Topic: error - page 30. (Read 360507 times)

Hello, peoples.  It's 11:58 EDT and the spike's in swing:

48 coins in 10 minutes, that's a coin every 12 seconds, a full factor of 10x out-of-spec.

I'm an RPC bagholder too -- never yet sold a share -- BUT THERE'S A HOLE IN MY BAG!

33275   2014-03-21 03:48:05   5   5.99691215   4.698   33276   39.9777   87.2489   20.4162%
33274   2014-03-21 03:47:48   1   1   4.698   33275   39.9798   87.2487   20.4141%
33273   2014-03-21 03:47:45   1   1   4.698   33274   39.981   87.2486   20.4141%
33272   2014-03-21 03:47:42   1   1   4.698   33273   39.9821   87.2486   20.4141%
33271   2014-03-21 03:47:39   1   1   4.698   33272   39.9833   87.2486   20.4141%
33270   2014-03-21 03:47:28   1   1   4.698   33271   39.9844   87.2484   20.4142%
33269   2014-03-21 03:47:26   1   1   4.698   33270   39.9855   87.2484   20.4142%
33268   2014-03-21 03:47:10   3   6.74751769   4.698   33269   39.9866   87.2482   20.4143%
33267   2014-03-21 03:46:44   1   1   4.698   33268   39.9875   87.2479   20.4143%
33266   2014-03-21 03:46:36   1   1   4.698   33267   39.9886   87.2478   20.4143%
33265   2014-03-21 03:46:24   1   1   4.698   33266   39.9897   87.2477   20.4144%
33264   2014-03-21 03:46:18   4   3.67816768   4.698   33265   39.9908   87.2476   20.4144%
33263   2014-03-21 03:46:03   1   1   4.228   33264   39.9919   87.2475   20.4145%
33262   2014-03-21 03:45:50   1   1   4.228   33263   39.9929   87.2473   20.4146%
33261   2014-03-21 03:45:36   1   1   4.228   33262   39.9939   87.2472   20.4146%
33260   2014-03-21 03:45:30     1   1   4.228   33261   39.9951   87.2471   20.4146%
33259   2014-03-21 03:45:13   1   1   4.228   33260   39.9961   87.2469   20.4147%
33258   2014-03-21 03:45:10   1   1   4.228   33259   39.9972   87.2469   20.4147%
33257   2014-03-21 03:45:07   2   1.2661781   4.228   33258   39.9984   87.2468   20.4148%
33256   2014-03-21 03:44:55   1   1   4.228   33257   39.9995   87.2467   20.4148%
33255   2014-03-21 03:44:54   1   1   4.228   33256   40.0007   87.2467   20.4148%
33254   2014-03-21 03:44:39   1   1   4.228   33255   40.0017   87.2465   20.4149%
33253   2014-03-21 03:44:24   2   4.0002   4.228   33254   40.0027   87.2463   20.415%
33252   2014-03-21 03:44:19   2   1.21448695   4.228   33253   40.0039   87.2463   20.4149%
33251   2014-03-21 03:43:55   1   1   4.228   33252   40.0048   87.246   20.415%
33250   2014-03-21 03:43:47   1   1   4.228   33251   40.0059   87.2459   20.415%
33249   2014-03-21 03:43:26   1   1   4.228   33250   40.0069   87.2456   20.4151%
33248   2014-03-21 03:43:08   1   1   4.228   33249   40.0079   87.2454   20.4152%
33247   2014-03-21 03:43:07   5   10.18466639   4.228   33248   40.0091   87.2454   20.4152%
33246   2014-03-21 03:42:49   2   2.02329224   4.228   33247   40.0102   87.2452   20.415%
33245   2014-03-21 03:42:46   1   1   4.228   33246   40.0114   87.2452   20.415%
33244   2014-03-21 03:42:32   1   1   4.228   33245   40.0125   87.245   20.415%
33243   2014-03-21 03:42:29   1   1   4.228   33244   40.0137   87.245   20.415%
33242   2014-03-21 03:42:05   3   3.00166341   4.228   33243   40.0146   87.2447   20.4152%
33241   2014-03-21 03:41:19   1   1   4.228   33242   40.0153   87.2442   20.4153%
33240   2014-03-21 03:41:09   2   6.63722356   4.228   33241   40.0164   87.2441   20.4153%
33239   2014-03-21 03:40:41   1   1   4.228   33240   40.0174   87.2437   20.4152%
33238   2014-03-21 03:40:34   1   1   4.228   33239   40.0185   87.2437   20.4152%
33237   2014-03-21 03:40:29   2   3.0001   4.228   33238   40.0197   87.2436   20.4152%
33236   2014-03-21 03:40:18   3   3.02159661   4.228   33237   40.0209   87.2435   20.415%
33235   2014-03-21 03:40:03   1   1   4.228   33236   40.022   87.2433   20.4149%
33234   2014-03-21 03:39:59   1   1   4.228   33235   40.0232   87.2433   20.415%
33233   2014-03-21 03:39:55   1   1   4.228   33234   40.0243   87.2432   20.415%
33232   2014-03-21 03:39:49   2   3.0001   4.228   33233   40.0255   87.2431   20.415%
33231   2014-03-21 03:39:18   1   1   4.228   33232   40.0263   87.2428   20.4151%
33230   2014-03-21 03:39:15   1   1   4.228   33231   40.0275   87.2427   20.4151%
33229   2014-03-21 03:39:15   1   1   4.228   33230   40.0287   87.2427   20.4151%
33228   2014-03-21 03:39:10   2   1.79804272   4.228   33229   40.0298   87.2427   20.4151%
33227   2014-03-21 03:38:11   3   3.98148348   4.228   33228   40.0304   87.242   20.4154%

In the same boat...

I wonder if things would have been better if we never forked at all? 

BTW, I never sold any RPC. Bagholder all the way.

RPC diff currently 4.6. This is the lowest I can ever remember seeing it aside from shortly after launch.

I'm in. I think BTC has been infiltrated by the wrong people who want to see it regulated and centralized. I hope that this doesn't happen, but in the meantime I am holding and buying RPC as I believe in the message of liberty and respect Dr Paul as well.

Coins are an investment but they also should be a fun hobby. I'm proud to have RPC in my crypto collection. Is there anyone making physical RPC's?

Yes, I know there is digishield out there, but is it better?  The precise algorithm isn't specified there.  Doge's adoption of it will be an interesting test, and the code is out there, so it is hopefully easy to adopt, which is a plus.

Earlier, I remarked that DigiShield seems to retarget diff downwards more slowly than upwards, so the question remains: is a two-window system that distinguishes spikes like what I proposed superior?

Total net hash rate is 20MH/s right now, folks, with half of it on Chunky Pools.  Next spike is going to be brutal.

Because our present difficulty adjustment algorithm only copes by capping maximum swing up or down to 11%, we're seeing 15+ min block times instead of 2.  I suggest we get the issue debated and solved quickly.

And it's indisputable at this point that a single user could launch 51% attack.  40MH under one hood is not hard to garner (60%), 80MH would give them 80% (80MH out of 100MH), a nice advantage.

The time to act and promote is now, my friends.

http://www.cryptocoinsnews.com/2014/03/13/dogecoin-forks-again-to-avoid-multipool-exploit/


http://www.reddit.com/r/Digibyte/comments/1zjeso/eli5_what_is_digishield/

New Block Difficulty Algorithm Proposal:

The Problem: We need to differentiate a "spike" from "regular" total net hash rate to keep block rate steady at 1 block every 2 minutes, or 30 blocks an hour, in order to respond to large miners jumping in and out of mining RPC.

Considerations: The block difficulty needs to adapt quickly, so it needs to evaluate blocks solved more frequently than it does.  If it doesn't, it misses spikes and blocks fly by every few seconds.  The second challenge, however, is how to return to baseline.  We must flexibly handle spike difficulty settings vs. regular difficulty settings so that a return to lower difficulty doesn't take a long time when the spike is gone, otherwise confirmations take a long time and small miners lose interest.  This second challenge requires evaluation of the block rate over a long time to establish a non-spike period's "baseline" mining rate.

So we need two windows of time consideration, a short-term window to detect spikes and a long-term window to establish regular mining rates.  The math should be as simple as possible for testing and transparency.  I propose using quartiles, because it's as simple as it gets for identifying outliers:

http://en.wikipedia.org/wiki/Quartile

Proposal:
Evaluate total net hash every 5 blocks, or, if feasible given network communication speeds and interface with MPOS, every block.  Then distinguish whether we are in a spike based on the net hash rates over the past 180 blocks (=6 hours of blocks according to the desired rate of 30 blocks per hour.)  If we are in a spike, adjust difficulty upwards quickly.  Stay there until spike passes.  When spike passes, revert difficulty to a factor based on the median block rate for the past 180 blocks.

Algorithm:
Every 5 blocks (or 1 block if possible), calculate:

Quartiles for the total estimated net hash rate based on the past 180 blocks.

Are we in the upper quartile = in a spike?  (Is the present net hash rate in the top 25% of net hash rates we've seen in the past 6 hours?)

Yes: Adjust difficulty immediately based on present net hash rate, to retarget to 2 mins per block.

No: Set difficulty based on the median net hash rate over past 180 blocks to target to 2 mins per block.

Advantages:
Rises quickly to meet rapid increase of hashrate, preventing plundering of blocks every few seconds.

Falls back to normal quickly when the spike is gone, preventing long delays in confirmations when net hash rate returns to normal.

Distinguishes a spike from a long-term increase of hash rate if the coin gets popular.  In 3 hours (90 blocks), a long-term high hash rate becomes the median normal hash rate.  Interim adjustments should be smoother, based on the median, rather than the average.

Verification:
Anyone have a CSV formatted list of timestamps and block numbers handy?  I could grab a list from altexplorer.net, but I would need to muck around reformatting it.  This could be tested in a spreadsheet first, and then on a testnet if RPC has one.

Tweaks:
(a) Defining a spike: Instead of looking at the upper quartile, we could define a spike more narrowly as a net hash rate in the uppermost 10% or other percentage of the blocks we've seen over the long-term window.

(b) Length of long-term window: I picked 180 blocks / 6 hours as a safe range of time to establish the "regular" net hash rate.  This window could be longer or shorter, but it should remain a high multiple of how many blocks we wait before recalculating the difficulty.

(c) Evaluating every 5 blocks would be pretty fast (10x faster than present) but the quicker the algorithm adjusts, the fewer blocks are lost to plundering.

Your thoughts??

+1 Turpoman!   So, who wants to do what?

I'm interested in:
Establishing a link to a precious metals merchant
then, making a video to advertise the merits of the coin
then, FB marketing.

BTW, my wife works for a free-market research center at a local university with a specialist in cryptocurrency.  If the coin can polish its image, maybe we can turn more libertarians on to it as a serious alternative.

Turpoman, your English is flawless.  If you hadn't said it, I don't think anyone would have guessed.

More than anything, RPC community need individuals like this: https://www.youtube.com/watch?v=I14b-C67EXY

We have most excellent brand and our community should take full advantage of it! We need passionate people to promote the coin!

Developers, developers, developers, developers, developers, developers, developers, developers, developers. Promotion, promotion, promotion, promotion, promotion, promotion, promotion, promotion, promotion. Marketing, marketing, marketing, marketing, marketing, marketing, marketing, marketing, marketing, marketing, marketing.

I would do it, but I'm Finnish and my english sucks. We need real Americans who are passionate on liberty and share Ron Paul's ideals.

The truth is: the coin has existed for about 3 1/2 months. That's a very short period of time!

It's amazing how many things can change in such a short period of time in the crypto world, and it's also easy to get caught up in the little changes.

Bitcoin had plenty of its own crashes in its lifetime. I shouldn't need to remind anyone but it seems easily forgotten.

At the same time, RonPaulCoin will be around for many years. So this first 3 months is a drop in the bucket. We will evolve as necessary to roll with the punches and handle the blows the that evolving crypto world might hand out. Who knows what exploit will be discovered in the weeks or months to come. It's about being able to adapt that will ensure survivability. So once we figure out the best solution to the current issues of multi-pools we will implement them.

Hi Profjski,

Some pools handle what has been happening better than other, the Co-op was my
preferred site but wont update though it still seems to be mining and discovering
new blocks, but will stay away until it updates just in case.

I dont think these attacks are from asic's, if the coins are earned then they will in all
probability be multipools, some of these pools are ginormous, but what is mining in
these middlecoin and clevercoin pools is probably increasingly asic gripseed miners.

This will be the next challenge to us, when these asic's become widespread they will drive
the profitability of mining with a GPU to below break even (we are near or at that
now here in Australia with around 26cKWH).

So you will have the dominant BTC, and then LTC will keep going up in difficulty
as more asics move there (and the price should follow it up), but the rest of script coins
will be driven down in price as the coins will be dumped by those for profit asic miners,
and switched into BTC or LTC coins.

Then the remaining coins will evolve into "N" adaptive or similar where more memory
will be required to mine these(to fight the asic's), I think that being one of the first
adapters of the move to more memory intensive forms, will stand us in better chance to be not swept under by these new asic miners.  These "N" adaptive coins increase the memory
required to mine them as technology improves. What will slow the development of asic
"n" script miners is the cost of this memory, but it looks likely that "N" script could stay
one step ahead of present and future asic's, and allow the GPU miners a chance to mine
a profitable coin.



The no premine was a blessing, but now because there is no coin available to spend
on development and support, it is working against us IMO, yep the coin is cheap, but so
are most others, it will be survival of the fittest and strongest communities.

Not sure about digishield, but at least they are doing something, it looks pretty good to
me but really need to see how it performs in the real world, not just in tests.

Not sure about the technical side of things with MPOS and the like, but to me a one block
change will use a lot of computer power every change, i am willing to let more informed
heads work this one out, but I believe a 1 to 4 minute fixed minimum and maximum time
would solve most of our problems, if this was technically achievable.

I have noticed that these multipools are now earning considerably less than a few weeks ago
dropping to below LTC mining returns, this is probably due to all the different coins finding
ways to decrease the influence of these pools, clevermining by the looks of it is down 3GHs
from when I last checked, so maybe the window that these pools had to exploit the coins
is closing, and we might be being picked on more often because we have not adapted fast
enough, we still have time on our side to make a good choice for the future direction of this
coin, but things are moving and evolving so fast, a week of two is like a normal lifetime.


cheers   

I'm leaning toward the fact that someone is doing that timestamp attack like what was discussed earlier about the possibility of attacking a KGW coin.  I doubt someone is spending 5GH on ruining a coin.  It's likely a script being used to silently take advantage of a loophole. 

Is there some way to filter/block that IP range or connection when it is recognized, via the stratum server software or something?  Any way to find out which connection is submitting those counterfeit blocks?  Seems as though it's not coming from a pool.  Probably NSA, they would like to censor anything Ron Paul lol.

rEVOLution and I have also proposed this on the RPC forum.  Initially I think devs shied away from this because it was called a "centralized" pool.  The proposal is not about centralization; it's about a dedicated home.  Right now, I'm holing up at Chunky's Pool, because it's been stable throughout these hashrate spikes, and is doing fine now at low diff and a few blocks a minute.

Regarding the megahasher(s): How do you know it's ASIC and not just a p2pool or multipool powered by a couple hundred people running traditional multi-GPU rigs?  It's not hard to command 10-12 MH/s personally.  I have only 1.2 MH/s, but rev has 12, and neither of us run dedicated rigs, just gaming PCs with a few extra cards.

I'm skeptical that a move to Scrypt-N would solve any such problems for long; if people are sinking money into developing ASICs, they will sink money into developing Scrypt-N ASICs.  The technology will follow opportunity for easy mining.  To me, it's about establishing a community of miners behind the coin.  I'd be happy to buy an ASIC and park it on RPC all day, every day.  When the regular daily total net hash is good, a 1 GH/s spike means a lot less.  The problem is a low number of regular miners.  Until that is solved, anyone with coin mining power will be a major player, and I don't think there is a long-term way to shield from specialized mining power.  Maybe I'm wrong, but I'd like to see the argument.


I don't see how no premine hurts any.  When the coin is dumped, everyone with cash to invest is in a position to invest easily.  Having a chunk of premined coins only means the holders lose value because they spent time to premine coin before market price was established.

Re: DigiShield:
I thought about our problem the other day, and thought of this, but dismissed it.  The solution DigiShield proposes seems to be: permit difficulty to rise quickly during hashrate spikes, but let it decrease slowly after that.  Upwards difficulty changes happen fast, downward difficulty changes fals off slowly.  Isn't the obvious con what ColinIsTheMan encountered before the first difficulty fix?  Namely, when the large megahashers leave, and difficulty falls off more slowly, coin transactions will take a looong time?  Maybe people are now willing to accept this as a trade-off, but if lots of people jump ship and mine something else during long difficulty fall-offs, the cycle will just repeat based on how many blocks per retarget, no?

If anything, a diff retarget every 5 blocks instead of 48 might help.  Difficulty retarget every block seems like it would cause major problems for pooled mining.  As far as I've seen it, when blocks are solved quickly, MPOS pools have a hard time keeping up with the network status updates.  Would a diff retarget every block hang MPOS pools?  And if a block retarget hard fork is going to be done, could we update it on MPOS too, since the next estimated difficulty hasn't been accurate since the last hard fork?

JSki

If it comes down to it, something like this looks promising: http://www.reddit.com/r/Digibyte/comments/1yn6t1/digibyte_v_20_code_name_digishield/   (DigiShield v2.0)

If nothing else it at least offers a helpful information toward a good solution.

I would want to see how it pans out first though. It's interesting that the "big solution" of KGW that so many coins implemented a short while back is now backfiring and proving to be an exploitable "fix". So I would want to make sure that any new fix is actually an improvement and doesn't open up any future issues/holes.

G'day RPC folk,

My take on the last 48 hours or so regarding the mega hashes.

I still have not witnessed the net hashometer go over 500MHs, I have estimated
that probably around 5 GHs would be required to win the number of blocks at the
speed they do, (I know the hashometer goes over one gig as it did say a month ago,
when we had more miners mining), so my question is still out there, are these
really fast coins being earned or stolen, is there a glitch in the existing code.

Has anyone witnessed the net hashrate go over 1GH in the last week?


For most of the last 48 hours during attacks, it seems that the megahasher was
a bit different than normal, they appeared to be keeping around 100 RPC and selling
them slowly when they were not mining our pool, they would then mine and replenish
their stash of RPC, then do the same thing again, this allowed the price of RPC to
be more stable than what has been happening over the previous week or so, this megahasher
has found nearly 3000 RPC over the last 20,000 or so blocks, so they have been doing
it for a while.

The other megahasher appears to be just a straight mine and when confirmed, dump at market
kind of multipool.  But as I said I still have not seen proof of their hashes, the blocks just
seem to fly to them.

I am not a technical expert, I just do a bit of detective work at put 2 and 2 together, check
the blockchain and follow it up on cryptsy, it is pretty simply really.

So, if these megahashers are earning their coin with huge 5GH plus power, then is it a bad
thing for them to be doing what they are doing to RPC and other coins.

I believe it is, it strips the easy or less difficult coins out, and throws our algorithm out,
so we then go from feast to famine, when a block will take 10 to 40 times what the coin
was meant to do, it also drives the price of RPC lower because these coins are sold on
cryptsy, either slowly or dumped, but both ways still work against the price eventually.

I think we all can see the problem they are creating, this is being exacerbated by the low net
hash rate, that in itself is being caused by miners giving up mining, leaving because of what
these mega pools are getting away with.

Now that Colin and the devs understand what is happening, we are in a pretty good position.

Strengths

There has not been that many coin mined, no premine.

We have a purpose and as has been said the desire to see this coin thrive is also linked to
its ideals, we are all here for a similar reason, honest ideals.  This is not a kitty kat or penguin coin.

The original parameters of 1 coin every two minutes, seem to be pretty fair, if we can get it
back there, these parameters seem good for both miners and investors.

Weaknesses

No premine, we are not in a financial position to support the coin in the market, so if the
coin gets dumped the price gets hit hard.

We dont have our namesakes blessing, without this we will always be in a shadow of being
asked to disassociate our coin from him, if his blessing is not forthcoming we really need
a plan B, also Ron Paul appeals mainly in USA, the altcoin market is a worldwide market.

The ability and ease of these mega hash attacks.

The fact that there are now gripseed altcoin miners available, these will IMO drive all altcoins
(that are not in some way immune from them) down in price, we have already seen evidence
that most coins now take between 100 and 200 days to mine one BTC worth per 1MH, when
this figure was between 50 and 100 days only a few weeks or a month back.

I believe that we will see most of the current coins disappear, and we will be left with probably
six to ten coins, our objective is to be one of those few coins that make it through.

We have a chance now to make some changes in this coin to give us the best possible
opportunity to be one of the survivors.

What I would like to see

A mining pool be set up for us to mine RPCs, (at present the pools are all over the place,
and are being shut down by these attacks).  Where committed miners dont have to switch
between pools.  This could help build a sense of community and solidarity between miners,
it could also have a mandatory percentage donation to help fund the further development
of this coin, and could help  prevent a 51% attack.

Look at changing the coin to maybe an "N" adaptive, similar to vertcoin, I can see vertcoin
being one of those surviving coins because of the features it has to prevent the threats
that most altcoins have that I mentioned above, if this is not acceptable then change the
48 block difficulty changes to somewhere around 5 to 10 block difficulty change, but I dont
see this fixing the problem, just somewhat like putting a bandaid on it.
I still think that if we could put a minimum and maximum timeframe  in the algorithm, it would
make us somewhat immune from the megahash attacks.

Get Ron Pauls blessing, and use his image on the coin, if he says no, we look for a plan B.
but be proactive and not reactive, I am sure we could come up with some great ideas for a name
that could honour Ron Pauls and many of our own ideals and values.

We need a reason for this coins existence, we need to distinguish ourselves from the
other now multitude of altcoins.


cheers



 
  
 
 
  

+1 to relb: my desire to see this coin thrive is also linked to its ideals.

I am wondering what other readers think about the botnet hypothesis I wagered?

Even if the burst in mining speed is traditional pump-and-dump by large megahasher(s), there are two sides to the equation: supply and demand.  Mine-and-dump gluts the market with supply, lowering coin price.  From the looks of it, today I finally saw the dump on Cryptsy, where the coin fell from 0.0028 briefly past the buy wall at 0.0022, and then the wall came up again.  A few hundred coins were dumped.

Demand is the other side of the equation.  A vendor who will sell precious metals for RPC would create demand, since the coin now has a direct use, and one of interest to currency speculators, not to mention Ron Paul fans.  Advertising the ideals of the coin would also create demand.  If demand translates into more interested people mining, the spikes caused by a few large megahashers will have less of an effect, cause less disruption, and ultimately mean less % of coins mined to be dumped.

What think you all?

Jski