4

Suppose a transaction is deemed confirmed at some moment and it is put in the milestone.

Somewhat later new tips come in which are not referencing (direct/indirect) this particular transaction and its confirmation-probability drops to 0.0000001

However since it is already in the milestone, it will still be considered confirmed. Is my line of thought correct and if so this might be a problem right?

  • This cannot happen. Confirmation (in the current implementation) is based solely on milestones (everything that is below a milestone is confirmed, everything else not). Milestones have to approve the previous milestone (and no conflicting transactions) to be considered valid. – mihi Dec 29 '17 at 20:29
  • ok ok so a milestone is considered final then, but it can still contain a false transaction right? and then there is no way anymore to get rid of this false tx. without milestones the confirmation-probability would have a dynamic nature and would be able to drop under your treshold at some point..... – GJEEE Dec 29 '17 at 21:04
  • 1
    The coordinator, which issues the milestones, does of course not include wrong transactions. If it would do, the full nodes would ignore the milestone. – Werner der Champ Dec 29 '17 at 21:43
  • @WernerderChamp so the coordinator actually checks every transaction himself before he puts it in a milestone???? now i understand. this goes against the distributed nature but i realize it is only for the time-being till the tangle matures. i am very interested to know how the consensus-mechanism will look like in the post-coordinator age. – GJEEE Dec 29 '17 at 21:55
  • iota.stackexchange.com/questions/48/… you might want to read this one – Werner der Champ Dec 29 '17 at 22:13
2

The relationship between confirmation and milestones is still unclear in the current implementation. While the confirmation process still depends on the Coordinator, milestones are not necessary the indicator of valid transactions.

(see the attached image *full-credit for @Abmushi)

"Milestone" is a bundle that contains two txs and references the previous milestone and another random tx. Every tx inside a transaction bundle, expect the last index one, references a milestone. If all valid transactions that reference a milestone can get confirmed or only transactions that are (directly/inrectly)referenced by a milestone get confirmed; in other words, if milestones matter so much, I think there is no point of getting referenced by other transactions and the confirmation rate of the current network should be much different: either much better or worse.

Bundle Cheat Sheet

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.