Arbitrator is not an Alligator

Written by

Qsilver

Jan 2, 2024

Yes, they both start with “A” and end with “ator” but amazingly the qubic arbitrator is not anything like an alligator.

Of course it would be very silly to even think that such could be true, however it seems some people that should do a bit of research before posting have claimed that qubic arbitrator is centralized as it is like the IOTA coordinator. Notice both end with “ator” so the false assumption maybe can be forgiven.

Anyway, rather than rely on random things people (or competitors) on the internet post, I find it much more effective to just read the code and see what it says.

So I did. core/src/qubic.cpp at main · qubic/core (github.com)

I looked for all the amazing things the arbitrator can do. At first it looked like it could steal the 1 trillion mining reward for the epoch. However upon a deeper look the arbitrator transactions happen without any arbitrator signature and just directly update the balances of the computors according to a performance scale. Basically for this the arbitrator address gets the left over funds (around 1% or less recently) without doing anything. The better the computors do, the less the arbitrator gets. Over the lifetime the arbitrator address has accumulated almost a trillion and a half QU, about 2% but at current rates it seems it will take many years to get to 2 trillion QU.

OK, so no free money for the arbitrator, though getting a few billion QU each week from underperforming computors is a nice gig! Guess there are perks to making such a new innovation as Qubic.

Arbitrator also sets the 8 random bytes for the miners each epoch. This can be automated once the RANDOM SC is completed along with the main Arbitrator function of publishing the list of computors for the epoch.

Here is real power! To be able to blackmail any computer to do whatever the evil Arbitrator dictates. Except for the fact that all computor nodes are tracking every tick the performance of all the computors, candidates and newcomers. What that means is that in a future update, the list of computors for the next epoch can be set by the quorum itself and the Arbitrator would just sign it if it matches what it determines to be the rankings.

Such superpower it is not, basically just a rubber stamp on a deterministically determinable list of computors.

The one problem that could happen is if the quorum generated list of next computors is not approved by the Arbitrator. Then we would need to have a QU stake weighted election to determine which slate of computors had the best performance. No politics, just a sorting of scores. This would only happen if the quorum or Arbitrator is infiltrated by an attacker and it would be resolved by the Qubic stakeholders. At worst a temporary few day pause before the new epoch starts. Not good but no disaster and it could be planned on how to deal with such an event ahead of time.

Once the QU vote is done, then the computors would update with a new address for Arbitrator if that is what is decided. Arbitrator is totally replaceable and the job is super easy. Anyway the screams of any computor that is dropped due to Arbitrator malfeasance will be heard very loudly as the competition with the mining power growing so fast is very expensive. A very large scandal would ensue.

So, yes, as of this epoch there is some leeway for Arbitrator misbehavior that can be detected, but when the seamless epoch change is implemented the window for even the detectable misbehavior slams shut, like the jaws of the alligator.

CfB posted the clarification that already Arbitrator misbehavior is not allowed: Come-from-Beyond on X

CfB clarified on discord: “Arb publishes list for newcomers and those without full nodes. Quorum relying on the list (and verifying it) shows the list is correct. Invalid list event would become known to users very soon (by seeing no tick progress).”

So protection from evil arbitrator is already working and the tx with the list of new computors would not be included in the tick, and we enter into the need for a QU stake based election to resolve the split. Then being the scientist he is, CfB said it is possible to get a fork if the election is tied! Seems very unlikely to me but even in such an event we end up with two versions of Qubic (a fork!) like we have BCH forked from BTC and ETC from ETH (or is it ETH from ETC!?)

English

© 2024 Qubic.

English

© 2024 Qubic.

English

© 2024 Qubic.