Subj : FidoNews submission
To   : Alex Galiyev
From : Michiel van der Vlist
Date : Fri May 30 2025 06:01 pm

Hello Alex,

On Friday May 30 2025 10:25, you wrote to me:

MV>> Then Comcast and/or their users are doing something wrong. Native
MV>> IPv6 is definitely not slower than IPv4.

AG> It's definitely slower by design:

AG> - Bigger headers (40 B vs. 20 B) add per-packet overhead.

Hardly noticeable on modern hardware and well compensated by the packages themselves that can be larger.

AG> - Path MTU Discovery failures (blocked ICMPv6 "Too Big") lead to
AG> drops/retries.

Blocking ICMPV6 is a configuration error, not a design flaw.

AG>  - Transitional tunnels (e.g. 6in4, Teredo) introduce
AG> extra hops and encapsulation cost.

Agreed, but those tunnel mechanisms were only meant to facilitate the transition. Should have been fased out long ago. And mostly are. I enjoy native IPv6 from my ISP for about a decade now.

AG> - Less hardware offload and OS-stack tuning for IPv6 vs. decades of
AG> IPv4 optimizations.

IPv6 has been in use for well over a decade now. It has gone through the optimasation fase as well. Add to that the overhead in IPv4 like multiple levels of NAT and other tricks and work around to keep it working despite the IPv4 exhaustion.

If IPv6 is noticeably slower than IPv4 someone is doing something wrong.


Cheers, Michiel

--- GoldED+/W32-MSVC 1.1.5-b20170303
* Origin: http://www.vlist.org (2:280/5555)

You are viewing proxied material from cvs.synchro.net. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.