The Interpeer Project’s mission is to enable a human centric next generation internet.
Interpeer gUG is a tiny R&D non-profit run by @jens . See also: @interpeer
#Interpeer #HumanRights #NextGenerationInternet #DigitalCommons
---
This account is less likely to follow back than its owner; it's more of an outlet for official news and related interests.
| Website | https://interpeer.org/ |
| Code | https://codeberg.org/interpeer |
| Mailing List | https://lists.interpeer.io/mailman/listinfo/interpeer |
... trips you up reading them. There are no wrong comments in this. At worst, there may be comments I consider but won't adopt.
This isn't deeply technical effort. I would like these documents to be high level enough that they make some kind of sense to most people.
So it's effort, yes. A lot to ask. But it's not as if it requires arcane knowledge of the Internet's protocol stack.
Thank you, for anything you put into this.
... finalize them next, to clear a mental burden.
The first two are "ready", unless someone needs clarification. So I'd love a bit of feedback on them: here, a PR over on codeberg, or something on the discusssion list mailto:[email protected]
Otherwise I'll likely declare them final by end of March in this version. The last one I'm reworking more, though it's mostly reshuffling of bits to be clearer. A few additions.
It's the perfect time for telling me what...
... but that is going into hobby projects I've neglected. It's a "put on your own oxygen mask before helping others" kind of thing. Part of me feels bad about it, but I also really need it.
If you want to help out, the next steps are pretty simple. I've restructured a huge document I produced some years ago on the architecture into three somewhat more coherent and complete pieces:
https://specs.interpeer.org/draft-pies/
Starting with problem statement and gap analysis, and finally architecture, I want to...
I've been putting off posting stuff here for a while. The TL;DR is, I burned out last year.
This isn't about this project, it's about life and this project.
I'm good enough to realize it'll be a bit of time before I feel comfortable enough to put a lot of time in here. Meanwhile, ploddingly, I'm preparing a few more specs.
Some are just rewrites/restructuring based on things I learned about specs in the last few years. Some are new, drafts for the next things.
I am writing code again, too...
Kaufst du zu Silvester 2025 Feuerwerk?
Gern teilen für mehr Reichweite.
It's been a bad year for the Interpeer Project. Personal stuff has prevented almost any real progress.
I feel like giving up.
What timing: Roland pushed a new KIRA draft today:
This document describes the Kademlia-directed ID-based Routing Architecture KIRA. KIRA is a scalable zero-touch distributed routing solution that is tailored to control planes. It prioritizes scalable and resilient connectivity over route efficiency (stretched paths are acceptable vs. routing protocol overhead). KIRA's self-assigned topological independent IDs can be embedded into IPv6 addresses. Combined with further self-organization mechanisms from Kademlia, KIRA achieves a zero-touch solution that provides scalable IPv6 connectivity without requiring any manual configuration. For example, it can connect hundreds of thousands of routers and devices in a single network without requiring any form of hierarchy (like areas). It works well in various topologies and is loop-free even during convergence. This self-contained solution, and especially the independence from any manual configuration, make it suitable as resilient base for all management and control tasks, allowing to recover from the most complex failure scenarios. The architecture consists of the ID-based network layer routing protocol R²/Kad in its Routing Tier (using source routing) and a PathID-based Forwarding Tier (using PathIDs as labels for paths). KIRA’s tightly integrated add-on services (e.g., name resolution as well as fast and efficient topology discovery) provide a perfect basis for autonomic network management solutions.
... contain fairly specific effort to work within the underlay's limitations.
The bulk of the work lies there. It's another reason I'm a big fan of #librecast, because they're working out one solution here.
I'd also like to see an implementation on KIRA, a zero touch overlay over Ethernet or other Layer 2 networks. KIRA may be aimed at different problems, but it has enough properties to make a likely underlay.
Look, now I did post more 😁
... however is that we've managed to evolve the Internet along a different path. Rather than groups as the communications primitive, we've arrived at point-to-point links.
Layering groups on top, e.g. via IP multicast, is running into problems sometimes created at a lower stack level specifically to optimize point-to-point behaviour.
So the implementation of a theoretically complete system will need to be able to work with underlays of various kinds that either follow a similar model, or...
So the core of the project is: how do you design a system that spans the extremes of those dimensions?
They're boundless, phrased as I put them. So to rephrase: where do you put the reasonable bounds, and why? And then design a system that covers at least the range between those bounds?
What started out as "let's do p2p again" turned to this problem over time. So if we manage to solve that (which seems feasible, i.e. I think I know how), then we've come a long way.
The interesting part...
Thomas Cloer