P2P Matrix • Hacker News Board

Re: P2P Matrix

#141

Post by Arathorn » Wed, Jun 03, 2020, 1:28 AM UTC

josh2600 wrote: post #111

Arathorn wrote: post #93

Earlier quoted context omitted.

So it's true that P2P Matrix is currently full mesh (which is why it's staggering a bit as everyone trying it from HN piles on). However, you can absolutely do better than full mesh without going straight back to hub-and-spoke: you can use spanning trees (like Yggdrasil), or gossiped segmentation as libp2p's Gossipsub does ( https://blog.ipfs.io/2020-05-20-gossipsub-v1.1 ). Also, you don't need to scale larger than t…

The problem with hub and spoke is that the hubs hold the social graph and then become the target for censorship. I don't think there are any techniques employed by Matrix to mitigate that threat at this time (please correct me if I'm wrong). In short, if you have a small network (under 10k nodes), I think P2P can work, but for networks of large scale (>10k nodes) I think you need hub and spoke, at which point the rou…

Matrix mitigates the threat of a hub & spoke model by not being hub & spoke (particularly in P2P!) :)

> I am not sure that spanning trees or gossip protocols solve the problem I'm describing, but, if they do, I'd appreciate if you could elucidate further.

Perhaps the confusion here is the expression "hub and spoke" which sounds to me like a static centralised star topology routing model.

My point is that if you have 1M nodes trying to share data (e.g. follow a celebrity's personal pubsub topic), you clearly need a smarter routing algorithm than full mesh (where the celeb's node would have to do 1M parallel pokes to send its messages). You're completely right that one solution is a hub-and-spoke model (where the celeb's node would poke a big centralised hub somewhere, which would then relay the poke out to 1M followers on spokes) - but as you point out, the hub becomes a centralised chokepoint of failure/control/privacy-violation etc.

So, the main other options I'm aware of are either to arrange your 1M nodes into a spanning tree (or overlapping spanning trees) of some kind, as Yggdrasil does (their current one looks like https://yggdrasil-map.cwinfo.org/#, although only ~300 nodes are live atm)... or you let the nodes self-organise into some kind of hierarchy based on gossiping and fan out the messages that way (as per https://blog.ipfs.io/2020-05-20-gossipsub-v1.1, and some of the experimental routing stuff we've been doing with Matrix).

Re: P2P Matrix

#142

Post by WhatIsDukkha » Wed, Jun 03, 2020, 1:28 AM UTC

Arathorn wrote: post #124

WhatIsDukkha wrote: post #99

Earlier quoted context omitted.

It's really not that bad or I certainly wouldn't be using it. The matrix home server has horrible lag spikes at some times of the day. The e2e implementation was pretty miserable until a few weeks ago but is now pretty smooth sailing. Been using it for 3-4 years and quite happy.

As of last week the matrix.org homeserver should no longer be laggy, as per https://twitter.com/matrixdotorg/status/1265412147737223174 fwiw :)

It has seemed much better but anecdotally I was just about to talk to my friend about how much better Matrix had been in the last few weeks and....

Unusuable between 14:20 and 14:28pst today :)

Is there a public ops page for those perf numbers?

-thanks for fixing e2e, cross signing was huge improvement

Re: P2P Matrix

#143

Post by holler » Wed, Jun 03, 2020, 1:30 AM UTC

zozbot234 wrote: post #119

holler wrote: post #85

Earlier quoted context omitted.

Should the very talented and accomplished progressive rock band Isis change their name? I don’t think so.

Um, they did exactly that, in 2018. And it was the right decision to make.

ok true they did a reunion show as Celestial (had tickets to it even), but iirc they didn’t change their name up to that point.

Re: P2P Matrix

#144

Post by hahahahahka » Wed, Jun 03, 2020, 1:50 AM UTC

dpc_pw wrote: post #100

omnimus wrote: post #90

Earlier quoted context omitted.

What is Urbit for? I mean i never understood whats it's real use. From what i've read i imagine it as some new stack that replaces networking and servers but then i found out that you have to buy namespaces that are finite? Why?

So when you blacklist a troll, they can't just create a new account for free.

sounds perfect for state sponsored troll armies, who have way more money/btc/eth than any of us.

Re: P2P Matrix

#145

Post by hahahahahka » Wed, Jun 03, 2020, 1:54 AM UTC

gonehome wrote: post #116

omnimus wrote: post #90

Earlier quoted context omitted.

What is Urbit for? I mean i never understood whats it's real use. From what i've read i imagine it as some new stack that replaces networking and servers but then i found out that you have to buy namespaces that are finite? Why?

I actually don't think it's that hard to grasp, but people often don't explain it very clearly. It's a from the ground up design for a serverless p2p 'overlay' operating system. Overlay just means that you run an Urbit VM on top of a mac/linux OS and people can write applications for your Urbit VM. Your Urbit VM communicates with other Urbit VMs in such a way that the application design is p2p by default so the compl…

and how does one urbit vm finds other urbits vm to establish the p2p?

Re: P2P Matrix

#146

Post by Arathorn » Wed, Jun 03, 2020, 2:15 AM UTC

MayeulC wrote: post #117

josh2600 wrote: post #76

Earlier quoted context omitted.

I understand the goal. I don't understand how a peer P2P system is going to route large groups at scale. I don't think phones are powerful enough to deal with the kinds of routing you need for large scale comms services. In general, as the number of nodes in a comms graph increases, the overhead of synchronizing the comms between those nodes increases to the point that noise dominates signal, which is why most comms…

The good thing with Matrix is that it's room-based, and I haven't really seen rooms with over 10k users. Presumably most of these nodes would be offline most of the time. I can easily see routing being delegated to more powerful homeservers. That said, you do not need perfect routing to be better than hub and spoke at using the network effectively, and it seems to be what most current mesh networks aim for.

the only time you need to coordinate beyond rooms is for presence and for synchronising device list updates, fwiw.

Re: P2P Matrix

#147

Post by Arathorn » Wed, Jun 03, 2020, 2:20 AM UTC

WhatIsDukkha wrote: post #142

Arathorn wrote: post #124

Earlier quoted context omitted.

As of last week the matrix.org homeserver should no longer be laggy, as per https://twitter.com/matrixdotorg/status/1265412147737223174 fwiw :)

It has seemed much better but anecdotally I was just about to talk to my friend about how much better Matrix had been in the last few weeks and.... Unusuable between 14:20 and 14:28pst today :) Is there a public ops page for those perf numbers? -thanks for fixing e2e, cross signing was huge improvement

hm, weird - not seeing anything around 14:20 on the graphs. might be network connectivity or something else going wrong, but not an overloaded synapse for once.

status.matrix.org exists but is fairly useless - we need to publish the graphs publicly.

Re: P2P Matrix

#148

Post by Arathorn » Wed, Jun 03, 2020, 2:23 AM UTC

Andrew_nenakhov wrote: post #53

zaroth wrote: post #15

Earlier quoted context omitted.

I think decentralized DNS with machine readable addresses is basically a solved problem (e.g. Magnet links or Tor Hidden Services). A domain in this case is akin to a hash of a public key, or something like that, just enough to securely identify the target of the communication so there’s nothing to “revoke” although that’s not to say the peer discovery systems could not try to blacklist you. Decentralized DNS with hu…

People want a permissionless (like bitcoin) name service, and they also don't want good names / trademarks to be squatted. However, between these two you can only pick one. You can't solve squatting problem without some kind of authority over issued names.

Social attestations are another solution to this, unless you count them as an authority over issued names. “If all my friends are convinced this person is called Bob, that’s good enough for me”. Doesn’t help with uniqueness though, but that’s what we have keys for.

Re: P2P Matrix

#149

Post by drdeca » Wed, Jun 03, 2020, 2:52 AM UTC

voltagex_ wrote: post #138

core-questions wrote: post #55

Earlier quoted context omitted.

How is the new Urbit stuff, since you mention it? Is it usable for a casual but interested user?

Looks like it's still relying on Ethereum, which IMO is irresponsible from a climate change / energy use point of view.

Do you think that ethereum won’t move to PoS soon? (I haven’t looked into the design, can’t comment on whether the construction seems sound to me)

Re: P2P Matrix

#150

Post by setr » Wed, Jun 03, 2020, 4:34 AM UTC

hahahahahka wrote: post #144

dpc_pw wrote: post #100

Earlier quoted context omitted.

So when you blacklist a troll, they can't just create a new account for free.

sounds perfect for state sponsored troll armies, who have way more money/btc/eth than any of us.

That's true for anything where difficulty is measured by a continuous function -- that is, everything is easiest for a state sponsored anything, good or bad, unless it can't be done at all. You're really just saying states have more resources than individuals... which is true almost by definition (they aggregate individuals within their realm)