Adam, there is really no justification I can see to lower the interblock interval on the Bitcoin blockchain, primarily due to the effects of network latency. Lowering the interblock interval and raising the block size are not equal alternatives - you can always get more throughput in bitcoin by raising the block size than by lowering the interblock time. And that's without considering the effect shorter intervals would have on e.g. SPV client bandwidth or sidechain connectivity proofs. So I find it very unlikely that such granularity would ever be needed on the Bitcoin block chain, although if were to happen then extra bits from nSequence could be used in a soft-fork compatible way.
However it is true that various sidechains such as Liquid will have a much shorter interblock interval than 10min, as well as customer demand for protocols with shorter timeouts. It would be nice if such systems did not HAVE to resort to complex bit shifting to support more precision, and if protocols written for bitcoin could be reused on such systems with minimal or no modification. To that end, it might be preferable to move the flag bit indicating use of seconds from bit 16 to bit 23 and (by convention only) reserve bits 17..22 to provide higher granularity in a sidechain environment. This keeps the size of a stack push to 3 bytes while also keeping sufficient room for high-order bits of relative lock-time in a sidechain that supports shorter block intervals. Another alternative is to put the units flag in the least significant bit, which has the advantage of allowing both units of lock-time to make use of 1-2 byte pushes, but the disadvantage of making lock times of 64..127 2-bytes instead of 1-byte. Thoughts? On Thu, Oct 15, 2015 at 9:37 AM, Adam Back via bitcoin-dev < firstname.lastname@example.org> wrote: > Does that pre-judge that block interval would never change from > 10mins? Eg say with IBLT or fountain codes etc and security arguments > for the current limitations of them are found, such that orphan rates > can remain low in a decentralised way with 1min blocks, then the > locktime granularity would be coarse relative to the block interval > (with 512s locktime granularity. > > Adam > > On 15 October 2015 at 18:27, Btc Drak via bitcoin-dev > <email@example.com> wrote: > > Alex, > > > > I am sorry for not communicating more clearly. Mark and I discussed your > > concerns from the last meeting and he made the change. The BIP text still > > needs to be updated, but the discussed change was added to the PR, albeit > > squashed making it more non-obvious. BIP68 now explicitly uses 16 bits > with > > a bitmask. Please see the use of SEQUENCE_LOCKTIME_MASK and > > SEQUENCE_LOCKTIME_GRANULARITY in the PR > > https://github.com/bitcoin/bitcoin/pull/6312. > > > > /* If CTxIn::nSequence encodes a relative lock-time, this mask is > > * applied to extract that lock-time from the sequence field. */ > > static const uint32_t SEQUENCE_LOCKTIME_MASK = 0x0000ffff; > > > > /* In order to use the same number of bits to encode roughly the > > * same wall-clock duration, and because blocks are naturally > > * limited to occur every 600s on average, the minimum granularity > > * for time-based relative lock-time is fixed at 512 seconds. > > * Converting from CTxIn::nSequence to seconds is performed by > > * multiplying by 512 = 2^9, or equivalently shifting up by > > * 9 bits. */ > > static const int SEQUENCE_LOCKTIME_GRANULARITY = 9; > > > > I am also much happier with this last tightening up of the specification > > because it removes ambiguity. 512s granularity makes sense within the > > context of the 10 minute block target. > > > > Thank you for spending so much time carefully considering this BIP and > > reference implementation and please let me know if there there are any > > remaining nits so we can get those addressed. > > > > > > > > > > > > On Thu, Oct 15, 2015 at 2:47 PM, Alex Morcos via bitcoin-dev > > <firstname.lastname@example.org> wrote: > >> > >> Mark, > >> > >> You seemed interested in changing BIP 68 to use 16 bits for sequence > >> number in both the block and time versions, making time based sequence > >> numbers have a resolution of 512 seconds. > >> > >> I'm in favor of this approach because it leaves aside 14 bits for > further > >> soft forks within the semantics of BIP 68. > >> > >> It would be nice to know if you're planning this change, and perhaps > >> people can hold off on review until things are finalized. > >> > >> I'd cast my "vote" against BIP 68 without this change, but am also open > to > >> being convinced otherwise. > >> > >> What are other peoples opinions on this? > >> > >> On Thu, Oct 8, 2015 at 9:38 PM, Rusty Russell via bitcoin-dev > >> <email@example.com> wrote: > >>> > >>> Peter Todd <p...@petertodd.org> writes: > >>> > On Tue, Oct 06, 2015 at 12:28:49PM +1030, Rusty Russell wrote: > >>> >> Peter Todd via bitcoin-dev <firstname.lastname@example.org> > >>> >> writes: > >>> >> > However I don't think we've done a good job showing why we need to > >>> >> > implement this feature via nSequence. > >>> >> > >>> >> It could be implemented in other ways, but nSequence is the neatest > >>> >> and > >>> >> most straightforward I've seen. > >>> >> > >>> >> - I'm not aware of any other (even vague) proposal for its use? > >>> >> Enlighten? > >>> > > >>> > There's three that immediately come to mind: > >>> > > >>> > Gregory Maxwell has proposed it as a way of discouraging miners from > >>> > reorging chains, by including some of the low-order bits of a > previous > >>> > block header in nSequence. > >>> > > >>> > A few people have proposed implementing proof-of-stake blocksize > voting > >>> > with nSequence. > >>> > >>> Excellent, thanks! It's good to have such ideas as a compass. PoS > >>> voting seems like it won't be a problem in 5 bits. > >>> > >>> The "prevbits" idea would want more bits; naively 64 would be good, but > >>> I think there are some tricks we can use to make 32 work OK. We would > >>> have to then split between nLocktime (if available) and multiple > >>> nSequence fields, and it would weaken it for some txs. > >>> > >>> There is one easy solution: change the BIP wording from: > >>> > >>> -For transactions with an nVersion of 2 or greater, > >>> +For transactions with an nVersion of 2, > >>> > >>> And on every tx bump, we decide whether to keep this scheme (mempool > >>> would enforce it always). > >>> > >>> Cheers, > >>> Rusty. > >>> _______________________________________________ > >>> bitcoin-dev mailing list > >>> email@example.com > >>> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > >> > >> > >> > >> _______________________________________________ > >> bitcoin-dev mailing list > >> firstname.lastname@example.org > >> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > >> > > > > > > _______________________________________________ > > bitcoin-dev mailing list > > email@example.com > > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > > > _______________________________________________ > bitcoin-dev mailing list > firstname.lastname@example.org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev >
_______________________________________________ bitcoin-dev mailing list email@example.com https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev