192.168.255.237
4 24 ms 147 ms 24 ms 192.168.255.234
5 120 ms 24 ms 24 ms default-rdns.vocus.co.nz
6 * * * Request timed out.
7 * * * Request timed out.
8 * * * Request timed out.
9 * * * Request timed out.
10 24 ms 24 ms 142 ms ec2-3-102-141-118.ap-southeast-6.compute.amazonaws.com
----------------------
Traceroute to AWS Sydney (52.64.156.81):

-------------------------

Tracing route to ec2-52-64-156-81.ap-southeast-2.compute.amazonaws.com
over a maximum of 30 hops:

1 2 ms 2 ms 2 ms HB810
2 15 ms 131 ms 3 ms v1.cpcch-ric-bng1.tranzpeer.net
3 24 ms 23 ms 24 ms 192.168.255.237
4 80 ms 23 ms 23 ms 192.168.255.234
5 106 ms 24 ms 24 ms default-rdns.vocus.co.nz
6 * * * Request timed out.
7 * * * Request timed out.
[Truncated - Hops 8 to 30 all timed out]

------------------------- Right after Hop 5, the international trans-Tasman link completely black-holes again, timing out for the remaining 25 hops and failing to reach Sydney entirely. This confirms the routing path remains highly unstable and broken at the edge gateway level. Support Update:
I called 2degrees again after receiving a text from them to get in touch. The ticket is still actively being processed by their advanced teams, and they are currently attempting to apply a Static IP profile to my account as a troubleshooting measure to see if opting me completely out of the CG-NAT layer clears the local queuing and international bottlenecks.

RunningMan

9329 posts

Uber Geek
+1 received by user: 4968

#3499948 3-Jun-2026 17:11

mysticalxo: [snip] However, I managed to capture two completely clean traceroutes where my local 6GHz Wi-Fi connection behaved perfectly. As you can see below, Hop 1 to my HB810 router was rock solid at a flat 2ms across the board. The cross-tests I've run over the last X amount of hours show that the Wi-Fi is a symptom rather than the cause, and strongly indicate a CG-NAT layer bottleneck:
The Mobile Hotspot Test: When I connect this exact same PC via Wi-Fi to a mobile hotspot on a different cellular network, the baseline ping to the AWS servers drops straight back down to a stable ~55ms. If my PC's Wi-Fi card or local environment was causing the 130ms bottleneck, the hotspot loop would lag exactly the same way.

There's some incorrect assumptions about wifi in the logic there.

1) Just because you can get a stable connection to the AP at one point in time doesn't mean it is that way all the time. Intermittent interference from something nearby on a similar frequency can cause the very issues you see.

2) The test over wifi to a hotspot is unlikely to be be a good comparison. You've mentioned more than once that you're using 6GHz wifi, and most phone hotspots would use 5GHz at best - something interfering with the 6GHz band may not affect 5GHz.

That's not to say that your wifi is the definitive fault, but for the sake of grabbing a cable temporarily it's a fairly quick thing to isolate.

mysticalxo

11 posts

Geek

#3499976 3-Jun-2026 20:04

To test your theory, I moved my PC directly next to the router. I also dug out a cable to try and hardwire into the router's high-speed LAN ports. Unfortunately, the PC and router completely refused to negotiate a local IP handshake (stuck on 'Unidentified Network / No Internet' via DHCP), so a wired test is physically off the table, I sat in the middle of the hallway with my desktop and a monitor, peripherals, used both the Cat5e cable provided with the router, and my own Cat7 Ethernet cable to no avail. I disabled my archer so there could be no potential interference, I then ran;

netsh winsock reset
netsh int ip reset
ipconfig /flushdns
ipconfig /release
ipconfig /renew

Still no connection.

I then did another Wireless ping test with the Archer. I don't have a current screenshot, but the continuous ping results right next to the router on the 6GHz band were... telling. I am still averaging 40ms and spiking violently up to 350ms directly to the local gateway (192.168.1.1).

On an isolated 6GHz AND 5Ghz band with zero neighbor contention, sitting literal inches from the router and ont, a 350ms local spike... Maybe textbook symptom of Router CPU Max-Out / Bufferbloat?

The router's processor is either clearly choking and hitting 100% utilization trying to sort, unpack, and handle the heavily queued, throttled data coming off the potentially broken/queued/congested 2degrees CG-NAT network layer. This also explains why my Steam downloads are hitting a hard, flat, artificial ceiling at approx 250 Mbps on a Hyperfibre 2 connection.

So while you're right that the Wi-Fi link is currently a mess and can be unreliable, it's seemingly a secondary symptom of the main upstream infrastructure failure hammering the router's brain. This also can explain me going from completely fine on the same hardware, drivers, network etc for the past 4 months with no issues.

I also ran the typical Aginet internet test through the app and I am getting 900 down, 900-1200 up over multiple tests (In the past I was getting flat 2gb up 2gb down). But my pc isn't getting the full 900 despite being the only one on the network currently? I am not a techy, just a script kiddie who dropped out of CompSci so bare with me.

Just pray that 2degrees figure it out?

MaxineN

Max
2111 posts

Uber Geek
+1 received by user: 1733

ID Verified

Trusted

Subscriber

#3499987 3-Jun-2026 20:50

I think our configuration is a bit fubar'd and we need to start from the basics.

Let's forgo achieving our HF speeds and get a router in there that has been fully factory resetted configured for VLAN 10 DHCP and then hooking up our desktop to that. Please ensure there are no static IP configurations on your desktop.
If your latency issue is resolved, great. Now reset your HB810 and continue to add stuff in one by one until you have found your problem.

If by some unfortunate circumstance that your HB810(even wired) is still a problem then considering it is an ISP supplied product the ISP should be supporting you in getting it right.
We'll figure it out pretty quickly if it's 2degrees CG-NAT, or your equipment. Your ISP will also be thankful that you've done this.

I know this is a massive bother and I know you have better things to do, but we need to get the basics right before we can return to normality.

Ramblings from a mysterious lady who's into tech. Warning I may often create zingers.

Want broadband cheap and made for tech enthusiasts? Go look over at Quic. Use R177510EBNVXP for free setup at check out.

mysticalxo

11 posts

Geek

#3500008 3-Jun-2026 21:22

I appreciate the back-to-basics logic Max, and normally I'd agree that a full factory reset is the best way to clear the slate, I WOULD do a full factory reset, however whilst I wait for 2degrees to get back to me I don't think touching it would help my case. However, a few factors make that test impossible on my end right now:

1. The LAN Port Lockup: As mentioned in my hallway test, the HB810's LAN ports are completely refusing to negotiate a local IP handshake via DHCP over multiple cables on both 10gbps port, and the 2.5gbps prots (stuck on 'Unidentified Network'). I physically cannot run a wired baseline test to any device because the router's internal local switch is locked up or there is something with my Z690 Mobo.

2. Hardware Restrictions: I don't have a spare, high-performance router lying around that can handle the specific VLAN 10 provisioning for a Hyperfibre 2 line. I don't have any at all lying around unfortunately. Getting chorus here on time initially was a month late and was a pain as it was, even with a 3 month notice for installation.

3. The Hotspot Control Test: If the issue were a hidden static IP config or a corrupted setting on my desktop, the PC would experience the same throughput choke and jitter on any network. But when I tether to a mobile hotspot on a different cellular network via Wi-Fi, the latency drops straight back to a stable ~55ms with zero packet grouping (This applies to both 6ghz and 5Ghz bands which has been tested.)

The fact that the router is pulling 900+ Mbps internally via the Aginet app, but a CLI speedtest ran through terminal shows the actual data stream hitting my client is choked down to an absolute crawl at 87 Mbps, with a horrifying 2134ms (over 2 seconds) of upload latency under load, strongly indicates the router's processor is (To my belief) completely maxed out trying to handle a corrupted/queued incoming profile stream

Here is one of them that I copied.

"

Speedtest by Ookla

Server: 2degrees - Christchurch (id: 7317)
ISP: 2degrees
Idle Latency: 37.53 ms (jitter: 45.52ms, low: 37.38ms, high: 128.41ms)
Download: 87.12 Mbps (data used: 111.4 MB)
146.61 ms (jitter: 68.07ms, low: 37.68ms, high: 666.04ms)
Upload: 271.81 Mbps (data used: 456.8 MB)
1105.47 ms (jitter: 108.98ms, low: 36.50ms, high: 2134.44ms)
Packet Loss: 0.0%

"

And unfortunately It is currently a waiting game with 2degrees right now.

Geekzone

Want to support Geekzone and browse the site without the ads? Subscribe to Geekzone now (monthly, annual and lifetime options).

MaxineN

Max
2111 posts

Uber Geek
+1 received by user: 1733

ID Verified

Trusted

Subscriber

#3500011 3-Jun-2026 21:25

And that is exactly why we need to go back to basics.

You don't need a spare high performance router right now. You just need something to do some sanity checking and to ensure it's absolutely not you.
Right now it looks like the issue is at your end, specially since you've mentioned you're having issues with the HB810.

Ramblings from a mysterious lady who's into tech. Warning I may often create zingers.

Want broadband cheap and made for tech enthusiasts? Go look over at Quic. Use R177510EBNVXP for free setup at check out.

michaelmurfy

meow
13665 posts

Uber Geek
+1 received by user: 11071

Moderator

ID Verified

Trusted

Lifetime subscriber

#3500014 3-Jun-2026 21:35

It looks like you’re using AI to diagnose this and while it can be helpful, it’s being misleading right now.

I seriously don’t think there is a fault with 2degrees. It’s on your end somewhere. Go back to the basics of replacing Ethernet cables and factory resetting your router. If you’re using WiFi then understand this is causing you further issues, you need to use Ethernet if latency is important.

The latency to “their internal core nodes” seriously don’t matter as routers deprioritise pings. The thing that matters is the end destination latency which is looking fine.

And please stop posting AI responses on here. It’s not helping anyone.

Michael Murphy | https://murfy.nz
Referral Links: Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.

mysticalxo

11 posts

Geek

#3500015 3-Jun-2026 21:35

Done a factory reset.

Speedtest by Ookla

Server: 2degrees - Christchurch (id: 7317)
ISP: 2degrees
Idle Latency: 8.84 ms (jitter: 0.65ms, low: 8.50ms, high: 9.58ms)
Download: 125.26 Mbps (data used: 218.6 MB)
88.65 ms (jitter: 70.93ms, low: 9.50ms, high: 663.20ms)
Upload: 286.11 Mbps (data used: 403.6 MB)
491.85 ms (jitter: 105.57ms, low: 8.91ms, high: 1183.17ms)
Packet Loss: 0.0%
Result URL: https://www.speedtest.net/result/c/cb0e51fc-71b8-4a3a-8177-72112a268ed9
PS C:\Users\riley\Downloads\ookla-speedtest-1.2.0-win64>

However Aginet is now 2k Up, 2k Down again, Progress. What now?

mysticalxo

11 posts

Geek

#3500017 3-Jun-2026 21:38

And here is the 6G only band wifi. (Previous was MLO)

Speedtest by Ookla

Server: 2degrees - Christchurch (id: 7317)
ISP: 2degrees
Idle Latency: 8.54 ms (jitter: 0.38ms, low: 8.32ms, high: 9.63ms)
Download: 118.35 Mbps (data used: 109.6 MB)
117.66 ms (jitter: 75.86ms, low: 9.59ms, high: 694.99ms)
Upload: 169.68 Mbps (data used: 240.0 MB)
625.95 ms (jitter: 112.08ms, low: 8.52ms, high: 2168.81ms)
Packet Loss: 0.0%
Result URL: https://www.speedtest.net/result/c/cf2d2c69-ecf3-4ed5-864f-fb49d05f5af0

michaelmurfy

meow
13665 posts

Uber Geek
+1 received by user: 11071

Moderator

ID Verified

Trusted

Lifetime subscriber

#3500019 3-Jun-2026 21:39

We don’t care about WiFi. Only use Ethernet.

And based on those responses you’ve confirmed the issue is with your PC / Router and not 2degrees. Boot it into Linux and test under that.

Geekzone

Want to support Geekzone and browse the site without the ads? Subscribe to Geekzone now (monthly, annual and lifetime options).

quickymart

15117 posts

Uber Geek
+1 received by user: 14336

ID Verified

#3500020 3-Jun-2026 21:40

Those results are over wifi or ethernet?

edit, Michael beat me to it.

mysticalxo

11 posts

Geek

#3500027 3-Jun-2026 21:45

michaelmurfy:

It looks like you’re using AI to diagnose this and while it can be helpful, it’s being misleading right now.

I seriously don’t think there is a fault with 2degrees. It’s on your end somewhere. Go back to the basics of replacing Ethernet cables and factory resetting your router. If you’re using WiFi then understand this is causing you further issues, you need to use Ethernet if latency is important.

The latency to “their internal core nodes” seriously don’t matter as routers deprioritise pings. The thing that matters is the end destination latency which is looking fine.

And please stop posting AI responses on here. It’s not helping anyone.

Regardless of how I format my data into a post, the raw console data I'm pasting is 100% real and coming directly from my machine, I am not going to sit and type essay after essay when I can plug and play my results.

I literally just said in the previous post that I did the full factory reset on the HB810. I also already replied that the router's physical LAN ports are completely locked up and refusing to assign local IPs via DHCP to my desktop (which is why a wired test is physically impossible right now).

You say the end destination latency looks fine, but an internal hop spiking to 200ms+ followed by a final hop showing 2168ms (over 2 seconds) of upload bufferbloat (Which is disgusting for a hyperfibre connection REGARDLESS of wired vs wireless.) Which also, by the way, was ruled out considering I spent 4 months with the exact same setup, with NO issues. Under load 2kms is absolutely not fine for a 2Gbps Hyperfibre line. Not to mention the previous ping tests, the previous traceroutes, the current ticket with 2Degrees, amongst ALL other evidence.

More importantly, you're completely ignoring the main point: The Aginet app confirms the router itself is pulling 2k/2k at the WAN port, but a raw CLI test right next to the hardware on a flawless 6GHz band (0.38ms jitter) is hard-capped at 118 Mbps. A local wireless link doesn't create a clean, flat 90% bandwidth restriction ceiling like that—upstream packet shaping or a choked CG-NAT pool profile does.

I've done the basics, I've ruled out my environment, and the data is what it is. I'm leaving it with 2degrees Network Assurance to sort out the routing table via a Static IP as of now.

MaxineN

Max
2111 posts

Uber Geek
+1 received by user: 1733

ID Verified

Trusted

Subscriber

#3500029 3-Jun-2026 21:52

mysticalxo:

Done a factory reset.

Speedtest by Ookla

Server: 2degrees - Christchurch (id: 7317)
ISP: 2degrees
Idle Latency: 8.84 ms (jitter: 0.65ms, low: 8.50ms, high: 9.58ms)
Download: 125.26 Mbps (data used: 218.6 MB)
88.65 ms (jitter: 70.93ms, low: 9.50ms, high: 663.20ms)
Upload: 286.11 Mbps (data used: 403.6 MB)
491.85 ms (jitter: 105.57ms, low: 8.91ms, high: 1183.17ms)
Packet Loss: 0.0%
Result URL: https://www.speedtest.net/result/c/cb0e51fc-71b8-4a3a-8177-72112a268ed9
PS C:\Users\riley\Downloads\ookla-speedtest-1.2.0-win64>

However Aginet is now 2k Up, 2k Down again, Progress. What now?

Localized problem.

If you're seeing full speed from the HB810's inbuilt Ookla speedtest app which would actually pick the closest speedtest server. Then the problem is still your configuration.

I also do not believe there is anything wrong with the server either as I just smashed it from an off-net connection.

I understand you do not like hearing this, but you have to go back to basics.

If we don't want to do that then that is your choice and if 2degrees come back and say there is no issue then you will have to accept that and prove it to them that it's still them at fault and not your network.

Ramblings from a mysterious lady who's into tech. Warning I may often create zingers.

Want broadband cheap and made for tech enthusiasts? Go look over at Quic. Use R177510EBNVXP for free setup at check out.

michaelmurfy

meow
13665 posts

Uber Geek
+1 received by user: 11071

Moderator

ID Verified

Trusted

Lifetime subscriber

#3500033 3-Jun-2026 21:59

Right.

You literally replied to me telling you not to use AI with a response that is fully AI written that also made no sense seeing your previous response has proven the issue is on your end.

If you want to use AI to diagnose your issue then go ahead. But that issue is on your end, and those AI responses don’t belong on here. The correct information has been provided to you from members on here who have experience in this field.

1 | 2