Дан Хюз за Sybill атаката - SMS
---------------------------------------
I know the latest version of the eco is in the pipeline atm and I'm currently working on finishing up the SMS paper (and over next week) now that I've hit it with as many big sticks as I can find in simulators and stuff. Not sure on the eco paper release date, but
SMS will be this side of xmas.
Gotta deliver *something* before Q5.
Context
----------------------------------------
We had a problem with atoms becoming weak after a conflict that an attacker could leverage ... a bit similar to meta stability that RSS models face, but not quite the same due to the temporal proofs. We were working pretty hard to a solution for that and were making good progress,
but then while doing that an economic realisation also surfaced.Due to the way it was used, the energy/mass combo that nodes "own" and we were using to weigh conflict resolutions is worthless.
This led to a secondary market issue where because of the lack of value, an adversary could in theory go and buy up lots of node keys to gain significant influence. Because energy / mass is worthless, sellers have no reason to ask a high price for it.
Now sure, there is quite a bit of logistics involved there, you gotta find all the people to buy from via a pub key, you gotta negotiate with them, pay them, then take the risk they actually hand over the keys and don't keep copies of them.
I spent a good hunk of time trying to quantify those parameters so I could take the decision of the risk of moving forward with it ... too many variables
Because of that, decided it wasn't worth the risk to move forward with it in its current state and to spend some time looking for a proper solution instead of rushing one out where does IOTA come in? well I'll be very surprised if mana doesn't have the same economic problem (based on my current understanding of how its used). Its abundant, easy to obtain and has no intrinsic value.
Nodes do not benefit if they have lots of it or not, only the adversary does. Its probably worse in IOTA because there are no fees, so there is even less depends if its balanced ternary or not incentive to not sell it balance = -1 0 1 unbalanced is usually 0 1 2 fractional = 0 0.5 1 generally depends on the application which is required, but they can obviously be mutated with a simple function similar to big / little endian I guess
if you guys need the full context it's been discussed around Nov 12th
We'll probably roll out with Sybil Mitigation Scheme in public though - SMS.