justusranvier
Legendary

Activity: 1400
Merit: 1069
|
 |
July 05, 2015, 05:43:22 PM |
|
Thanks for pointing out the difference between the black-box approach and "looking inside the box." I think the black box approach is useful because it forces us to consider the actual behaviour of the complete system, and then ask if our understanding of the inner pieces make sense given these measurements. The #bitcoin-wizards have fallen into that trap before, when the btcd team announced the results of their benchmarking on large simnet blocks. Their complete system measurements were quite a bit slower than what the individual component microbenchmarks suggested. If you look at the comments made regarding their blog post, you'll see the same dynamic playing out as is here.
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 06:28:37 PM |
|
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 06:30:53 PM |
|
Conclusion: As the average blocksize gets larger, the time to verify the previous block also gets larger. This means that miners will be motivated to improve how quickly their nodes can perform the ECDSA operations needed to verify blocks or that they will be more motivated to trick the system.
you missed the most important alternative of all. miners will pare down block sizes for maximum verification times and propagation times.
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 06:52:52 PM |
|
Don't get put-off from posting by Greg's feedback as it is easy to be silenced by someone-who-knows-best and let them run the show. My big takeaway from his comment is how you got him into arguing from a position that big blocks are OK (handclap and kudos to you). In fact, I have learnt now that even 7.5GB blocks are today theoretically tolerable in at least one respect ( validation where tx are pre-validated), although l suspect not in quite a number of other crucial respects. Wasn't Gavin's doubling end-point 8GB in 2036? Effectively the same end-point! I am reminded of Mike Hearn's words of wisdom:Scaling Bitcoin can only be achieved by letting it grow, and letting people tackle each bottleneck as it arises at the right times. Not by convincing ourselves that success is failure.
One less bottleneck to expect with larger blocks. ok, we've all been lead to believe up to now that validation of tx's had to occur twice by full nodes. first, upon receipt. second, upon receipt of the block. this was crucial to the FUD scare tactic of decrementing full nodes and thus "centralization" we were getting last month by the BS core devs via the "large miner w/ superior connectivity attack on small miners". now we here about this new mechanism of "pre-validation" that allows tx's to only have to be validated once upon receipt but not necessarily on receipt of a block? what am i missing?
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 07:02:59 PM |
|
I am reminded of Mike Hearn's words of wisdom:Scaling Bitcoin can only be achieved by letting it grow, and letting people tackle each bottleneck as it arises at the right times. Not by convincing ourselves that success is failure.
One less bottleneck to expect with larger blocks. exactly right. sometimes it takes a hundred steps to walk across the street.
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 07:12:19 PM |
|
interesting, isn't it? We will continue do SPV mining despite the incident, and I think so will AntPool and BTC China.
Another very good reason people should not mine on Chinese pools. This is EXTREMELY bad for the Bitcoin network. yes they will. and do you blame them? until the 1MB limit stops jacking up the unconf tx set thru either spamming or real demand, they can't be bothered to go thru the computation to order or pare down this larger than normal set. SPV is the easiest, simplest way to get around constructing a full block that has the potential to be orphaned. 
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 07:18:14 PM |
|
We just need someone to figure out how to constantly feed them invalid blocks.  I'm surprised to hear this from you, Holliday. Can you explain why you think mining on the blockheader during the short time it takes to validate the block is bad? It is clearly the profit-maximizing strategy, and I also believe it is ethically sound (unlike replace-by-fee). Are they mining on the blockheader during the short time it takes to validate the block, or are they simply skipping validation entirely? Miners are supposed to secure the network, ehh? If they are just blindly mining on whatever they are fed, they aren't doing a very good job, are they? i think they're reacting to the larger unconf tx sets being created as a result of the 1MB cap. they don't want to have to both verify the full blocks coming thru nor construct efficient block sizes from that set. if the cap was lifted, the unconf tx set should drop in size as faster validating miners can swallow the large unconf sets in one gulp and stuff it into a single block to keep the size of the set down thus discouraging further spamming and the deviant SPV behavior.
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 07:23:34 PM |
|
I see no problem with SPV mining while verifying the previous block. It would also make my above suggestion pointless.
Yes. Moreover, I don't see why the time spent verifying is very significant. The vast majority of transactions would have been verified by a full node before being included in the block, and then it is just a case of checking the hash. The verification time for a block would then be insignificant relative to its download time. Great point. Does anyone know how Bitcoin Core currently works in this regards? Is every transaction in a block (re-) verified? Or are the transactions first hashed, and if matched to a TX in mempool not re-verified, thus saving time. but, but pwuille told us that a large block attack from a large miner would choke off small miners! don't forget that the top 5 largest miners in the world have inferior connectivity, not superior, like he told us.
|
|
|
|
|
justusranvier
Legendary

Activity: 1400
Merit: 1069
|
 |
July 05, 2015, 07:44:04 PM |
|
I've learned loooooooooong ago that miners are bottom feeders who have no concept of the idea of protecting their investment. This is a self-correcting problem.
|
|
|
|
|
Peter R
Legendary

Activity: 1162
Merit: 1034
|
 |
July 05, 2015, 07:46:28 PM |
|
and do you blame them?
Yes, it's called cutting off the nose to spite the face. I've learned loooooooooong ago that miners are bottom feeders who have no concept of the idea of protecting their investment. Remember, there are two issues here: SPV mining empty blocks while you're validating the previous block, and what F2Pool appeared to do that caused the problem: never getting around to actually validating the previous block at all! I don't see why SPV mining empty blocks for the short amount of time it takes to validate the previous block is harmful to the network's health (I think it is actually helpful). On the other hand, I see what F2Pool did as hurtful to the network, and they were punished accordingly with a loss of 100 BTC.
|
|
|
|
justusranvier
Legendary

Activity: 1400
Merit: 1069
|
 |
July 05, 2015, 07:51:53 PM |
|
Somebody is buying a lot of bitcoins today.
|
|
|
|
|
molecular
Donator
Legendary

Activity: 2800
Merit: 1023
|
 |
July 05, 2015, 08:09:36 PM |
|
Somebody is buying a lot of bitcoins today.
I think it's at least 2 guys.
|
PGP key molecular F9B70769 fingerprint 9CDD C0D3 20F8 279F 6BE0 3F39 FC49 2362 F9B7 0769
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 10:30:27 PM |
|
I have a feeling that fiber (eu) and equities are going to rally hard on a grexit... Makes sense cause we still haven't hit our long term target and makes sense from a market perspective too.
i don't think so: 
|
|
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 05, 2015, 11:49:07 PM |
|
I have a feeling that fiber (eu) and equities are going to rally hard on a grexit... Makes sense cause we still haven't hit our long term target and makes sense from a market perspective too.
i don't think so:  in case you hadn't noticed; we saw all sorts of these types of gap downs back in 2008. absolutely no chance to get out.
|
|
|
|
|
thezerg
Legendary

Activity: 1246
Merit: 1018
|
 |
July 05, 2015, 11:54:05 PM |
|
Very dangerous. ECB could respond by claiming all Y series notes not legal tender. Of course greece could presumably print other serial numbers easily. And it could invoke the nuclear option where all greek overseas bank accounts are frozen and ultimately confiscated. Where is qoute about double txn validation? Should be unnecessary...
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 06, 2015, 12:15:39 AM |
|
Very dangerous. ECB could respond by claiming all Y series notes not legal tender. Of course greece could presumably print other serial numbers easily. And it could invoke the nuclear option where all greek overseas bank accounts are frozen and ultimately confiscated. Where is qoute about double txn validation? Should be unnecessary... but Euros are fungible, are they not?
|
|
|
|
|
thezerg
Legendary

Activity: 1246
Merit: 1018
|
 |
July 06, 2015, 02:02:35 AM |
|
Very dangerous. ECB could respond by claiming all Y series notes not legal tender. Of course greece could presumably print other serial numbers easily. And it could invoke the nuclear option where all greek overseas bank accounts are frozen and ultimately confiscated. Where is qoute about double txn validation? Should be unnecessary... but Euros are fungible, are they not? They aren't exactly the same so there's a tiny possibility to break fungibility. Any serial # beginning with a Y was printed by the Greek central bank... other countries could give their citizens 1 week (say) to exchange any Y notes that made it across the border for equivalent value. You'd show up at any bank (say) with photo ID and the bills. Of course the press in Greece could probably be modified to change serial numbers pretty easily...
|
|
|
|
|
|
laurentmt
|
 |
July 06, 2015, 02:39:22 AM |
|
miners will pare down block sizes for maximum verification times and propagation times.
I agree that it could be a positive outcome but I "prefer" my option for 2 reasons: - The theoretical reason: there's no solution faster than no validation - The practical reason: from the recent fork, we can see that some (many ?) mining pools prefer no validation than paring down block sizes (e.g. F2Pool mines small or big blocks).
|
|
|
|
|
cypherdoc (OP)
Legendary

Activity: 1764
Merit: 1009
|
 |
July 06, 2015, 03:03:36 AM |
|
miners will pare down block sizes for maximum verification times and propagation times.
I agree that it could be a positive outcome but I "prefer" my option for 2 reasons: - The theoretical reason: there's no solution faster than no validation - The practical reason: from the recent fork, we can see that some (many ?) mining pools prefer no validation than paring down block sizes (e.g. F2Pool mines small or big blocks). what's interesting is that we've never seen it done to the degree it is now. we had the Mystery Miner a few years ago but he stopped it pretty quick. also, despite many upgrades added to the protocol previously, we've never had a fork as a result of SPV mining before either. what's different this time is the consistently full blocks and the fact that Wang Chun told us they create SPV blocks in response to large blocks as a defense. it seems they consider full blocks large blocks so the excessive SPV mining created last nights fork in light of BIP66 and the upgrade to 0.10.x. so in that sense, the 1MB cap is the direct cause of what is happening.
|
|
|
|
|
|