Gavin Andresen [ARCHIVE] on Nostr: š Original date posted:2015-05-31 š Original message:On Sat, May 30, 2015 at ...
š
Original date posted:2015-05-31
š Original message:On Sat, May 30, 2015 at 9:31 PM, Chun Wang <1240902 at gmail.com> wrote:
>
> If someone propagate a 20MB block, it will take at best 6 seconds for
> us to receive to verify it at current configuration, result of one
> percent orphan rate increase.
That orphan rate increase will go to whoever is producing the 20MB blocks,
NOT you.
Or, we can mine the next block only on
> the previous block's header, in this case, the network would see many
> more transaction-less blocks.
>
Are you sure that is the best strategy? If a big block is slow to
propagate, I suspect it will be better to punish the miner that created it
by refusing to build on it until it has been fully validated.
I'll try to find time to run a couple of simulations.
>
> Our orphan rate is about 0.5% over the past few months. If the network
> floods 20MB blocks, it can be well above 2%. Besides bandwidth, A 20MB
> block could contain an average of 50000 transactions, hundred of
> thousands of sigops, Do you have an estimate how long it takes on the
> submitblock rpccall?
>
I can benchmark it. It should be pretty fast, and sipa has a couple of
patches pending to make the UTXO cache much faster.
It can be fast because the vast majority of the work of validating all
those transactions can happen as they are received into the memory pool.
> For references, our 30Mbps bandwidth in Beijing costs us 1350 dollars
> per month.
You should be able to handle 20MB blocks no problem; if I round up to 100MB
per block that works out to 1.3Mbps.
We also use Aliyun and Linode cloud services for block
> propagation. As of May 2015, the price is 0.13 U.S. dollars per GB for
> 100Mbps connectivity at Aliyun.
That speed will handle 20MB blocks no problem.
If each 20MB block is 100MB of data up/down the wire (I'm vastly
over-estimating, after optimization it should be 40MB) then you'll be
paying...uhhh:
0.1 GB / block-data-on-wire * 144 blocks/day * 30.5 days/month * 0.13 $ /
GB = $57
Less than $2 per day in bandwidth, surely you can afford that.
> For a single cross-border TCP
> connection, it would be certainly far slower than 12.5 MB/s.
That's OK, you'll 1.3Mbps or less.
> I think we can accept 5MB block at most.
>
Are you worried about paying too much, or do 20MB blocks "feel like too
much" ?
--
--
Gavin Andresen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20150531/41d199dc/attachment.html>;
š Original message:On Sat, May 30, 2015 at 9:31 PM, Chun Wang <1240902 at gmail.com> wrote:
>
> If someone propagate a 20MB block, it will take at best 6 seconds for
> us to receive to verify it at current configuration, result of one
> percent orphan rate increase.
That orphan rate increase will go to whoever is producing the 20MB blocks,
NOT you.
Or, we can mine the next block only on
> the previous block's header, in this case, the network would see many
> more transaction-less blocks.
>
Are you sure that is the best strategy? If a big block is slow to
propagate, I suspect it will be better to punish the miner that created it
by refusing to build on it until it has been fully validated.
I'll try to find time to run a couple of simulations.
>
> Our orphan rate is about 0.5% over the past few months. If the network
> floods 20MB blocks, it can be well above 2%. Besides bandwidth, A 20MB
> block could contain an average of 50000 transactions, hundred of
> thousands of sigops, Do you have an estimate how long it takes on the
> submitblock rpccall?
>
I can benchmark it. It should be pretty fast, and sipa has a couple of
patches pending to make the UTXO cache much faster.
It can be fast because the vast majority of the work of validating all
those transactions can happen as they are received into the memory pool.
> For references, our 30Mbps bandwidth in Beijing costs us 1350 dollars
> per month.
You should be able to handle 20MB blocks no problem; if I round up to 100MB
per block that works out to 1.3Mbps.
We also use Aliyun and Linode cloud services for block
> propagation. As of May 2015, the price is 0.13 U.S. dollars per GB for
> 100Mbps connectivity at Aliyun.
That speed will handle 20MB blocks no problem.
If each 20MB block is 100MB of data up/down the wire (I'm vastly
over-estimating, after optimization it should be 40MB) then you'll be
paying...uhhh:
0.1 GB / block-data-on-wire * 144 blocks/day * 30.5 days/month * 0.13 $ /
GB = $57
Less than $2 per day in bandwidth, surely you can afford that.
> For a single cross-border TCP
> connection, it would be certainly far slower than 12.5 MB/s.
That's OK, you'll 1.3Mbps or less.
> I think we can accept 5MB block at most.
>
Are you worried about paying too much, or do 20MB blocks "feel like too
much" ?
--
--
Gavin Andresen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20150531/41d199dc/attachment.html>;