Micropayments incentivized bridges

[ple1n]

Various ways of censorship-evasion have been tried, but I haven't seen an actually useful tokenomics-based project. Existing ones are merely VPNs, and some of them don't even support Linux.

Bridge providers should be incentivized to tweak their protocol parameters or offer their own protocols. The recommended ones would be obfs4, and its variant, widely used by Chinese right now, sosistab, or meek which exploits censors' whitelist. There are three ways of censorship-circumvention, obfuscation, mimicry (parrot legit protocols), and tunneling (includes domain fronting and video streams). If bridge providers are to invent their own protocols, how shall they ship the clients. maybe running them in a wasm container.

Is it possible to construct a micropayment scheme on Oxen ? Orchid protocol has a probabilistic nanopayment system, but it requires smart contracts. Should the micropayment be made chain-agnostic ? The users of bridges tend to have anonymity need too.

[majestrate]

Is it possible to construct a micropayment scheme on Oxen ?

on chain payment logic with oxen is the next major feature we want to add to lokinet going forward. right now it is primarily meant to be used to facilitate service discovery and payments with 1 hop snapps on lokinet (exit market). i would like to figure out how to do something similar for network bridging infra. adding that on chain logic to bridging may be slightly more involved vs the exit market but i dont know if it is or by how much, just a gut feeling.

[ple1n]

exit market

I think I read about paid exits somewhere. Is there docs on exit node payment ? Are they paid directly on Oxen layer one, which is spammy if doing micropayments ?

Btw, maybe wOxen should be deployed on eth layer 2 to lower gas fees.

[majestrate]

exit market

I think I read about paid exits somewhere. Is there docs on exit node payment ? Are they paid directly on Oxen layer one, which is spammy if doing micropayments ?

we would want it to be using oxen as that allows us to do really cool things like zk proofs to exits themselves without the exit knowing who is authenticating, just that they paid for transit. microtx arent really needed, as those can be negotiated via offchain side channel (like lokinet) and then have the aggregate published on chain.

Btw, maybe wOxen should be deployed on eth layer 2 to lower gas fees.

i do not see any compelling reasons to involve eth.

[ple1n]

i do not see any compelling reasons to involve eth.

It's listed oxen.io https://ethereum.oxen.io/. It's not to involve. The token bridge can increase liquidity. Not all people can use centralized exchanges to deposit funds.

also, censors probe IPs for bridges, so bridge information is a scarce resource

microtx allows for more fine-grained control. changing bridges should be often

[majestrate]

i do not see any compelling reasons to involve eth.

It's listed oxen.io https://ethereum.oxen.io/. It's not to involve. The token bridge can increase liquidity. Not all people can use centralized exchanges to deposit funds.

this looks like it is a problem addressed by chainflip and is at best an implementation detail that should remain far out of scope from the protocol designs in lokinet.

also, censors probe IPs for bridges, so bridge information is a scarce resource

yes it is, and this is much more relevent to the design of incentivised bridging.

microtx allows for more fine-grained control. changing bridges should be often

i dont think it is a scalable way to do it, you would be better off negotiating in a side channel with all of the bridges you plan on using and transmit a net result. i dont think putting awareness of most of this kind of state into lokinet is desirable.

[nullchinchilla]

so I think the hardest part of bridge distribution is not protocol obfuscation: protocol obfuscation is naturally incentivized if users just pay bridges for traffic, since bridges are therefore incentivized to not b e blocked.

the hardest part is making sure different users automatically retrieve different bridges, over some unblockable channel. the way Geph, Snowflake, etc do it is inherently based on a centralized, non-transparent, permissioned rendezvous server.

[majestrate]

so I think the hardest part of bridge distribution is not protocol obfuscation: protocol obfuscation is naturally incentivized if users just pay bridges for traffic, since bridges are therefore incentivized to not b e blocked.

This is my assessment on it too, the incentives are all about wide accessibility. more ASN diversity more better.

the hardest part is making sure different users automatically retrieve different bridges, over some unblockable channel. the way Geph, Snowflake, etc do it is inherently based on a centralized, non-transparent, permissioned rendezvous server.

i dont think the normal network bridging setup will work out after seeing the pain tor has gone through over the past almost decade now. the idea i have in my head for our net censor mitigation is to do a kind of recursive bridging, basically a full of mesh net which i capable of transiting over the l3 or l2 of your choosing. routing around the DPI by using an alternative l2 looks to be really the only way to truly getting out of networks like chinanet going forward.

as for the incentives, maybe if there was a way to prove network diversity (with zero knoledgeness) given your set of transit bridges you offer in a future theoretical setup, we could structure incentives such that it promotes a wide diversity of ASN.

[ple1n]

basically a full of mesh net which i capable of transiting over the l3 or l2 of your choosing

wdym by l2, l3 ? yggdrasil ?

making sure different users automatically retrieve different bridges

so, the censor is collecting bridges through the faucet. it reminds me of proof of personhood. even PoP isn't enough, because censors can always hire/coerce/bribe people to pass the test. and censors may buy up bridges and block them.

maybe https://appliedzkp.org could be of use. users may anonymously import reputation from sources.