https://homeostasis.scs.carleton.ca/wiki/index.php?action=history&feed=atom&title=DistOS_2021F_2021-11-23 2026-05-12T23:28:38Z Revision history for this page on the wiki MediaWiki 1.42.1 https://homeostasis.scs.carleton.ca/wiki/index.php?title=DistOS_2021F_2021-11-23&diff=23534&oldid=prev 2021-11-24T01:56:22Z

<p>Created page with &quot;==Notes== &lt;pre&gt; Lecture 19 ---------- - experience 2 is going up shortly - all set, william just has to post - will be due last day of class - but will be accept...&quot;</p> <p><b>New page</b></p><div>==Notes==<br /> <br /> &lt;pre&gt;<br /> Lecture 19<br /> ----------<br /> - experience 2 is going up shortly<br /> - all set, william just has to post<br /> - will be due last day of class<br /> - but will be accepted until a few days before<br /> the final exam<br /> <br /> - Nov. 30: Ethereum (will come out tonight)<br /> - Dec. 2: cryptocurrency critiques, futures<br /> - Dec. 7: present your projects<br /> - looking for volunteers<br /> - ideally 5 minute presentation<br /> <br /> I will ask William about solutions to experience 1 &amp; grading<br /> <br /> <br /> Byzantine Generals problem<br /> - really, the problem of consensus in the presence of bad actors<br /> - so some may send false information, how do we deal<br /> - basically, we don&#039;t trust all the &quot;generals&quot; (CPUs),<br /> so how can we achieve consistency?<br /> <br /> <br /> Key assumptions<br /> - assume majority are honest (follow the protocol)<br /> (necessary, not sufficient)<br /> <br /> To have a solution, you have a way to achieve consensus in the face of failure/malicious action (depending on the solution)<br /> <br /> We&#039;ve seen this often<br /> - node failures are common in distributed systems,<br /> and they can fail in arbitrary ways (not just crashing)<br /> - whenever we see a consensus solution, we&#039;re seeing<br /> a kind of solution to the byzantine general&#039;s problem<br /> - almost<br /> <br /> In the paxos-type solutions, nodes can fail or behave a bit badly, but they aren&#039;t going to flood the network or pretend<br /> to be other nodes<br /> - faulty, not malicious<br /> <br /> Note we can&#039;t deal with faked messages<br /> - no cryptographic assurances on node identity<br /> <br /> Bitcoin, what&#039;s the goal?<br /> - &quot;digital cash&quot;<br /> <br /> Value is not inherent in any of this<br /> <br /> What is inherent is it is a distributed ledger<br /> - record of transactions<br /> - no trusted third party<br /> - for updates or checks<br /> <br /> Is it for privacy? How private is bitcoin?<br /> - normally, in cryptography, to get privacy you want<br /> perfect forward secrecy<br /> - confidentiality is maintained even if keys are broken<br /> <br /> If someone saves all your communication and later breaks in<br /> and steals your private key, then do they get all your past<br /> communication?<br /> - with forward secrecy, NO<br /> - with bitcoin, YES<br /> <br /> The key to forward secrecy is using ephemeral keys<br /> - so the key the attacker wants isn&#039;t on the system long term<br /> - you can&#039;t compromise what you don&#039;t have<br /> <br /> As long as the algorithms remain secure and you dispose of<br /> ephemeral keys as you should, perfect forward secrecy works<br /> <br /> But Bitcoin can&#039;t do it<br /> - public keys are in the permanent record<br /> - and people need them to keep ownership of their money<br /> <br /> You have pseudonyms rather than anonymity<br /> - and the pseudonyms can be compromised anytime<br /> <br /> Advantage is who controls the pseudonyms<br /> - the participant, not a central third party<br /> - but, will need to reveal it when connecting Bitcoin to the real world<br /> <br /> Exchanges are the worst<br /> - they know who you are!<br /> - they can take your stuff!<br /> <br /> You have to trust the exchange<br /> - but bitcoin assumes that you don&#039;t trust anyone!<br /> <br /> But is bitcoin really &quot;no trust&quot;?<br /> NO<br /> - depends on the code being correct<br /> - depends on the cryptography being correct<br /> - NO PROOFS HERE<br /> - and depends on 50%+ being honest<br /> - of computational power, not people<br /> <br /> Have there ever been problems with hash functions?<br /> - before SHA-256 there was SHA-1 and then MD5, MD4, MD2...<br /> <br /> What would happen if someone came up with a way to generate<br /> collisions for SHA-256 quickly?<br /> - you could double spend potentially, breaking<br /> everything<br /> <br /> I&#039;m not saying that decentralized finance is inherently bad<br /> - lots of potential benefits!<br /> - but you have to understand the assumptions<br /> <br /> no such thing as zero trust<br /> - something is ALWAYS trusted<br /> - at most, can have checks and balances<br /> <br /> <br /> idea of bitcoin is trust the math/code, not people<br /> - but what happens when bugs are found in the math<br /> or in the code?<br /> - you have to then trust people to update the system<br /> - are majorities inherently more trustworthy than<br /> a centralized authority?<br /> &lt;/pre&gt;</div>

Soma