#nat

2025-06-15

Kafka через NAT

Иногда появляется необходимость обращения к брокерам Kafka из другой сети через NAT. Но, даже если NAT настроен корректно, то при попытках подключения, обратно возвращается ошибка о том, что брокер недоступен. Хотя ping и telnet по нужному порту проходят. В статье мы подробно разберем, как необходимо настроить Kafka для того, чтобы к системе можно было обращаться через NAT.

habr.com/ru/articles/918520/

#kafka #nat

Hubu.dehubude
2025-06-13

⚡ Gericht: Trumps Einsatz der Nationalgarde in Kalifornien illegal: Ein Bezirksgericht in San Francisco hat die Entsendung der Nationalgarde durch die US-Regierung unter Präsident Donald Trump nach Kali... hubu.de/?p=282515 |

@lycanmatriarch That does sound weird. The LAN-side-of-NAT IP (typically RFC 1918 space) should only show up on the LAN side of NAT. That's kind of the point of NAT.

Are you *sure* it's your Plex box originating that traffic, and not just some other machine on the network you're on that happens to have the same IP address? Maybe something in the Plex web UI which happens to include a request to its internal IP address and which (obviously) does not get rewritten by the NAT at your home?

#NAT

Rynn the Cyberwitch 🌙lycanmatriarch@furry.engineer
2025-06-12

Hey other #network folks I have a weird #firewall and #NAT question. I'm looking at the firewall traffic log where my work laptop (I'm in the office) is the source, and I'm looking at the traffic connecting to my home plex server.

The plex server is a private IP behind NAT. In the log I show the putbound traffic from my work laptop to my home router IP, but I also see a following entry with the private IP that my router is forwarding plex's port traffic to. I thought anything done behind NAT wouldn't show up, especially because I'm only looking at the initial communication out, and not the return communication from my server to my me. Is my understanding of NAT wrong?

2025-06-09

What Docker did was to intellectually limit the creativity that users could have had with containers, funneling everyone into the most trivial of network use-cases. When everything is server-client, it's really hard to develop peer-to-peer or avant guard applications.

The parallels to #LegacyIP and #NAT are staggering, as they too stymied progress in other, not-yet imagined scenarios. #IPv6 should have been the go-to for hyperscalar from day-one. That it wasn't is forever a travesty.

7/n

Ponder Stibbons 🇧🇷🇩🇪blackcoffeerider@social.saarland
2025-05-27

I think i just made the first timid steps on understanding some core concepts behind #ipv6 that i always struggled to understand.
I just started to think about it like a more sane version of #IPX than a totally insane version of #ipv4 and suddenly things start to make sense...

Of course that doesn't help anyone who never used IPX, but it is a great help for me. To be honest I am still scared of IPV6 because #NAT feels like a security feature if all your FW-Rules look at ip headers anyway.

Mike Hindlemikehindleuk
2025-05-22

Hey Gemini:

Create a black and white photo of a misty morning in woodlands filled with old gnarly oak trees.

A black and white photo of a misty morning in woodlands filled with old gnarly oak trees.
2025-05-21

NAT — как наследие старого интернета мешает будущему

Поставили новый роутер, запустили онлайн-игру или развернули облачный сервер — и снова натыкаетесь на «двойной NAT», бесконечный порт-форвардинг и вместо своего IP видите чей-то 203.0.113.45. Причем железо и провайдеры уже готовы к IPv6, а мы все еще буксуем в прошлом. Давайте посмотрим, почему наследие старого интернета — повсеместный NAT — тормозит нашу сетевую эволюцию и что с этим можно сделать. Детали под катом.

habr.com/ru/companies/selectel

#selectel #ip #https #ipv4 #ipv6 #nat

2025-05-21

NAT — как наследие старого интернета мешает будущему

Поставили новый роутер, запустили онлайн-игру или развернули

habr.com/ru/companies/selectel

#selectel #ip #https #ipv4 #ipv6 #nat

2025-05-16

I had a thought earlier: is stateful #NAT guaranteed to never scale as well as stateless routing? My thinking is yes, since stateful NAT creates entries for each connection (a pair of hosts), whereas routing tables have entries per host destination. For N hosts, the worst case routing table would have N entries, but worse case connections would be N(N-1), which is basically N^2.

Now I wonder if there are distributed state #CGNAT implementations. Though that might be harder than moving to #IPv6

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst