From e7a77ef5024714cd61d7b0b0ba862da007056219 Mon Sep 17 00:00:00 2001 From: Green Sky Date: Fri, 2 Aug 2024 12:17:09 +0200 Subject: [PATCH] update readme --- README.md | 29 ++++++++++++++++++----------- 1 file changed, 18 insertions(+), 11 deletions(-) diff --git a/README.md b/README.md index 8856e75..7312815 100644 --- a/README.md +++ b/README.md @@ -3,7 +3,7 @@ As long as you can contribute an unpredictable (random) number of your choosing, the algorithm will make sure the outcome is unpredictable. -First a public "vote" is held, where everyone secretly generates an unpredictable number and shares a digest to later verify it (HMAC in this proposal). +First a public "vote" is held, where everyone secretly generates an unpredictable number and shares a digest to later verify it (HMAC in this lib). After receiving all HMACS, one sends out the number and starts receiving all the others and verifies them. Now we have random numbers that need to be combined in a predetermined way that mangles them seemingly randomly and temper proof (very expensive). For this a hashing chain is chosen. @@ -11,36 +11,43 @@ For this a hashing chain is chosen. # Algo ## init -InitialState (IS) contains a unique(-ish) id, to uniquly identify this random number -+ any extra data thats usecase dependent. (like the set of cards we are choosing from) +InitialState (IS) contains a unique(-ish) id, to uniquly identify this random number generation ++ a list of all peers participating, ideally with a unique id ++ any extra data, usecase dependent. (like a set of cards we are choosing from) --> hashed to get a fixed sized SI -SI = H(id + user data) - -## rng? -use an unpredictable rng. Simple prng dont cut it, since their state can be reconstructed form very few numbers. +## rng +Use an unpredictable rng. Simple prng dont cut it, since their state can be reconstructed form very few numbers. (use system crng or seed own chacha crng with system crng ...) +This library provides `p2prng_gen_and_auth()`, that provides you with a random number and also directly computes mac and key. +IS is mixed into the random number. + ## hmac message is the rng (while possibly variable in size, should be same as output of H() ) key is random send HMAC to everyone +`p2prng_auth_create()` can be used here, if you dont already use `p2prng_gen_and_auth()`. + ## collection wait for everyone elses HMAC send out secret message (rng) and key -verify everyone elses message +verify everyone elses message using `p2prng_auth_verify()` do not proceed until everything is verified. either hardblock if someone is not responding (to prevent a retry-attack) or exclude unresponsive/lying peer in next (retry) generation. -## post processing +## post processing (combining) +Combine IS with all the numbers -combine IS with all the numbers +Using `p2prng_combine_init()` and `p2prng_combine_update()`. +effectively doing: +``` for each M do H(M + prevH) result = H(IS + prevH) +```