69

The network stats show a constant average blocktime of around 17 seconds. Whenever you check it is around 16~18 seconds.

enter image description here

Reading the blog post by vitalik buterin it states the block time should be 12 seconds on average. (Update: Ok, more recent resources state a 15 second target time.)

Why is the average block time constantly at around 17 seconds?

Improve this question
1
  • 2
    Great question, and a clear answer. Don't forget to upvote these people! We want to have lots of clear shining examples for other people to follow on the site.
    – Jeff Coleman
    Jan 21, 2016 at 0:19

1 Answer 1

Reset to default
59

Due to advances in blockchain research, it was shown that significantly lower block times were possible and perhaps beneficial given the current connectivity of the internet. One of the potential risks of a low block time is a higher rate of orphaned blocks (competing blocks that do not make it into the main chain). To counter this, a GHOST protocol is used which pays for these orphaned blocks (known as uncles), adding to the security of the main chain. Instead of the main chain being "longest", it is instead "heaviest".

During the development of Ethereum, several different times and targeting algorithms were tested, eventually settling around 15s. Due to the specific total difficulty calculation used, the actual time between blocks is currently around 17 seconds, however a change scheduled for Homestead will hopefully alter this closer to 15 seconds. This is only proposed as a temporary solution, as Ethereum intends to transition to proof-of-stake, potentially offering near-instant finality.

Improve this answer
6
  • 1
    Good answer. But it's a bit misleading to talk of potentially near-instant finality, because it's really difficult to go below 5 seconds block time. Also, chain finality is only probabilistic until a certain threshold, which may be (much) longer.
    – Symeof
    Jan 21, 2016 at 23:43
  • I disagree @Symeof. It is conceivable that eventually proof-of-stake and other factors such as network size will be able help achieve near-instant finality. Obviously that goal may not be achieved this year, or even multiple years from now, but the changes coming to Serenity seem to indicate future levels of blockchain abstraction that can facilitate making changes to the consensus pieces of the blockchain much simpler. As newer innovations in the space crop up, the network can react.
    – Hudson Jameson
    Jan 24, 2016 at 8:03
  • @Hudson: I certainly hope so. But I don't see how an increase in network size would help finality. Do you have a reason to believe that?
    – Symeof
    Jan 24, 2016 at 9:17
  • 1
    @Hudson: latency plays a major role indeed. It doesn't seem likely that this could be pushed down past a certain point, since the network is global. Also we may need to have a working definition of finality: because in most system it's just a probabilistic threshold (i.e the likelihood of having a transaction reverted is lower than Epsilon)...
    – Symeof
    Jan 24, 2016 at 10:07
  • 7
    "near-instant finality" -> Are we talking 1s, 100ms, 10ms?
    – Randomblue
    Jan 26, 2016 at 6:39

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.