Gregory Maxwell [ARCHIVE] on Nostr: đź“… Original date posted:2014-07-19 đź“ť Original message:On Fri, Jul 18, 2014 at ...
đź“… Original date posted:2014-07-19
đź“ť Original message:On Fri, Jul 18, 2014 at 8:06 PM, Emin GĂĽn Sirer <el33th4x0r at gmail.com> wrote:
>
>> Most things I've seen working in this space are attempting to minimize
>> the data transfered. At least for the miner-interested case the round
>> complexity is much more important because a single RTT is enough to
>> basically send the whole block on a lot of very relevant paths.
>
> Agreed. Yaron's scheme is magical because it is non-interactive. I send you
> a packet of O(expected-delta) and you immediately figure out the delta
> without further back and forth communication, each requiring an RTT.
Oh that does sound interesting— its the property I was trying to
approximate with the FEC.. It achieves the one-shot, but there is
overhead. One plus we have is that we can do some tricks to make some
computational soundness arguments that we'd actually get average
performance on average (e.g. that someone can't author transactions in
such a way as to jam the process).
> In any case, I have no horse here (I think changing the client so it's
> multithreaded is the best way to go), but Yaron's work is pretty cool and
> may be applicable.
Thank you, I've certantly queued the paper for reading.
đź“ť Original message:On Fri, Jul 18, 2014 at 8:06 PM, Emin GĂĽn Sirer <el33th4x0r at gmail.com> wrote:
>
>> Most things I've seen working in this space are attempting to minimize
>> the data transfered. At least for the miner-interested case the round
>> complexity is much more important because a single RTT is enough to
>> basically send the whole block on a lot of very relevant paths.
>
> Agreed. Yaron's scheme is magical because it is non-interactive. I send you
> a packet of O(expected-delta) and you immediately figure out the delta
> without further back and forth communication, each requiring an RTT.
Oh that does sound interesting— its the property I was trying to
approximate with the FEC.. It achieves the one-shot, but there is
overhead. One plus we have is that we can do some tricks to make some
computational soundness arguments that we'd actually get average
performance on average (e.g. that someone can't author transactions in
such a way as to jam the process).
> In any case, I have no horse here (I think changing the client so it's
> multithreaded is the best way to go), but Yaron's work is pretty cool and
> may be applicable.
Thank you, I've certantly queued the paper for reading.