Hacker Newsnew | past | comments | ask | show | jobs | submit | pjf's commentslogin

NB: this is not "IPv6 traffic crosses the 50% mark" but "availability of IPv6 connectivity among Google users", which is a very important difference. This means roughly half of Google users have IPv6 capability, which does not 1:1 correspond how much traffic is actually transferred over IPv6, which is what this submission says in the title.

Yeah and this distinction explains the fact that because China's Great Firewall blocks Google, this website shows 4.66% adoption as a reflection of that. I think China's IPv6 support rate is actually much higher than that, maybe a little over 50% because of its central initiative to increase IPv6 adoption?

EDIT: Apparently it's 77% https://pulse.internetsociety.org/en/news/2026/01/china-hits...


Wouldn’t it be close? AFAIK modern network libraries on modern OSs default to IPv6 when available.

It also means you're excluding China, who has has it as a long-term priority to deploy IPv6 and have made huge strides.

great post! key points for me:

1. 100 IXes alone would get 56% IPv4 and 61% IPv6 prefixes, but ~14% reachability

2. little uniqueness between exchanges: not many new prefixes after the top 5

3. for outbound-heavy networks IXes are great, but to attract traffic they are not (edit: applies to automatic peering via route servers)


Post author here

For inbound traffic, they're completely fine. This is only looking at the route servers. You can almost certainly receive 50/50 traffic ratios if you do bilateral peering. This post only covers the " automatic peering " services that IXs offer


Kind of "funny" affected service is BGP RouteViews CLI access, still running over telnet: https://archive.routeviews.org/ (scroll to bottom of the page)

Isn't this one of the remaining, "legit" uses of the Telnet protocol on TCP/23 port over the public Internet?


I'd advise to first compare with:

- https://bgp.tools/ - https://bgproutes.io/ - https://bgp.he.net/ - https://radar.qrator.net/ - https://github.com/nttgin/BGPalerter

...all of which are (usually) free. IMHO you should have a competing product + money strategy before you continue. Many people have tried (and failed) to make money off BGP.

BTW, author of https://bgpipe.org/ here, an open-source BGP data tool


Awesome - exactly the feedback I was looking for. Thanks a lot!


Just out of curiosity, why did you use Go for bgpipe and Python?


One reason is there already was exabgp, written in Python, which in my experience is slow and resource hungry. Golang is much faster, easily portable, and produces static binaries (easy to deploy).

Another thing is bgpipe speaks JSON to background (or even remote) packet processors, so basically you can use whatever language you want with it to drive your BGP routers.


https://bgpipe.org/

I'm working on a man in the middle proxy for BGP, which can fix and inspect routing sessions on the fly. Like a firewall for the BGP control plane.


Some time ago I was building a mitm proxy myself, then I found out about: https://www.mitmproxy.org/ Maybe you already had it in your radar


That's for HTTP/S and related, as parent said his is for BGP which is a completely different protocol


...prepare your pagers for an April 1st "experiment"


Do you know - and can share - how this compares vs the Onyx Boox Tab Ultra C Pro?

I guess it might be a very competitive alternative to your product, yet it can't reach 60+fps and is sketchy in terms of security, imho.



What are the limits on the number of databases (say sqlite files) one can have attached (and queried through a UNION) at a single time?


Small world : - ) Great work, Larry!


:laugh: Indeed! Thanks!


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: