Received: from sog-mx-2.v43.ch3.sourceforge.com ([172.29.43.192] helo=mx.sourceforge.net) by sfs-ml-1.v29.ch3.sourceforge.com with esmtp (Exim 4.76) (envelope-from ) id 1SXvZQ-0007i4-Hn for bitcoin-development@lists.sourceforge.net; Fri, 25 May 2012 14:32:48 +0000 Received-SPF: pass (sog-mx-2.v43.ch3.sourceforge.com: domain of coinlab.com designates 209.85.213.47 as permitted sender) client-ip=209.85.213.47; envelope-from=peter@coinlab.com; helo=mail-yw0-f47.google.com; Received: from mail-yw0-f47.google.com ([209.85.213.47]) by sog-mx-2.v43.ch3.sourceforge.com with esmtps (TLSv1:RC4-MD5:128) (Exim 4.76) id 1SXvZK-0002RB-MW for bitcoin-development@lists.sourceforge.net; Fri, 25 May 2012 14:32:48 +0000 Received: by yhjj56 with SMTP id j56so736917yhj.34 for ; Fri, 25 May 2012 07:32:37 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type:x-gm-message-state; bh=AvAXbrwNTJbmPdlbDYoQ/T5iyCFWVivcQWwLlayZ3+0=; b=VNKnmul57bpT1X57QEeXXQWSJ9GA8H18XVFedDuY3H9MMlDX8HU20fHS37pVt5X0Gd HFbhbzMKIMgNN/InEHDtzEc64726zAXgfUW0xsbSq5nvzYaIUeRKgivi9Gy1w7GLTePO yQB/CA8Y46G4ia4hj/zO/ylTdysVEHr+4ae95wxuUHdgZKr/tNAwphBhtKu0zekLlnu6 bZbP6G+4ZIqpT5GZ3PRyDBDX2nCyxj2XM6K6UNdsbv67F5TuebNPKHF/tbBFkub/EJ1f K88xqSlnKQ9c02XEVM5VaqlvkbHNK3tLmnnB0uFuKmfRN3xs9M1ul57biN5B8RAtyOki 9spQ== Received: by 10.60.0.193 with SMTP id 1mr3208518oeg.39.1337954482316; Fri, 25 May 2012 07:01:22 -0700 (PDT) MIME-Version: 1.0 Received: by 10.182.41.200 with HTTP; Fri, 25 May 2012 07:00:09 -0700 (PDT) In-Reply-To: References: <201205250045.24540.luke@dashjr.org> <201205250057.39749.luke@dashjr.org> From: Peter Vessenes Date: Fri, 25 May 2012 10:00:09 -0400 Message-ID: To: Alan Reiner Content-Type: multipart/alternative; boundary=e89a8fb20116a4e79704c0dccd5b X-Gm-Message-State: ALoCoQmIC8atDMkBkzcUsdvkt/0TfEVOyUJhn84fPRddtJOKl43cKczfvtlp9kIEc/2CykkPwTSl X-Spam-Score: -0.5 (/) X-Spam-Report: Spam Filtering performed by mx.sourceforge.net. See http://spamassassin.org/tag/ for more details. -1.5 SPF_CHECK_PASS SPF reports sender host as permitted sender for sender-domain -0.0 SPF_PASS SPF: sender matches SPF record 1.0 HTML_MESSAGE BODY: HTML included in message X-Headers-End: 1SXvZK-0002RB-MW Cc: bitcoin-development@lists.sourceforge.net Subject: Re: [Bitcoin-development] Punishing empty blocks? X-BeenThere: bitcoin-development@lists.sourceforge.net X-Mailman-Version: 2.1.9 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 25 May 2012 14:32:48 -0000 --e89a8fb20116a4e79704c0dccd5b Content-Type: text/plain; charset=ISO-8859-1 We just implemented our own mining tool, soup-to-nuts, and I would say that the likely motivation for what I presume are botnet owners is just economic. It's a lot more work to make sure your merkleing and keeping up-to-date are happening than it is to just get an 80 byte header from a central server, and re-calc a single transaction merkle client-side. Not to mention the extra work to keep track of what version of getmemorypool output you're receiving work for in a broadly distributed pool. For what it's worth, we did this extra engineering work since we care about Bitcoin, but if I just wanted to pull value out of the ecosystem, we would have skipped it. The only solutions to this are economic solutions -- making such 'cheater' blocks less valuable, or increasing the value of the transactions. Also note that botnet operators likely care, in the end, about fiat currency, so going to 25 btc per block in what I think of as transaction fee subsidies won't necessarily impact this -- it's a matter of what happens to exchange rates vs generation rates that will matter. I think we also have to moderate this consideration against the rights (and arguable benefits) of someone wanting to build an express-delivery mining service, one that will provide provably faster certification for those adding a transaction fee of, say, 1 btc. My own experience now in the MMO world is that we have to carefully understand how we deal with griefers who control massive resources (compute or gold-farmers). It may not be a winning battle to choose a solution which harms the rest of the network in exchange for harming the griefers. This is definitely out of the box, but one solution might be to change the difficulty calculations to just ignore 1tx blocks; that would minimize impact on others to a great extent, and would let someone set up an express block service if they chose. I guess we'd have to settle on whether or not such blocks counted towards the issuance countdown as well. Or, we could allow only 1/10 generation fees on such blocks. Peter On Fri, May 25, 2012 at 9:44 AM, Alan Reiner wrote: > I like the concept except that it only works if every node connected to > the miner enforces the rule (if it works). Once any one of the nodes > forwards the block, other nodes see it coming from a node that can pass > the challenge. > > I don't think any solution based on node queries will succeed, especially > if it requires spontaneous super-majority-of-nodes acceptance. I think > it's gotta be based on the block itself and each nodes' own info. > > If you could spontaneously get all miners to agree not to build off of > anti-social blocks (however that is defined) , it would have a chance of > making a difference, but individual miners would have an advantage > building off the antisocial block because they only need to produce one to > create the longest chain (and collect reward) while the miners following > the rules need two blocks. > > --Sent from my overpriced smartphone > On May 25, 2012 3:48 AM, "Christian Decker" > wrote: > >> How about a simple proof of work test? This one though does not ask for >> CPU work but asks the miner for a random old transaction. If the miner >> really stores the entire blockchain he will not have any problem answering >> to that getdata request, whereas a botnet would have to ask someone else >> for it, which could be detected if the response time deviates too much from >> what has been previously measured (compare it against getdata for the block >> they advertise). It's not perfect but it allows an estimate of whether it >> is a chainless miner. >> >> Regards, >> Chris >> -- >> Christian Decker >> >> >> >> On Fri, May 25, 2012 at 3:17 AM, Jeff Garzik wrote: >> >>> On Thu, May 24, 2012 at 8:57 PM, Luke-Jr wrote: >>> > Block times are not accurate enough for that. >>> >>> The times in your log are very accurate, assuming your system clock is >>> remotely accurate. >>> >>> -- >>> Jeff Garzik >>> exMULTI, Inc. >>> jgarzik@exmulti.com >>> >>> >>> ------------------------------------------------------------------------------ >>> Live Security Virtual Conference >>> Exclusive live event will cover all the ways today's security and >>> threat landscape has changed and how IT managers can respond. Discussions >>> will include endpoint security, mobile security and the latest in malware >>> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ >>> _______________________________________________ >>> Bitcoin-development mailing list >>> Bitcoin-development@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/bitcoin-development >>> >> >> >> >> ------------------------------------------------------------------------------ >> Live Security Virtual Conference >> Exclusive live event will cover all the ways today's security and >> threat landscape has changed and how IT managers can respond. Discussions >> will include endpoint security, mobile security and the latest in malware >> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ >> _______________________________________________ >> Bitcoin-development mailing list >> Bitcoin-development@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/bitcoin-development >> >> > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development > > -- Peter J. Vessenes CEO, CoinLab M: 206.595.9839 Skype: vessenes Google+ --e89a8fb20116a4e79704c0dccd5b Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable We just implemented our own mining tool, soup-to-nuts, and I would say that= the likely motivation for what I presume are botnet owners is just economi= c.

It's a lot more work to make sure your merkleing = and keeping up-to-date are happening than it is to just get an 80 byte head= er from a central server, and re-calc a single transaction merkle client-si= de.

Not to mention the extra work to keep track of what ver= sion of getmemorypool output you're receiving work for in a broadly dis= tributed pool.=A0

For what it's worth, we did = this extra engineering work since we care about Bitcoin, but if I just want= ed to pull value out of the ecosystem, we would have skipped it.=A0

The only solutions to this are economic solutions -- ma= king such 'cheater' blocks less valuable, or increasing the value o= f the transactions.

Also note that botnet operator= s likely care, in the end, about fiat currency, so going to 25 btc per bloc= k in what I think of as transaction fee subsidies won't necessarily imp= act this -- it's a matter of what happens to exchange rates vs generati= on rates that will matter.

I think we also have to moderate this consideration aga= inst the rights (and arguable benefits) of someone wanting to build an expr= ess-delivery mining service, one that will provide provably faster certific= ation for those adding a transaction fee of, say, 1 btc.=A0

My own experience now in the MMO world is that we have = to carefully understand how we deal with griefers who control massive resou= rces (compute or gold-farmers). It may not be a winning battle to choose a = solution which harms the rest of the network in exchange for harming the gr= iefers.

This is definitely out of the box, but one solution mig= ht be to change the difficulty calculations to just ignore 1tx blocks; that= would minimize impact on others to a great extent, and would let someone s= et up an express block service if they chose. I guess we'd have to sett= le on whether or not such blocks counted towards the issuance countdown as = well. Or, we could allow only 1/10 generation fees on such blocks.

Peter


On Fri, May 25, 2012 at 9:44 AM, Alan Reiner <<= a href=3D"mailto:etotheipi@gmail.com" target=3D"_blank">etotheipi@gmail.com= > wrote:

I like the concept except that it only wo= rks if every node connected to the miner enforces the rule (if it works).= =A0 Once any one of the nodes forwards the block,=A0 other nodes see it com= ing from a node that can pass the challenge.

I don't think any solution based on node queries will succeed,=A0 es= pecially if it requires spontaneous super-majority-of-nodes acceptance.=A0 = I think it's gotta be based on the block itself and each nodes' own= info.

If you could spontaneously get all miners to agree not to build off of a= nti-social blocks (however that is defined) ,=A0 it would have a chance of = making a difference,=A0 but individual miners would have an advantage build= ing off the antisocial block because they only need to produce one to creat= e the longest chain (and collect reward) while the miners following the rul= es need two blocks.

--Sent from my overpriced smartphone

On May 25, 2012 3:48 AM, "Christian Decker&= quot; <d= ecker.christian@gmail.com> wrote:
How about a simple proof of work test? This one though does not ask for CPU= work but asks the miner for a random old transaction. If the miner really = stores the entire blockchain he will not have any problem answering to that= getdata request, whereas a botnet would have to ask someone else for it, w= hich could be detected if the response time deviates too much from what has= been previously measured (compare it against getdata for the block they ad= vertise). It's not perfect but it allows an estimate of whether it is a= chainless miner.

Regards,
Chris
--
Christian Decker


On Fri, May 25, 2012 at 3:17 AM, Jeff Ga= rzik <jgarzik@exmulti.com> wrote:
On Thu, May 24, 2012 at 8:57 PM, Luke-Jr <luke@dashjr.org> wrote:
> Block times are not accurate enough for that.

The times in your log are very accurate, assuming your system clock i= s
remotely accurate.

--
Jeff Garzik
exMULTI, Inc.
jgarzik@exmulti.co= m

-----------------------------------------------------------= -------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122= 263/
_______________________________________________
Bitcoin-development mailing list
Bitcoin-development@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bitcoin-de= velopment


-----------------------------------------------------------------------= -------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122= 263/
_______________________________________________
Bitcoin-development mailing list
Bitcoin-development@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bitcoin-de= velopment


-----------------------------------------------------------= -------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122= 263/
_______________________________________________
Bitcoin-development mailing list
Bitcoin-develo= pment@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bitcoin-de= velopment




--
Peter J. Vessenes
CEO, CoinLab
M: 206.595.9839
Skype: vessenes
= Google+

--e89a8fb20116a4e79704c0dccd5b--