The author of a paper that describes a new attack on the Bitcoin protocol says that criticisms of the paper are misguided and that there are serious problems with Bitcoin that need to be addressed.

Ermin Gun Sirer, a professor at Cornell, published the paper earlier this week along with his co-author Ittay Eyal, and in it the researchers describe an attack in which a pool of Bitcoin miners would mine their own blocks and keep them secret and only publish their chain when it’s longer than the public one, which make theirs the authoritative one. This so-called “selfish” mining situation would then theoretically snowball and attract more miners to the pool in the hopes of gaining more rewards.

Critics have said that the paper relies on a flawed assumption that miners would act in the best interests of this new “selfish mining” group rather than in their own best interests, which could lead to them bouncing back and forth between the selfish mining group and the main pool of miners. Others, including the lead developer of Bitcoin, downplayed the attack, saying that it’s not a major issue in the real world.

“A very common response to security flaws is for the stalwarts to minimize the problem. The comments I’ve seen from reasonable people acknowledge that this is a serious issue, but that an attack will likely not be successful overnight, and therefore there will be time, on human-scales, to respond to the problem. We hope, and suspect, that this is true, as we want Bitcoin to be a viable currency,” Sirer said via email.

“But it would be a folly to ignore the problem, or to rely on manual intervention to deal with an automated, mechanical attack. Distributed systems are robust to the extent that their strengths and weaknesses are well-characterized. And Bitcoin’s strengths were overstated until our findings.

“And currency systems are robust to the extent that they provide the right monetary incentives to participants. At the moment, Bitcoin seems to be in a territory where it’s relying on the kindness of strangers.”

The idea of a group of miners gaining enough power to take over the Bitcoin system isn’t a new one, but the attack proposed by Sirer and Eyal relies on a much smaller pool of miners to accomplish the task, one-third of all users. If such an attack happened, or was ongoing now for that matter, trying to detect who these miners were might become a priority. Sirer said it’s unclear whether that’s possible right now.

“Selfish miners may have a network signature where they seem to be offering blocks concurrently with others. And if membership in the selfish pool is open to all, they may be infiltrated. But for every measure, there are countermeasures. For instance, the selfish miners could hide behind throwaway addresses, they limit the amount of information they reveal to participants, etc. Identifying selfish miners and excising them from the network is unlikely to be easy,” Sirer said.

Categories: Cryptography, Vulnerabilities

Comments (3)

  1. oda.krell
    1

    Please read up on the details of a technical discussion before publishing half-truths about said discussion.

    Your article presents the “defending” side as saying that the attack described by Sirer is not a risk because of a “flawed assumption” about miners’ behavior.

    That’s a strawman. Maybe someone, somewhere made that (admittedly stupid) argument, but in reality, the response by Bitcoin developers was a much more technical one, essentially saying that: a) the attack described by Sirer is a special case of an already well-known attack, and b) that this effect is not at all easy to pull off in reality because of certain characteristics of the Bitcoin network.

    A concise version of the *actual* answer to/refutation of the Sirer paper can for example be found here: https://bitcointalk.org/index.php?topic=324413.msg3476697#msg3476697

  2. SnowDog
    2

    The attack is a fallacy. It assumes that people are irrational and will seek a small temporary gain at the price of a huge loss with network instability, if the attack is carried out.

  3. R. U. Sirius
    3

    The researcher’s bold statement that “Bitcoin’s strengths were overstated until our findings” is indicative of attention-seeking academic sensationalism. I say this because any researcher who has been studying Bitcoin is well aware of many potential or real weaknesses that have been and are being discussed and analyzed in depth. The theoretical attack described here is based on a combination of two strategies that have been considered by many others in the community as early as 2010, namely the sybil attack and the mining cartel attack. It is certainly worth further analysis, including an independant peer review and detailed modeling. However, calling Bitcoin “fundamentally broken” based on not-yet-peer-reviewed, single study, and claiming that “Bitcoin’s strengths were overstated until our findings” is inappropriate to say the least.

Comments are closed.