On Sun, May 31, 2015 at 6:55 PM, Alex Mizrahi <alex.mizr...@gmail.com>

> Yes, if you are on a slow network then you are at a (slight) disadvantage.
>> So?
> Chun mentioned that his pool is on a slow network, and thus bigger blocks
> give it an disadvantage. (Orphan rate is proportional to block size.)
You said that no, on contrary those who make big blocks have a disadvantage.
> And now you say that yes, this disadvantage exist.
Did you just lie to Chun?

Chun said that if somebody produced a big block it would take them at least
6 seconds to process it.

He also said he has nodes outside the great firewall ("We also use Aliyun
and Linode cloud services for block

So I assumed that he was talking about the "what if somebody produces a
block that takes a long time to process" attack -- which doesn't work (the
attacker just increases their own orphan rate).

If the whole network is creating blocks that takes everybody (except the
person creating the blocks) six seconds to broadcast+validate, then the
increase in orphan rate is spread out over the whole network. The
network-wide orphan rate goes up, everybody suffers a little (fewer blocks
created over time) until the next difficulty adjustment, then the
difficulty drops, then everybody is back in the same boat.

If it takes six seconds to validate because of limited bandwidth, then he
should connect via Matt's fast relay network, which optimize new block
announcements so they take a couple orders of magnitude less bandwidth.

If it takes six seconds because he's trying to validate on a raspberry
pi.... then he should buy a better validating machine, and/or help test the
current pending pull requests to make validation faster (e.g.
https://github.com/bitcoin/bitcoin/pull/5835 or
https://github.com/bitcoin/bitcoin/pull/6077 ).

If Chun had six seconds of latency, and he can't pay for a lower-latency
connection (or it is insanely expensive), then there's nothing he can do,
he'll have to live with a higher orphan rate no matter the block size.

Gavin Andresen
Bitcoin-development mailing list

Reply via email to