[bitcoin-dev] Hard fork proposal from last week's meeting

Wed Mar 29 09:16:43 UTC 2017

> I suggest you take a look at this paper: http://fc16.ifca.ai/
bitcoin/papers/CDE+16.pdf  It may help you form opinions based in science
rather than what appears to be nothing more than a hunch.  It shows that
even 4MB is unsafe.  SegWit provides up to this limit.

I find this paper wholly unconvincing.  Firstly I note that he assumes the
price of electricity is 10c/kwh in Oct 2015.  As a miner operating and
building large farms at that time, I can guarantee you that almost no large
mines were paying anything even close to that high for electricity, even
then.  If he had performed a detailed search on the big mines he would have
found as much, or could have asked, but it seems like it was simply made
up.  Even U.S. industrial electricity prices are lower than that.

Moreover, he focuses his math almost entirely around mining, asserting in
table 1 that 98% of the "cost of processing a transaction" as being
mining.  That completely misunderstands the purpose of mining.  Miners
occasionally trivially resolve double spend conflicts, but miners are
paid(and played against eachother) for economic security against
attackers.  They aren't paid to process transactions.  Nodes process
transactions and are paid nothing to do so, and their costs are 100x more
relevant to the blocksize debate than a paper about miner costs.  Miner's
operational costs relate to economic protection formulas, not the cost of a
transaction.

He also states: "the top 10% of nodes receive a 1MB block 2.4min earlier
than the bottom 10% — meaning that depending on their access to nodes, some
miners could obtain a significant and unfair lead over others in solving
hash puzzles."

He's using 2012-era logic of mining.  By October 2015, no miner of any size
was in the bottom 10% of node propagation.  If they were a small or medium
sized miner, they mined shares on a pool and would be at most 30 seconds
behind the pool.  Pools that didn't get blocks within 20 seconds weren't
pools for long.  If they were a huge miner, they ran their own pool with
good propagation times.  For a scientific paper, this is reading like
someone who had absolutely no idea what was really going on in the mining
world at the time.  But again, none of that relates to transaction "costs."
 Transactions cost nodes money; protecting the network costs miners money.
Miners are rewarded with fees; nodes are rewarded only by utility and price
increases.

On Tue, Mar 28, 2017 at 10:53 AM, Alphonse Pace via bitcoin-dev <
bitcoin-dev at lists.linuxfoundation.org> wrote:

> Juan,
>
> I suggest you take a look at this paper: http://fc16.ifca.ai/
> bitcoin/papers/CDE+16.pdf  It may help you form opinions based in science
> rather than what appears to be nothing more than a hunch.  It shows that
> even 4MB is unsafe.  SegWit provides up to this limit.
>
> 8MB is most definitely not safe today.
>
> Whether it is unsafe or impossible is the topic, since Wang Chun proposed
> making the block size limit 32MiB.
>
>
> Wang Chun,
>
> Can you specify what meeting you are talking about?  You seem to have not
> replied on that point.  Who were the participants and what was the purpose
> of this meeting?
>
> -Alphonse
>
> On Tue, Mar 28, 2017 at 12:33 PM, Juan Garavaglia <jg at 112bit.com> wrote:
>
>> Alphonse,
>>
>>
>>
>> In my opinion if 1MB limit was ok in 2010, 8MB limit is ok on 2016 and
>> 32MB limit valid in next halving, from network, storage and CPU perspective
>> or 1MB was too high in 2010 what is possible or 1MB is to low today.
>>
>>
>>
>> If is unsafe or impossible to raise the blocksize is a different topic.
>>
>
>>
>> Regards
>>
>>
>>
>> Juan
>>
>>
>>
>>
>>
>> *From:* bitcoin-dev-bounces at lists.linuxfoundation.org [mailto:
>> bitcoin-dev-bounces at lists.linuxfoundation.org] *On Behalf Of *Alphonse
>> Pace via bitcoin-dev
>> *Sent:* Tuesday, March 28, 2017 2:24 PM
>> *To:* Wang Chun <1240902 at gmail.com>; Bitcoin Protocol Discussion <
>> bitcoin-dev at lists.linuxfoundation.org>
>> *Subject:* Re: [bitcoin-dev] Hard fork proposal from last week's meeting
>>
>>
>>
>> What meeting are you referring to?  Who were the participants?
>>
>>
>>
>> Removing the limit but relying on the p2p protocol is not really a true
>> 32MiB limit, but a limit of whatever transport methods provide.  This can
>> lead to differing consensus if alternative layers for relaying are used.
>> What you seem to be asking for is an unbound block size (or at least
>> determined by whatever miners produce).  This has the possibility (and even
>> likelihood) of removing many participants from the network, including many
>> small miners.
>>
>>
>>
>> 32MB in less than 3 years also appears to be far beyond limits of safety
>> which are known to exist far sooner, and we cannot expect hardware and
>> networking layers to improve by those amounts in that time.
>>
>>
>>
>> It also seems like it would be much better to wait until SegWit activates
>> in order to truly measure the effects on the network from this increased
>> capacity before committing to any additional increases.
>>
>>
>>
>> -Alphonse
>>
>>
>>
>>
>>
>>
>>
>> On Tue, Mar 28, 2017 at 11:59 AM, Wang Chun via bitcoin-dev <
>> bitcoin-dev at lists.linuxfoundation.org> wrote:
>>
>> I've proposed this hard fork approach last year in Hong Kong Consensus
>> but immediately rejected by coredevs at that meeting, after more than
>> one year it seems that lots of people haven't heard of it. So I would
>> post this here again for comment.
>>
>> The basic idea is, as many of us agree, hard fork is risky and should
>> be well prepared. We need a long time to deploy it.
>>
>> Despite spam tx on the network, the block capacity is approaching its
>> limit, and we must think ahead. Shall we code a patch right now, to
>> remove the block size limit of 1MB, but not activate it until far in
>> the future. I would propose to remove the 1MB limit at the next block
>> halving in spring 2020, only limit the block size to 32MiB which is
>> the maximum size the current p2p protocol allows. This patch must be
>> in the immediate next release of Bitcoin Core.
>>
>> With this patch in core's next release, Bitcoin works just as before,
>> no fork will ever occur, until spring 2020. But everyone knows there
>> will be a fork scheduled. Third party services, libraries, wallets and
>> exchanges will have enough time to prepare for it over the next three
>> years.
>>
>> We don't yet have an agreement on how to increase the block size
>> limit. There have been many proposals over the past years, like
>> BIP100, 101, 102, 103, 104, 105, 106, 107, 109, 148, 248, BU, and so
>> on. These hard fork proposals, with this patch already in Core's
>> release, they all become soft fork. We'll have enough time to discuss
>> all these proposals and decide which one to go. Take an example, if we
>> choose to fork to only 2MB, since 32MiB already scheduled, reduce it
>> from 32MiB to 2MB will be a soft fork.
>>
>> Anyway, we must code something right now, before it becomes too late.
>> _______________________________________________
>> bitcoin-dev mailing list
>> bitcoin-dev at lists.linuxfoundation.org
>> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>>
>>
>>
>
>
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev at lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170329/c0cf92f1/attachment.html>