2 weeks is not a long time, it's actually a good marker.  The shorter you look, the less meaningful the chart.  Blockchain's 24/48/96 [even though the 48/96 are broken half the time] are useless for getting meaningful data because luck can show a pool twice as large as they should be, or less than half as big as they should be.

Botnets are easy to make.  You can buy it for almost no money if you know how to spread it yourself.  Additionally, pools that can handle a botnet are significantly difficult to make.  You'll run into significant scaling problems trying to support a botnet.  Additionally, botnets tend to be only 5-100 GH/s.  Very few have ever grown much beyond that.  It's much easier to join a pool and just get regular payments than run your own pool, especially when there are so many options/backups to use.

Just as an observation from a pool owner:

The majority of botnets are mining on pools.  They get tossed around [banned] from one to another like a game of hot potato due to how much load they put on the server vs how much they actually contribute.  And yes, they are definitely afraid of ASICs.  I've had two botnet operators beg to be not be banned again because they were getting banned from every pool and they didn't have much time left to make any money now that ASICs are coming.

Glad to hear that you stopped at the 24 hour mark and not as the result of a restart.  Good luck on future testing!  Glad to see the unit is going to good use and not just being stuck on a single pool and ignored.

Regarding earlier comments about share difficulty, 60 GH/s actually hits ~32 on most pool vardiff implementations.  Most vardiff ipools are set between 12 and 30 SPM.  These settings leave 24 hour variance relatively low (+/- 1%).  While an argument could be made for more aggressive vardiff (1-6 SPM), the 24-hour variance does increase a bit when you go much under 12 SPM [unless I'm not remembering organofcorti's post about it].

Stratum servers on 50.31.149.57 were restarted at ~4:05 PM.  Back online as of ~4:08 PM.  The server has been added to the stratum.btcguild.com DNS entry as well, so using that DNS entry will now place you on one of the two stratum servers.


UPDATE:  Both US-based stratum servers were briefly restarted [downtime was about 3 seconds each] to fix a critical bug in the coinbase building.

Makes me wish Dominos still delivered to my house.  All the nearby Dominos went carryout only .  Definitely +1 to Pizza Hut, they have a pretty smooth online checkout system that you could probably automate.  Round Table is another one, though their online ordering system is sometimes a mess.

The Stratum servers on 50.31.149.57 (ports 3333 and 10332) which were the default X-Stratum redirect will be down briefly at 3 PM [PST].  This downtime is expected to last roughly 5 minutes, and is to deploy the updated Stratum/bitcoind code to the server.  After the server is back online, it will be added to the stratum.btcguild.com DNS entry, which is the default X-Stratum redirect for US-based getwork servers at this time.


EDIT/UPDATE:  Got stuck at my day job longer than expected.  Restart of 50.31.149.57 stratum server has been postpone to 4:00 PM [PST]

I think you'd be hard pressed to find a huge botnet made up of headless Ubuntu server machines with GPUs attached.  Just a hunch.

Not exactly.  It's the bitcoin client that does the work.  SatoshiDice introduced a big problem though, which is transaction size.

Satoshi Dice does have fees involved in its payments.  However, miners end up with MASSIVE wallets full of what amount to essentially unspendable inputs [lots of 0.00000001 amounts].  There really is no way to spend 0.00000001 inputs without incurring a fee greater than how much you're trying to send, unless you are able to mine your own blocks [solo mine].

The problem SD presents to the network is the size of blocks has grown tremendously.  Larger blocks take longer to verify, which means the time it takes blocks to spread across the network is increased [both due to the transmission size and the CPU time to verify before it is passed on].  This has increased the rate of orphan blocks.  At the same time, it has dramatically increased the size of the blockchain, which slows down how fast new users can setup and use the default bitcoin client.

As your inputs (payments from the pool) age, eventually the fees will go away.  Each payment from a pool is it's own input, and when you spend your coins, the bitcoin wallet will try to use the oldest coins/fewest inputs to generate a single output [payment/deposit somewhere else].  It's one of the ways the network helps reduce the ability to create transaction spam with insignificant amounts.

I've reported this a few times on BTC Guild.  Miners submitting blank/random byte data in ntime/extranonce occasionally.  This was a much older cgminer version that had the problem though, and I think that user stopped submitting bad data after upgrading.

Notice to All Stratum Users

Going forward, miners with native Stratum clients are encouraged to connect via the following address:

stratum.btcguild.com:3333


Currently this entry points to a single Stratum instance.  All getwork X-Stratum redirects now point to this same DNS entry.  Going forward, this will be first updated to a Round Robin DNS entry, to help share the load over multiple servers.  Because of how Stratum works [persistent TCP connections], this will be seamless to an end-user.  By connecting with stratum.btcguild.com:3333, you avoid interacting with a getwork server to obtain the Stratum IP/port.

This is not a mandatory change.  Getwork servers will continue to function and provide X-Stratum redirects, this will simply reduce one extra connection needed to begin mining, and avoid one extra failure point when starting up.

2 Avalons have been RECEIVED and proven to be real.  Aside from those 2, we only have figures about how many have been shipped.  The two that have already been received were treated differently in order to get them into the hands of major figures in the community as soon as possible, at extremely high shipping costs.

Fresh update for the stratum changes.

Minimum difficulty code is going to be deployed on the beta Stratum server in the next hour.  Settings will not be available at first due to needing time to test it on a server with other users to make sure nothing funny happens with how new Stratum jobs are being created.

A few notes about the updated Stratum code:
1) Difficulty changes now take effect as soon as you pass a share-per-minute threshold.  These thresholds use different averaging windows to make sure you aren't adjusted to a higher difficulty as a result of lucky share submissions.  This is a change over the old code, where a difficulty change could only occur when a new job was prepared [so up to 30 seconds after you crossed an SPM threshold].

2) Minimum difficulty will be modified on the fly through the website interface.  Since Stratum difficulty is a per-connection basis [this is part of the official Stratum spec], be aware that if you connect many workers through one connection [single cgminer/proxy instance], your connection will be adjusted to the highest minimum difficulty of your authorized workers on that connection.


The settings to adjust your worker minimum difficulty may not be available for another day.  I am working on a smart load balancer to link with the Stratum servers.  I've been embedding new API calls in my Stratum codebase so a load balancer can talk to the servers.  Using the X-Stratum header redirects, I'm hoping to create a load balancing DNS entry that will always point you to the lowest load server when you establish a new stratum connection.

A fun read.  Always nice to follow a miner's experience through the different pools.  I just had one comment, and that was related to one of your statements:

"After about 20 Hours (sleep for 8 hours and work for 12 hours) I came home to 0.0244958 BTC. That's pretty decent for my hardware, but I feel ready for some more regular small payments."

PPS is as regular as you could get for payments [paid for every share!], the limit being you can't continually withdraw extremely small amounts to your own wallet, which fits your statement.  My only comment is you should be careful about pulling out significantly small amounts from pools as payments.  This leads to a problem dubbed "bag of pennies", where you are unable to send your BTC anywhere without incurring above average fees.

With the way the bitcoin protocol works, the age and size of your coin inputs is what determines the fee.  Being able to spend a 0.01 payout will take approximately 100 days to be able to send without a fee [according to default rules].  Multiple small payouts like this can induce even larger fees as the size of your transaction grows.  On top of this, small transactions with many inputs tend to take longer to confirm, so spending your coins might take longer.

Thank you for pointing to the thread .  I don't always get a check to look at all the different subforums to see where a problem was brought up.

The server running the website and one getwork pool is locking up with a hard drive failure.  Rebooting the host node now and looking at options for moving them.


UPDATE:  Website is back online.  One getwork pool is still offline, working on redirecting traffic to a different server temporarily.  The website will be a bit sluggish while the RAID resyncs.

UPDATE2:  I've double checked the scripts, and the only thing that should have been interrupted during the website outage were idle emails and automatic payouts, not counting the getwork server that was unusable.  Users who set another BTC Guild server as backup, used a different server, or used Stratum, should not have experienced any reduced performance/earnings.

UPDATE3:  Stratum updates mentioned yesterday will likely be going live in about 6 hours.  Just working on making minimum difficulty work properly without leaving it open to withholding shares for greater rewards when a new worker is authorized at a higher difficulty.

I suppose I should have expected it [and I did try to edit my post but the forum is so damn slow lately].  I just don't understand why so many users send funds from a source they don't control.

You are mistaken.  There is a FROM address in bitcoin [inputs into the transaction].  There might be multiple FROM addresses, but they are there.  


EDIT: The only problem I would see with it is if somebody did something foolish like try to withdraw coins from an exchange directly to the "Pay to" address.

I'm disappointed your definition of high electricity cost country is $0.25/kwh, considering I'm getting bent over with 0.32/kwh in California.