IPV4 vs IPV6 Test results

Morning All

Any reason IPV6 shows slower speed than IPV4 ?

IPV4 Result

https://www.thinkbroadband.com/speedtest/16727395137...

IPV6 Result

https://www.thinkbroadband.com/speedtest/16727394547...

Thanks

Lots of possible reasons. IPv4 and IPv6 are two completely separate networks that happen to share some of the same wires and devices - otherwise they are ships that pass in the night.

* Transit and peering. There are networks which peer on IPv4 but not on IPv6, and vice versa. This leads to different Internet topology, i.e. your IPv4 traffic may take a completely different path across the Internet than IPv6 to the same destination.
* Related to this is peering politics. For example, Hurricane Electric peers widely on IPv6, in an attempt to turn themselves into a "tier 1" provider. Google does not announce its IPv6 routes to transit providers (in an attempt to force everyone to peer with them). Last time I checked, Cogent doesn't peer with Hurricane Electric or Google - so Cogent customers don't get a full IPv6 reachability.
* Router performance and optimisation. For example, Mikrotik routers are more optimised for IPv4 (with "fastpath" and "fasttrack") than they are for IPv6. The same could apply to the IPv4 and IPv6 stack on your client devices. It can also apply to hardware offloading on some devices.
* Packet processing. A router which does shaping or queue management for IPv4 traffic might not do it for IPv6 traffic. This could apply at the customer side, the ISP side, or both.
* IPv6 is inherently slower than IPv4 due to the larger IP headers - 40 bytes versus 20 bytes. However on a speedtest, this should only show as about 1.5-2% lower throughput.

Thanks

Nice comprehensive reply

Personally I would take the TBB speed test results with a pinch of salt.

This is my (IPV4) result on here: https://www.thinkbroadband.com/speedtest/16727575467...

This is a symmetrical gigabit connection.

If I run a test on speedtest.net (Going from my Hey!BB connection to the Hyperoptic test server), I get this: https://www.speedtest.net/result/14156293495

So, quite a disparity. Even using the speedtest.net windows client, I see a lot of variability just between local servers.

Maxing out a torrent is probably the best way to test download speeds at least.

In reply to a post by AndyPandy:

This is a symmetrical gigabit connection.

Just to check, that's domestic broadband, not a leased line?

Yeah, domestic broadband. HeyBroadband.

In reply to a post by AndyPandy:

Yeah, domestic broadband. HeyBroadband.

I would guess then some sort of congestion on the shared network between you and the various speed tests you've tried. I have had similar with my Virgin Media connection in the past. Hopefully Hey will add capacity.