It has been said many times that the Coordinator is not necessary because its confirmations will be rejected by other nodes if it acts maliciously.

It follows that if other nodes can tell if the Coordinator acts maliciously, then in fact they do not need the Coordinator.

According to The Transparency Compendium:

This does not mean that the IOTA ledger is currently in any way centralized, the network is 100% decentralized, every node verifies that the Coo is is not breaking consensus rules by creating iotas out of thin air or approving double-spendings

According to eukaryote31 (respected person within IOTA ecosystem):

What if the Coordinator started acting maliciously? Every node looks at the transactions it gets and only tells other nodes about transactions that are valid. The Coordinator is no exception, if the Coordinator starts issuing bad Milestones, nodes will just reject them.

According to iota.stackexchange high rated answer:

The Coordinator is no exception, if the Coordinator starts issuing bad Milestones, nodes will just reject them.

This begs the question, "what makes Coordinator Milestones a security measure if the other nodes are already doing this?"

  • Please link to your source for "the Coordinator is not necessary because its confirmations will be rejected by other nodes if it acts maliciously" – ben75 Aug 23 '18 at 11:15
  • Ok. None of your links state that the coo is not necessary. All your links state that the validity of milestones is verified by the IRI. In other words, the IRI prevent hypothetical malicious actions of the coo. – ben75 Aug 23 '18 at 18:33
  • I'm not posting an answer because the question is still unclear to me. (And regarding your last comment: there is already a similar question here : iota.stackexchange.com/questions/2015/…) – ben75 Aug 24 '18 at 16:28

The Coordinator is about transaction finality. Without the coordinator, "confirmed" would be a probabilistic and subjective notion based on the number of "good" tips which reference the transaction in question directly or indirectly (run the MCMC 100 times and see what percentage of "good" tips reference your tx directly or indirectly). This is not unlike blockchain where "confirmed" is always subjective and probabilistic. It really comes down to the cost of building a longer chain of blocks which has a different history, and with the tangle its the cost of building a heavier branch, neither of which could ever be infinite. In Bitcoin, the current rule of thumb is after 6 blocks or about 1 hour, you can safely consider a transaction confirmed because for someone to build a longer chain to re-write that history would be very expensive (but not impossible).

In the current low throughput regime of the Tangle, however, this cost is not very high even after waiting many days, so the coordinator acts as a tx finalizer so that transactions can be considered final in a reasonable time frame.

  • 1
    It does not reach transaction finality according to whitepaper spec. We can't say what it does for sure (because no open source). We can assume, it starts a random walk at the latest milestone, does the default IRI checks and does some extra secret COO checks to determine the transactions it references. – Zauz Aug 24 '18 at 8:21
  • The time frame is determined by the cumulative weight of a transaction. With current TPS, the time would be very big until a tx has enough weight to be considered irreversable. – Zauz Aug 24 '18 at 8:35
  • >Doesn't the Coordinator reach the decision of transaction finality according to the whitepaper spec? No. The whitepaper spec assumes no "authority". The Coo transactions, called milestones (or called 'checkpoints' in the white paper) are confirmed by definition, and everything they reference is also confirmed as a result. Without a Coordinator, there is no 'authority' like this to be the transaction finalizer. – John Licciardello Aug 24 '18 at 19:03
  • >Does that mean that without the Coordinator the real function of the Tangle would result in unreasonable delays confirmation time? Why is that also? What aspect of the confirmation takes so much time? No because the idea is with a lot of transactions the time would be short. Right now there isn't a lot of transaction. Each transaction requires a tiny amount of PoW, but this adds up quickly. To build a heavier branch when the transaction in question already has a lot of transactions built on top would be very expensive. – John Licciardello Aug 24 '18 at 19:04
  • @AustinPowers Yes, of course we can assume what software does without having it's source code by analyzing it's behaviour. – Zauz Aug 24 '18 at 19:59

Milestones are bundles that reference a certain part of the Tangle. All transactions within this part are considered confirmed by those who trust the COO. In my eyes, milestones behave very much like a transactions with infinite weight. Other nodes can only issue transactions with a weight of 1 (Maybe more in the future). The problem with that is, that malicious actors could issue a lot of transactions with weight 1 to build parasite chains or similar attacks. To counter that, we use transactions, currently issued by a central authority, with infinite weight.

This is assuming, that the Coordinator does not start acting maliciously. If it would, human intervention would be necessary.

One could argue, that the statement in The Transparency Compendium: "the network is 100% decentralized" is a bit of an overstatement. But the point is that it is still decentralized enough even with the Coordinator.

  • Your first comment isn't math related nor a valid argument. You're just naming some of the centralized aspects of the current IOTA protocol (ignoring the decentralized ones), then appending "== true" (what does that even mean?) and "== 100% Centralized" (where did you get the number 100% from?). ; I wouldn't suggest trying to calculate or estimate to what extent Iota is centralized. I think this question is easier to answer by just thinking of worst case/bad case scenarios and their short-term and long-term consequences. If you do that you may realize that most of them aren't that devastating. – Zauz Aug 24 '18 at 19:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy