#routing

10 posts · Last used Aug 24

I really don’t need any more peers for my AS201379! (Okay, maybe a few more...) Started back in December 2025, and now this "little" infrastructure is running: 14 individual eBGP sessions 3 Internet Exchanges Multiple transit providers 110+ direct peers All 100% IPv6 (2a06:9801:1c::/48) because legacy IP belongs in the last century. Powered entirely by FreeBSD :freebsd: 4 routers on 15.1-RELEASE running FRR and PF. #BGP #IPv6 #FreeBSD #Routing #Networking #sysadmin
25
4
10
0
Looks like Flatpak discovery will move from one mostly generally accepted hub to a gazillion of repos since the performative LLM complaints have befallen the possible single hub. (Partially half-understandable arguments, partially performative fluff for my taste) Which is ok as it triggers competition that might lead to better outcomes. (bazaar vs cathedral and all, rite? 😏 ) And removes a possible single point of failure. Now what I'd like to see is an approach that does discovery of those repos and enables search across those repos without a single point of failure. As e.g. a single, centrally hosted index for repos to register to would be. Flatpak already uses a reverse-dns-namespace structure internally. Anything that would leverage that? Wouldn't help with enumeration though. Something something DHT/Gossip/...? (as the cool thing from the past) Clearly a problem that will arise with more of the expectable decentralized systems that need uniform discovery and search. What's the Fedis wisdom on current academic approaches that might be interesting around this? Or bastardized routing tech, yet another idea on how to twist DNS into a pretzel, etc? Dare I say: Blockchain (as the unpopular thing before the LLMs 🤣 ) Throw you best Arxiv posts etc. at me. I'm just curious about the state-of-the-art techn for now, all smirk and sarcasm aside. #discovery #search #decentralized #routing #dns #fedimeta #flathub #dht #cooltechfromthepast #llm #selfhosting #fossdrama #spof #singlepointoffailure
3
0
2
0
Since this year I have massive trouble accessing anything hosted on AWS during the evening hours like these. My ISP is Deutsche Telekom AG (AS3320) and it looks like the only route to AWS (AS16509) leads through Cogent (AS174). MTR is showing more than 30% packet loss through Cogent to an AWS target. I know peering between DTAG and Cogent has been historically bad, but I never noticed issues until now. What happened? Do any BGP nerds know more? #BGP #Routing #Internet #Backbone #AskFedi
1
1
5
0
You've seen all posts