<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Aparna Krishnan on Medium]]></title>
        <description><![CDATA[Stories by Aparna Krishnan on Medium]]></description>
        <link>https://medium.com/@aparnalocked?source=rss-d4f4e8e73570------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*DPrGJ7xkoqoaOdADImK3fw.jpeg</url>
            <title>Stories by Aparna Krishnan on Medium</title>
            <link>https://medium.com/@aparnalocked?source=rss-d4f4e8e73570------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 23:15:18 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@aparnalocked/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[All about the Op(sh)yn (Part 1)]]></title>
            <link>https://medium.com/@aparnalocked/all-about-the-op-sh-yn-part-1-bf26dfea13ee?source=rss-d4f4e8e73570------2</link>
            <guid isPermaLink="false">https://medium.com/p/bf26dfea13ee</guid>
            <dc:creator><![CDATA[Aparna Krishnan]]></dc:creator>
            <pubDate>Sat, 18 Apr 2020 01:22:00 GMT</pubDate>
            <atom:updated>2020-04-18T03:59:21.245Z</atom:updated>
            <content:encoded><![CDATA[<h3>What is an Option?</h3><p>Financial options give the option holder the right but not the obligation to do something. What is that something? Depends on the type of option ;)</p><p>There are two types of options:</p><ol><li><strong>Put Options</strong>: Put Options give the holder of the option “the right but not the obligation to sell their assets at a pre-specified price”. Example: I have a bunch of Tesla shares and I’m afraid they are going to drop in price since I <a href="https://www.youtube.com/watch?v=hOEWUm6aGPg">doubt Elon’s vibe</a>. I don’t want to sell all my Tesla shares since I work there and hope they might go up in price once Elon’s music skills improve. To protect myself against Tesla shares crashing, I can buy put options on Tesla shares so that I can limit my downside but keep my upside!</li><li><strong>Call Options: </strong>Call Options give the holder of the option “the right but not the obligation to buy assets at a pre-specified price.” Example: I want to buy <a href="https://unisocks.exchange/">unisocks</a> but I have too many socks right now, so I should probably wait for my laundry machine to put more holes in my socks before I buy the unisocks. But, what if unisocks go up in price since the entire DeFi squad wants to buy them? I can buy a call option on Unisocks so that I can still buy them at today’s price a month from now.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rwHR5AbGjQY1Y8sBMkb96Q.png" /></figure><p>Opyn currently supports put options. So, for the rest of this blog, let’s dive deeper into put options.</p><h3>Put Options: The buyer’s perspective</h3><p>There are several reasons to buy a put option. At a high level, it makes sense for someone to buy a put option if they are long term believers in the asset, but are afraid of short term price crashes and volatility in the asset. In the following section, we dive deeper into what the payoffs could look like for someone who chooses to buy a put option. Before understanding how the whole portfolio performs, we will focus on how the put options alone perform.</p><p>Through the examples below, we’ll illustrate the difference between an option that can only be exercised and an option that can be exercised or sold. We’ll see<strong> </strong>why <strong>an option that can be both exercised and sold is a more powerful way to hedge.</strong></p><h4>Example 1: Payoff for a buyer who can only exercise her option</h4><p>Let us consider an example where Alice is afraid of ETH prices falling and buys an option on ETH that gives her the right but not the obligation to sell her ETH for $100. The pre-specified price of $100 is also called the strike price.</p><p>The payoff for the buyer of this $100 Strike Put Option on ETH looks like the following graph.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/519/0*JC3WjFV82V2bKfyt.png" /></figure><p>The buyer of the put option pays a premium upfront to obtain the option, thus the option holder is initially net negative. Let us assume that the premium paid upfront was $10. The more the price of ETH falls, the more money you make from exercising your option.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l9Xq_YXm6SsNHKhm1G8hFw.png" /></figure><h4>Example 2: Payoff for a buyer who can both sell and exercise her option</h4><p>Let us consider the same example where Alice is afraid of ETH prices falling and buys an option on ETH that gives her the right but not the obligation to sell her ETH for $100. The pre-specified price of $100 is also called the strike price.</p><p>Payoff for the buyer of this $100 Strike Put Option on ETH is shown in the graph below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/519/0*dROYUummh1gegque.png" /></figure><p>The buyer of the put option pays a premium upfront to obtain the option, thus the option holder is initially net negative. Let us assume that the premium paid upfront was $10. As the price of ETH falls, the option holder can “sell their ETH for $100” (exercise the option) OR “sell their right to sell their ETH for $100.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*doY7d44VZ1UbzTlI3guD_g.png" /></figure><p>While the option sale prices in the above chart are hypothetical, they illustrate the key point that an option that can be exercised <strong>or</strong> sold is a better hedge than an option that can only be sold.</p><p>This is because, in some scenarios, it is more profitable to sell the put option than it is to exercise it. The put option rises in value as your holdings in the underlying asset falls, allowing you to sell it and reduce your downside risk in your overall portfolio. It is more common to sell put options early in event of a crash than to hold them to expiry in traditional finance.</p><p>Opyn Options on ETH work exactly above as in Example 2 where the holder can either 1. exercise the option or 2. sell the option.</p><p>What if I’m long ETH but just want to protect my downside? How does that work? Fear not, this is just part one ;) We covered the basics of what are puts and how they work in this blog post. In the next one, we will talk more about how protective put options can be used as insurance!</p><p>If you are looking to hedge yourself against ETH falling in price, try out what you learned at <a href="https://opyn.co/">opyn.co</a>! :) Also <a href="https://discordapp.com/invite/2NFdXaE">join our discord</a> to join the Op(sh)yn squad and learn more about Options!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bf26dfea13ee" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to create, design, and measure life]]></title>
            <link>https://medium.com/@aparnalocked/how-to-create-design-and-measure-life-a635fd5ff0d3?source=rss-d4f4e8e73570------2</link>
            <guid isPermaLink="false">https://medium.com/p/a635fd5ff0d3</guid>
            <category><![CDATA[personal-growth]]></category>
            <category><![CDATA[meditation]]></category>
            <category><![CDATA[entrepreneur]]></category>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[self-awareness]]></category>
            <dc:creator><![CDATA[Aparna Krishnan]]></dc:creator>
            <pubDate>Fri, 11 Jan 2019 00:44:40 GMT</pubDate>
            <atom:updated>2019-01-11T00:44:40.938Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oz_mZMe-uegzQ5aHOLZlLQ.jpeg" /><figcaption><a href="https://twitter.com/AlexisGauba?lang=en">Image by Alexis Guaba</a></figcaption></figure><p>Right from the time we are born, we are taught to live life sensibly, within well defined structures that have already been created for us. Concepts of success and failure are defined within these structures. While these structures are useful, questioning and reorganizing these structures enables people to create an environment that allows them to feel fulfilled. For me, fulfillment comes through creating optimal value for society in a way that I care about. What “optimal value” and “society” mean are both determined by my individual views. This framework helps me discover what definition of these terms leaves me the most fulfilled.</p><p>When I first took the Thiel Fellowship, I gave up structure and embraced chaos. This was liberating yet scary. My life changed from constantly scrambling to meet all my test and project deadlines at UC Berkeley to having no externally-imposed deadlines that I really had to meet. The change felt drastic.</p><p>From being overwhelmed by work with no time to think, I was suddenly overwhelmed by the silence with plenty of time to think and no clear work. While unstructured thinking time was initially liberating, it eventually became confusing since it was unclear what I should have been doing at any given point of time. To me, it felt like I could flip a coin and act arbitrarily. I could pack my schedule with meetings, I could sleep all day, or I could (finally) read all the books I’d been meaning to read.</p><h3>Realizing the importance of self created structures</h3><p>This lead to a <a href="https://en.wikipedia.org/wiki/The_Paradox_of_Choice">paradox of choice</a>. The more time I spent thinking about all the things I could do, the number of options available to me compounded, making the process of deciding even harder. This gave me the illusion that I could do everything I had ever wanted and yet, when I did try to do everything, I ended up doing nothing.</p><p>This was a powerful turning point in my life. I realized that I DID need to create structures in order to succeed. I started to realize the <a href="https://fs.blog/wp-content/uploads/2016/12/Will-Durant.pdf">distinction between chaos and liberty</a>. Chaos provided me an unrestricted ability to do anything, while liberty gave me the ability do what I needed in order to meet my goals. As Will Durant, the Pulitzer Prize winning historian, philosopher, and author illustrates, “In my youth I stressed freedom, and in my old age I stress order. I have made the great discovery that liberty is a product of order”. I realized that I didn’t want chaos, instead I wanted sufficient structures to provide me stability around everyday activities so that I could use the unstructured time productively. For example, creating routines around my eating, sleeping, exercise schedules sucked out time spent on making trivial decisions and provided me more mental energy that I could channel into creative research. My work is quite non repetitive and requires constantly making decisions and having a routine in other parts of my life helps me deal with the constant uncertainty and variance in the professional aspect of my life. Finding the right balance between unstructured and structured time helped me be creative while still ensuring I was making progress. I found that forcing functions were extremely helpful in ensuring I was always making progress. A common forcing function that I used with self learning was if I wanted to learn about a new topic, I would sign up to give a talk about the topic at a meetup or to an audience. This ensured I sufficiently mastered the topic in a timely manner. While forcing functions were helpful, I realized back to back calendar entries infringed on my ability to think and be creative. Instead, I found that having 2–3 clear goals that I want to achieve in a day was sufficient structuring that enabled me to be both creative and produce output.</p><h3>What does success mean?</h3><p>Before I talk about the framework that I use to find that perfect balance i.e. self-created structures that enable me to have liberty, it is essential to first understand what success even means.</p><p>To me, there is no objective definition of success. And yet, success or value creation isn’t subjective to just an individual either. I don’t think value has a<a href="https://meaningness.com/neither-objective-nor-subjective"> completely objective or subjective</a> meaning; rather it is relative, emergent property from our interactions with each other and the world around us. Defining success required an understanding of what was important to me, and what was important to the communities and people that I cared about. Thus, it became increasingly important to be aware of both myself and the desires of the section of society that I cared about.</p><p>For me, I feel most fulfilled when I am creating “value” for society. As the first step in the process to feeling fulfilled, I find it important to have an understanding of what success means. While it is possible to have a basket of multiple small mutually independent successes (I don’t believe there is one way to live life), I personally find it optimal to have an overarching thesis on the kind of value I want to create. I find that in periods of uncertainty, this overarching thesis maintains a direction and prevents me from ever feeling completely lost.</p><h3>Why is it important to understand the definition of success?</h3><p>Each of us have about <a href="https://www.euromoney.com/article/b12kjkqsmvdc8h/impact-investing-the-big-business-of-small-donors">80,000 hours</a> in our careers. Thus in order to optimize value creation, it is important to optimize both magnitude of work and direction of the work. Hard work in the wrong direction, or just hard work without direction doesn’t lead to value creation. Optimizing value creation is like vector physics; every vector has a direction, in addition to movement. While understanding myself requires some time and reflection and is a constant process, this time investment is hugely valuable in ensuring I am always moving in the right direction.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/556/1*VmvZ4bE-4phoVhZmSDBa2g.png" /></figure><p>A common trap that I’ve fallen into is when I don’t create sufficient routines for myself, I spend hours everyday deciding what to eat, what to wear, what to do when I feel bored and such. Sequences of small decisions can cause mental fatigue since each of these decisions requires effort and feels like work. This is an example of work done without magnitude or direction. Having to make decisions about several small situations which are unrelated (not in the same direction) uses up small magnitudes of energy per instance. The overall result of adding up the energy expended on making each small decision is a substantial amount which makes me feel like I did work which didn’t amount to anything substantial.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/655/1*g_LgONBz6sKPI47FPEWfiw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/538/1*tmNFeKt0aK-eVuwlucJTaw.png" /></figure><p>Another common trap that I’ve fallen into in the past when I’ve thought about “hard work being the key to success” is mindless repetition of an activity. Repetitive practice without awareness is insufficient to see drastic improvement at any activity.</p><p>While technically I could create anything in the world, narrowing down on what value I wanted to create was important in order to ensure I made some progress towards actually executing on it. Figuring out if I had found what impact I want to make or if I should keep searching for a better definition of value creation was a hard problem. I found it useful to think of the value that I wanted to create in terms of <a href="https://conceptually.org/concepts/explore-or-exploit/">explore vs exploit tradeoffs</a>. The <a href="https://medium.com/galleys/optimal-stopping-45c54da6d8d0">optimal stopping theorem</a> is a useful framework to ensure I had sufficiently explored before I started to exploit. I feel I have lived a life sufficiently rich in experiences for me to know which causes I care about currently. With time and other experiences, my goals will probably get more refined, but as of now, I see a general direction to proceed in. I also find that being passionate about something and going all out in it makes it easy to understand the whole world from that perspective. From there, curiosity can enables me to connect that one thing to everything else in the world.</p><h3>Framework for coming up with an initial definition of success for myself:</h3><p>Below are some questions that help me <em>refine my own perspective of success</em>:</p><h4>What value do I want to have created in the world in my lifetime? How will this impact the lives of other people?</h4><ol><li>What are causes that I care about the most currently? (Eg. I personally care a lot about privacy and creating a future where we can’t track human beings. The fact that <a href="https://www.businessinsider.com/china-social-credit-system-punishments-and-rewards-explained-2018-4#2-throttling-your-internet-speeds-2">there exists social ranking and tracking systems in China</a> is an extremely scary reality to me.)</li><li>Who are the people whose lives I want to impact the most?</li></ol><h4>What does meaning in life for me mean?</h4><ol><li>Personally?</li><li>Professionally?</li></ol><h3>Framework for executing on my initial definition of success:</h3><p>Once I had a better understanding of the long term value I wanted to create, it was about <em>executing </em>on that. I use the following questions to break down how I can ensure I align my actions with my longer term goal.</p><h4>What personal and professional growth do I want to have achieved in trying to create the value that I want 1 year from now?</h4><ol><li>How do these goals bring me closer to my longer term value creation goals?</li></ol><h4>How am I going to achieve these goals?</h4><ol><li><strong>Choosing forcing functions / accountability metrics</strong>: What forcing functions / accountability metrics do I need to create for myself to reach my professional goals? (Eg. for me I find that having motivated co-Founders who I do weekly check ins with creates a strong accountability metric. )</li><li><strong>Understanding impact of environment / culture / setting on flow state: </strong>What kind of environment / culture / setting helps me be productive? What factors contribute to me getting to flow state? (Eg. a room with sunlight and silence might be helpful factors) How can I create this environment for myself for long periods of time for 5–6 days a week?</li><li><strong>Understanding factors which accelerate getting into flow state quickly: </strong>What are things that I can externally control / change to improve my chances of getting into a flow state every morning? What are factors that internally contribute to me going into flow state? (Eg. For me I find that a slightly stressful atmosphere helps me get into flow state. Anything more than slightly stressful puts me into stressed and can’t work mode. Anything less than that makes me complacent. I try to set goals and accountability checks to bring me into this mental frame.)</li><li><strong>Understanding factors which break flow state: </strong>What disrupts my flow state? (Eg. I find emails, messages etc to be extremely distracting. I ensure I set a fixed time every day to batch and do these and when I am in flow state I remain in Do Not Disturb mode where only emergency contacts can reach me). How am I going to deal with these factors that break my flow state?</li><li><strong>Maintaining motivation:</strong> What structure in my life will help me remain constantly motivated for the next x years to keep creating this value in the world? Good health, friendships and relationships are important factors for me.</li></ol><h4>What am I currently doing to set up these structures and forcing functions for myself?</h4><ol><li>What are the situations where my actions are aligned with my goals for the next year?</li><li>What are the situations where my actions are not aligned with these long term goals?</li><li>What actions am I going to take over the next month to align these actions to meet my goals?</li></ol><h3>Framework to keep iterating on the definition of success while executing on it:</h3><p>While the initial definition of success is helpful to have, this definition isn’t a fixed target. In fact, the process of execution will bring up more information that might alter / refine the initial definition of success. In such cases, it is important to get good at iterating on the definition of success while executing on this definition.</p><p>I like to take some time out of my every day life to ensure I understand and refine my ideas of success. Below is a set of questions that helps me find and keep direction as I keep executing:</p><h4>Am I completely honest with myself?</h4><ol><li>What are places I am disillusioning myself but not willing to accept it?</li><li>Why am I disillusioning myself?</li></ol><h4>What are my thought patterns like?</h4><ol><li>What thoughts do I have most regularly? Where do they stem from and what do they lead to? I have been trying the Vipassana style of meditation where I categorize my thoughts. This helps me understanding thinking patterns and loops. This also helps me separate out mood from action. Feelings and emotions exist, but I try to observe and accept them and function despite them. I also find that understanding thinking patterns helps me recognize that I am going into a bad mood and catch myself before the thoughts spiral downward. <a href="https://juliacameronlive.com/basic-tools/morning-pages/">Morning Pages</a> has also been a great way for me to surface and deal with thoughts and emotions that I would otherwise bottle down.</li></ol><blockquote>In conclusion, this is a framework that helps me to understand myself better. I believe that while this is a useful starting point, each individual is unique and self reflection is the best way to discover frameworks that work best. Processing life is just as important as experiencing many things. This is why the <a href="https://www.businessinsider.com/malcolm-gladwell-explains-the-10000-hour-rule-2014-6">10,000 hour rule</a> (10,000 hours of practice makes one successful) doesn’t blindly apply. In order to maximize value creation, I find that introspection is just as important as acting and experiencing external stimulus. Introspection is crucial since it helps me figure out if I am any closer to the direction I want to move in, similar to a powerful inner compass.</blockquote><p><em>Thank you to </em><a href="https://twitter.com/AlexisGauba"><em>Alexis Gauba</em></a><em>, </em><a href="https://twitter.com/snarkyzk"><em>Zubin Koticha</em></a><em>, Logan Ford and </em><a href="http://allysondias.com/"><em>Allyson Dias</em></a><em> for their valuable feedback.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a635fd5ff0d3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cryptographic Adversaries in Classical and Nakamoto Consensus: Part 1, Overview]]></title>
            <link>https://medium.com/mechanism-labs/cryptographic-adversaries-in-classical-and-nakamoto-consensus-part-1-overview-474d364c0bbe?source=rss-d4f4e8e73570------2</link>
            <guid isPermaLink="false">https://medium.com/p/474d364c0bbe</guid>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[cryptography]]></category>
            <category><![CDATA[ethereum]]></category>
            <dc:creator><![CDATA[Aparna Krishnan]]></dc:creator>
            <pubDate>Sat, 17 Nov 2018 22:22:47 GMT</pubDate>
            <atom:updated>2018-11-17T22:22:47.122Z</atom:updated>
            <content:encoded><![CDATA[<h3>Cryptographic Adversaries in Blockchain Consensus: Part 1, Overview</h3><p>Imagine we lived in a world where <em>everyone </em>was honest. Wouldn’t that make all situations where large groups of people have to come to agreement significantly easier? Utopia is quite far from reality, which is why we have computer security techniques, cryptographic primitives, and more recently, blockchain protocols.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/894/0*jLPPGWRqC1_pzf1Q.jpg" /></figure><p>Human beings are intrinsically complex and so are their actions. Thus, creating protocols that deal with interactions between many of them is quite a challenging task. Each person is motivated by a unique set of factors and thus responds differently to stimuli. Figuring out the right set of assumptions that apply to the majority of human beings is thus a major step in designing an algorithm to come to an agreement on a common worldview i.e. a consensus protocol.</p><p>In consensus protocols, any action that is outside of the expected behavior of a node and is not protocol following is considered malicious. Large degrees of malicious activity could potentially break fundamental assumptions of the protocol and prevent participants in the protocol from ever coming to consensus. This results in either the consensus or blockchain protocol halting in the best case (sometimes called a <em>liveness</em> failure). In the worst case, malicious actors can cause false agreement on multiple inconsistent views or even cause a reversion of agreements that were already made (<em>safety</em> failures). Such situations need expensive external intervention to reboot the consensus algorithm. Thus, creating protocols with the right set of assumptions is critical in order to minimize expensive operations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*s1VZo5n6m3YkUSWNcAGUMg.png" /><figcaption>From left to right, increasingly strong adversaries.</figcaption></figure><p>The first set of consensus protocols that were developed tended to operate under simple adversarial assumptions. The goal was just for a predetermined set of nodes to come to an agreement by assuming an environment which had no malicious actors. This is <a href="https://www3.cs.stonybrook.edu/~prade/Teaching/F15-CSE535/prez/L18-L19-L20-Consensus.pdf">a trivial problem</a> and once such protocols were designed, the natural next step was to improve them. Protocols that could not handle any sort of faults or malicious actions were impractical to implement in certain settings, such as the internet. These protocols required all the participating nodes to be known ahead of time and for all of these nodes to remain online in addition to always acting honestly in order for the protocol to terminate correctly. This was impractical because if even one node went offline because of an <a href="https://medium.com/mechanism-labs/synchrony-and-timing-assumptions-in-consensus-algorithms-used-in-proof-of-stake-blockchains-5356fb253459">unstable communication channel</a>, it could prevent consensus.</p><p>Neither fiber optic cables nor the internet itself provides <a href="https://astro.temple.edu/~stafford/cis320f05/lecture/chap3/deluxe-content.html">reliable enough communication channels.</a> In the case of a power outage or a disconnection of a wire, the system would fail.</p><p>In order to build a system that was practically more beneficial in unstable environments, the system had to at minimum tolerate nodes going offline. It couldn’t assume that all nodes were completely honest. Thus came the earliest assumptions on what kinds of security breaches protocols can tolerate through the creation of malicious actors called <a href="http://web.cs.ucla.edu/~rafail/STUDENTS/katz-thesis.pdf">adversaries</a>.</p><h3>Types of Adversaries</h3><p>The first method of classifying adversaries was based on <em>the most powerful action that adversary could take</em>. The simplest adversaries of this kind were just nodes that would go offline permanently. This sort of protocol-deviating behavior on the part of a node was a called<em> </em><a href="https://dl.acm.org/citation.cfm?id=152612"><em>crash fault</em></a>. Nodes going offline happens when there is any sort of disruption or termination in the internet connection. To expect multiple nodes globally to stay online all the time over the internet is impractical because any power outage can cause a node to go offline. Nodes may also go out of sync for a few seconds if packets over the internet are dropped or delivered late. The internet itself is <a href="https://www.washingtonpost.com/sf/business/2015/05/30/net-of-insecurity-part-1/?noredirect=on">not sufficiently secure</a> and works on a “try my best” to deliver packets <a href="http://www.skullbox.net/tcpudp.php">approach</a>. Therefore, it is unreasonable to expect that all information will be exchanged in a timely manner and that all nodes will always remain online. As such, protocols that were crash fault tolerant played an important role in the development of consensus algorithms that could actually work on the internet.</p><p>The next iteration of protocols handled adversaries that <a href="http://mahdiz.com/crypto/basics/">could eavesdrop</a> on messages sent between users. Those adversaries could infringe on the privacy of the communicating parties by reading messages that weren’t meant to be shared with anyone else. These adversaries were called <em>passive adversaries</em>. The<a href="http://www.cs.umd.edu/~jkatz/gradcrypto2/NOTES/lecture25.pdf"> goal of this type of adversary</a> was to impersonate the honest participants after eavesdropping on multiple rounds of the protocol. These adversaries were a threat to protocols where confidential information was supposed to be transmitted. These adversaries were also a threat because a minority of dishonest actors were able to affect the value that the majority of honest actors agreed upon since the minority could impersonate other honest actors. These types of adversaries were studied in the 1970s and 1980s. Encryption and <a href="http://www.cs.umd.edu/~jkatz/gradcrypto2/NOTES/lecture25.pdf">Public Key Infrastructure</a> are techniques used to retain privacy, data integrity, and security in the Passive Adversarial Model.</p><p>Naturally, the next step was for protocols to handle adversaries that could use the information which they gathered to act arbitrarily. This could include sending false or conflicting messages. These adversaries were called <a href="http://mahdiz.com/crypto/basics/"><em>Byzantine or active</em></a><em> adversaries</em>. The term Byzantine originates from a landmark paper written in 1982 by Leslie Lamport called “<a href="https://web.archive.org/web/20170207104645/http://research.microsoft.com/en-us/um/people/lamport/pubs/byz.pdf">The Byzantine Generals problem”</a>. The paper describes the problem of consensus that the Byzantine army surrounding a city faces as they try to come to an agreement on if they should all attack or retreat. As they try to come to an agreement, they face issues with messages getting dropped, duplicated and altered. These messages are delivered out of order by an unreliable messenger and the dishonest generals try to confuse the rest of the army with contradictory messages. The goal of the dishonest generals is to ensure that the army loses the war by splitting up the efforts of the army having some of the generals attack while having the others retreat or not act at all. Despite all of these challenges, the goal of all honest generals is to act in one common way.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/607/0*ecYV114eh78gBT68" /><figcaption><strong><em>To attack or not to attack?</em></strong></figcaption></figure><p>Similarly, any public blockchain consensus protocol would have to be able to grow and remain consistent despite arbitrary attacks. Users of certain applications may not want some transactions to ever be recorded on the blockchain and might actively act maliciously and try to censor these transactions. Examples of these include front-running attacks, double spend attacks, and transaction censorship attacks to prevent the settlement of payment channels. These attacks are to be anticipated once there is a huge financial ecosystem relying on the blockchain and the protocol needs to be able to tolerate these attacks in order for it to be practically usable at a large scale. Thus, public blockchain consensus protocols have to be Byzantine fault tolerant.</p><p>The strength of an adversary, however, is more than just the most drastic action that the adversary can take. It also involves consideration of the time it takes for the adversary’s action to take effect or the time it takes for adversaries to corrupt different participants of the network.</p><p>Another way of classifying adversaries is based on <em>how long it takes an adversary to corrupt another node after they have decided who to corrupt. </em>If the adversary can instantaneously corrupt participants, the adversary is called “strongly adaptive”. In “static” adversarial models, the adversary is allowed to choose the participants that they want to corrupt at the start before the protocol execution starts. After that point, no changes can be made. Adversaries which take some time (longer than instantaneous but less than infinite time) to corrupt participants of the protocol are called “mildly adaptive”. The time it takes for an adversary to corrupt a participant of the protocol is related to the <a href="https://medium.com/@aparnalocked/randomness-in-blockchains-part-1-79192b173816">process by which validators and committee members are selected.</a></p><p>The strength of an adversary is also closely tied with the computational and storage limits placed on the adversary. If the adversary is assumed to have some fixed amount of resources, then we can assume a model which is computationally secure. The common assumption of most protocols is the existence of a polynomially bounded adversary. This means that an adversary cannot in a polynomial number of steps / time /space break any of the primitives of the protocol. This also means that the adversary <a href="https://security.googleblog.com/2017/02/announcing-first-sha1-collision.html">can with a very small probability break the security </a>of the protocol. This is often sufficient since it translates to an adversary taking over 50–60 years to break primitives with this level of security. On the other hand, if the adversary is assumed to be extremely rich and powerful, the protocol would have to be statistically secure. This means that even with infinite computing resources and a billion years, the adversary shouldn’t be able to break the protocol. Most cryptographic primitives used in practice are only computationally secure. This is sufficient security until <a href="https://www.technologyreview.com/s/610250/serious-quantum-computers-are-finally-here-what-are-we-going-to-do-with-them/">quantum computers become widespread</a>.</p><h3><strong>Economic Adversaries</strong></h3><p>So far, we considered assumptions that stem from computation and storage considerations. All of these assumptions focus actors who have decided they want to attack the system. All public blockchain consensus protocols are a lot more complex and need to often deal with more than adversaries who are sure they want to attack the system. For public blockchain protocols to sustain themselves in the long run, economic assumptions that relate to how actors respond to incentives are an important aspect to get right. These are important because they determine what incentivizes someone to act honestly and follow the protocol in the first place. There are three kinds of actors in any system. The first is economically rational, the second is altruistic and the third kind is irrational malicious. Majority of the blockchain protocols designed don’t align incentives for the economically rational actors to always follow protocol and this leads to the several attack vectors that might be unaccounted for at merely the protocol level.</p><p>In part 2 of this blog post, we will cover an analysis of some protocols that we looked at in our <a href="https://github.com/Mechanism-Labs/MetaAnalysis-of-Alternative-Consensus-Protocols/blob/master/MetaAnalysis.pdf">Meta-Analysis of Alternative Consensus Protocol</a>s based on adaptiveness of the adversaries that each of them can handle. For a more detailed description of why these are the adversarial models, read about <a href="https://medium.com/mechanism-labs/randomness-in-blockchains-part-1-79192b173816">the randomness generation schemes of each of these protocols</a>.</p><p>Having an understanding of adversarial models is crucial to designing secure systems. There is still a lot of work to be done with respect to modeling humans behavior and creating blockchain protocols. While we have come quite far, there are still several assumptions about the systems we create that if broken could render these protocols useless. As Vitalik says “<a href="https://medium.com/@VitalikButerin/why-cryptoeconomics-and-x-risk-researchers-should-listen-to-each-other-more-a2db72b3e86b">Cryptoeconomics is about agents with IQ 5 trying to control agents with IQ 150 </a>”. Human behavior is always hard to model, but with the emergence of fields like Cryptoeconomics, we get closer to capturing the multiple layers of human complexity in the systems we design. Even so, there still are <a href="https://emacbrough.wordpress.com/2018/10/24/sybil-attacking-byzantine-faults/">a lot of solutions</a> that are outside the field of Cryptoeconomics. A combination of developing new fields and re-using old concepts and ideas in clever ways is going to yield several new ways of modeling adversaries and designing more secure systems.</p><p>Twitter: @<a href="http://twitter.com/aparnalocked">aparnalocked</a></p><p>Acknowledgments: I would like to thank <a href="http://twitter.com/snarkyzk">Zubin Koticha</a>, <a href="http://twitter.com/alexisgauba">Alexis Gauba</a>, and <a href="http://twitter.com/sunnya97">Sunny Aggarwal</a> for their help in writing this post.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=474d364c0bbe" width="1" height="1" alt=""><hr><p><a href="https://medium.com/mechanism-labs/cryptographic-adversaries-in-classical-and-nakamoto-consensus-part-1-overview-474d364c0bbe">Cryptographic Adversaries in Classical and Nakamoto Consensus: Part 1, Overview</a> was originally published in <a href="https://medium.com/mechanism-labs">Mechanism Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Randomness in Blockchains (Part 1)]]></title>
            <link>https://medium.com/mechanism-labs/randomness-in-blockchains-part-1-79192b173816?source=rss-d4f4e8e73570------2</link>
            <guid isPermaLink="false">https://medium.com/p/79192b173816</guid>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[randomness]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[ethereum]]></category>
            <category><![CDATA[proof-of-stake]]></category>
            <dc:creator><![CDATA[Aparna Krishnan]]></dc:creator>
            <pubDate>Sat, 15 Sep 2018 01:59:22 GMT</pubDate>
            <atom:updated>2018-09-15T02:00:46.261Z</atom:updated>
            <content:encoded><![CDATA[<p>Randomness exists all around us. Concepts of luck, probability and fate have all historically been intertwined with the concept of randomness. Everything that humans don’t understand or can’t predict tends to be classified as such. We’re also <em>physically</em> engulfed in a sea of randomness. It exists everywhere — from the movement of clouds to behaviors in particles and waves.</p><p>And yet, for all its familiarity and abundance, humans have a hard time translating the randomness around us into something that computers can use. When we talk about randomness in a computer system, what we really mean is <em>pseudorandomness</em> i.e. a simulation of the randomness of the world, hard to distinguish from “true randomness”. Take, for example, this <a href="https://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator">very powerful</a> simulation of randomness.</p><p>Randomness plays a huge role in privacy technologies and cryptography. It is amazing how randomly generated value XORed with a message provides a simple yet <a href="https://en.wikipedia.org/wiki/One-time_pad">powerful encryption scheme</a>. Even the simplest form of private communication between two parties (i.e. symmetric key encryption where the parties share a key in advance) requires the shared key to be random. If this shared key isn’t random, then the encryption isn’t useful anymore because <a href="https://crypto.stackexchange.com/questions/52692/why-in-one-time-pad-must-the-key-distribution-to-be-truly-random">anyone can determine the key if they knew the function and decrypt the message</a>.</p><p>While randomness is important in the creation of secure channels of communication between two people, randomness can also play an important role in determining who is communicating with whom. If multiple people are trying to communicate with each other over a channel that has limited bandwidth, randomness can be used to determine a fair order in which the channel will carry the messages.</p><p>Randomness is also useful in helping a group of people or computers come to agreement, or consensus. <a href="http://disi.unitn.it/~montreso/ds/syllabus/papers/randomized-consensus-survey.pdf">Randomized consensus</a> protocols are such an example. This post will explore the role of randomness in blockchains.</p><p>Blockchains are a perfect example of multiple parties trying to come to agreement, to consensus, on a certain update to a global view. The updates are often done in periodic discrete time intervals called rounds.</p><p>One thing that constrains consensus is that there is an upper bound on the number of messages that can be sent within a certain period of time on the internet (i.e. the throughput) and the time it takes for a message to be sent across the internet (i.e latency). The goal of any blockchain consensus protocol is to come to agreement within the above constraints. In a public blockchain where thousands of nodes participate in maintaining the blockchain, the network constraints make it expensive to come to consensus if every node needs to send messages to all other nodes and waits for responses from all other nodes. This is because the number of messages sent through the network in the period of a round is too large. Thus the communication scheme needs to be optimized by cutting down on the number of messages sent through the internet in order to ensure consensus.</p><p>The way that Bitcoin achieves this agreement (in a process known as Nakamoto consensus) is by using <em>Proof of Work</em> (PoW) as the source of randomness to determine which one block per round will get added to the blockchain, thereby reducing the message-passing overhead. Because PoW is computationally difficult to solve, only the person who first solves the puzzle can add their block to the ledger. The probability that multiple people will solve the puzzle at the same time is low, thus PoW acts as a mechanism that limits the number of messages in the network.</p><p>Naturally, any alternative consensus mechanism trying to replace PoW would also need a way to limit the messages passing through the network. The way most <em>Proof of Stake </em>(PoS) protocols approach this problem is by selecting a subset of validators (nodes that maintain/govern the blockchain) to form a subcommittee based on how many coins they have staked. This subset of validators can then reasonably communicate back and forth with each other within the network constraints to achieve agreement in a timely manner.</p><p>Of course, in order to choose the members of this subset fairly and in such a way that no one knows the members ahead of time, the algorithm must be able to access some fair, unbiasable source of randomness. Once this subset of validators reaches agreement on the next block, this block is broadcasted to everyone else in the network.</p><p>An ideal source of randomness that can be used in PoS protocols to select a subcommittee would need to be unbiasable i.e no one can tamper with the randomness seed. In order<em> to be unbiasable</em>, the randomness protocol needs to ensure the following:</p><ul><li>i) there is always some output from the randomness function</li><li>ii) the output of the randomness function has not been manipulated (Note: we will explore the trade-off that emerges between i) and ii) and how it relates to network synchrony models in a future blog post)</li></ul><p>None of the subcommittee selection protocols that we at <a href="https://twitter.com/mechanism_labs?lang=en">Mechanism Labs</a> have <a href="https://docs.google.com/spreadsheets/d/1IW5AuFQtL4z34HgIZrqpcVfZPnOiifsc3xmA9KIef6I/edit#gid=658801415">analyzed</a> satisfied both properties of being unbiasable. Thus, given the above tradeoff, a blockchain consensus protocol should favor a randomness source that always produces some output rather than a randomness source that favors producing only a tamper-free output. This is because a blockchain protocol would need to ensure that the chain keeps growing and cannot have the source of randomness be the bottleneck. Even for consistency-favoring protocols, the randomness source should not be the reason that the blockchain doesn’t make progress. If anything, the randomness source should enable the protocol to focus on mitigating other attacks like DoSing of the validator subcommittee which halt progress.</p><p>If the blockchain completely halted due to bias in the randomness function output, there would be a huge cost of coordination at the social layer to reboot the blockchain. This would require the community to agree on what the actual blockchain is through an external social media platform and would result in long periods of discussion. This cost is comparable to the cost of resolving the DAO Hack. This huge cost at the social layer would also shake the community’s faith in the security of the blockchain protocol. On the other hand, as long as the bias is small and there is a mechanism to recover from the state of bias, the effect of the slight bias for a few rounds on the security of the blockchain is low. This is because any in protocol rewards given out to validators in a public blockchain protocol every round are relatively small. Since there are subcommittee selections every round or epoch (a set of rounds), there would have to be a significant bias in the randomness function for a validator to game the protocol to make money and reduce the security of the blockchain protocol. On the other hand, a lottery game utilizing randomness would absolutely need to ensure that the randomness source was not manipulated because even a little bias can completely change the winner of the lottery. This is because the effects of tampering with a lottery outcome are extreme as there is a large amount of immediate reward to gain. Such a game would thus favor a randomness function that guaranteed only a tamper-free output, even it meant that there were times the randomness would halt.</p><p>This post will talk about the importance of an unbiased source of randomness in the context of designing blockchain protocols. For the rest of this post, when we consider unbiasable protocols, we favor protocols that do not halt.</p><p>The <em>ideal committee selection scheme</em> should i) be unbiasable and ii) only reveal the randomness seed at the start of a new round. A randomness function needs to be unbiasable by the definition discussed above. If the randomness can be manipulated (and there is an in protocol reward distribution mechanism), then this can cause an unfair distribution of rewards. A skewed reward distribution will mean some people will get rewarded at a higher rate. Since only those with stake can be validators and since voting power is proportional to the stake that a validator owns, an unfair distribution will ensure that some subset of validators eventually dominates the blockchain. Thus, the degree to which a protocol is unbiasable determines <em>who can remain part of the network</em> as a validator in the long run. On the other hand, the amount of time that the seed is revealed before the start of a new round determines <em>who can be part of the network in the first place</em>. Thus, the degree of both the above properties will determine how decentralized the blockchain is.</p><p>Since the protocol for selecting a subcommittee is run every round or epoch, the time from when the output of this randomness function is known to the the time when the round or epoch starts is crucial. In this time frame, an adversary can use the output of the randomness function to determine what the subset of validators was chosen. If this output was hidden, an adversary trying to attack the blockchain protocol wouldn’t be able to determine the validator subset ahead of time. The length of this time frame determines the strength of the adversary that a protocol can tolerate. A stronger adversary, i.e. an adversary with more computational power and resources to attack the network, will be able to compute the validator subset in a shorter time frame. If the adversary is given sufficient time, the adversary will be able to determine the validator subset by running the committee selection algorithm with the randomness seed for the given round. This enables the adversary to conduct a denial of service (DoS) attack on the validator subset for that particular round or epoch, leading to empty blocks and a lack of progress for that time period. Halting of a blockchain even for one round can be devastating. Think of what would happen to Bitcoin if no one in the world could make any transactions for a few hours! Thus, only nodes which are capable of being DoS resistant should become validators in the first place. Indirectly, revealing the seed ahead of time will mean the consensus protocol tolerates a weaker adversary, thus making the blockchain less decentralized.</p><p>Below we take a closer technical look at various randomness generation mechanisms. The section below is based on the that the different blockchain protocols that we analyzed in our <a href="https://docs.google.com/spreadsheets/d/1IW5AuFQtL4z34HgIZrqpcVfZPnOiifsc3xmA9KIef6I">Meta-Analysis of Alternative Consensus Protocols</a> use:</p><h3>Tendermint:</h3><p>Tendermint uses a deterministic round-robin protocol scheme to choose proposers; there is no randomness in the protocol. Proposers are elected deterministically, through a heap sort based on the voting power and number of times the validator was chosen. The protocol can only be biased by adding or removing stake since that is the only input the adversary can access. The protocol cannot be biased instantaneously because of the long time it takes for a validator to <em>unbond</em> i.e remove stake from the system or <em>bond </em>i.e add stake to the system. Nevertheless, an attacker can plan ahead to bias the choice of proposer in a longer time frame.</p><p>Use of a deterministic random algorithm means the randomness seed is revealed well ahead of every round and the proposer can be determined ahead of time.</p><h3>Algorand:</h3><p>The randomness scheme in Algorand is as below: each validator v selected as a proposer computes a seed for round r using the equation &lt; seedr, π &gt; ← VRFskv (seedr−1||r) where skv is the secret key of validator v and seedr−1 is the randomness seed of the previous round.</p><p>The VRF proof π and the seed for round r are contained in the block accepted in round r − 1. If the given VRF proof doesn’t verify the given seed, then everyone computes the seed for the new round as H(seedr−1||r) where H is a hash function. This means that the secret key of each validator has to be chosen well in advance to ensure that they can’t bias the seed.</p><p>When the network is not strongly synchronous, the attacker has complete control over the message passing links, and can therefore drop block proposals and force users to agree on empty blocks, such that future selection seeds can be computed. Algorand thus requires the<em> weak synchrony</em> assumption that in every period of length u, there must be a strongly synchronous period s of length s &lt; u for the protocol to be bias resistant. As long as s is suitably large enough that at least one honest block was produced in a time period of b, an adversary v′ choosing a key skv′ cannot bias the randomness seed for round r.</p><p>Only once a block has been proposed is the new seed and a VRF proof to verify it publicly known. This ensures that the proposer and the randomness seed is not leaked ahead of time. This guarantee ensures that Algorand can defend against DoS attacks on proposers and is adaptively secure in a model with erasures, even with instantaneous corruption.</p><h3>Dfinity:</h3><p>In several protocols adversaries usually abort the protocol to invoke the fall back mechanism, and thus introduce bias. The threshold scheme that Dfinity uses is bias resistant because the threshold is chosen in a way that ensures an adversary cannot abort the protocol by preventing a <em>threshold signature</em> from which the randomness seed is derived from being created. This requires t the threshold to be chosen according to the equation: t ∈ [f + 1, n − f ] where f is the number of signatures the adversary controls, n is the total number of signatures in the scheme, and t is the threshold of signatures required to generate randomness. This threshold is chosen to ensure that an adversary cannot predict the outcome of creation of a signature nor can the adversary prevent its creation.</p><p>One caveat is that if an adversary has more than 50% of all deposits in any BLS scheme they would be able to manipulate the final signature and randomness. However if an adversary had such a huge portion of stake there are other adversarial attacks and this breaks the fundamental assumptions of the Dfinity protocol.</p><p>The randomness seed is revealed once any honest validator has advanced to round r. While the time gap between an honest validator advancing and the new round officially starting is small, this gap is sufficient for an adversary with significant computational resources to identify and DoS proposers. This is the reason Dfinity can only tolerate mildly adaptive adversaries and not instantaneous corruptions.</p><h3>Thunderella:</h3><p>In the Random Oracle scheme instantiated with a Hash Function, a proposer is determined by the equation: H_nonce(pk,q) &lt; Dp where H is the Hash function which is used as the Random Oracle, pk is the public key of the validator, q is the given time-step and nonce is the source of entropy to the hash function. The nonce is chosen by the proposer of the previous block. The difficulty parameter D_p is set such that a committee member is elected as the proposer with probability w in a single time step.</p><p>If the adversary proposes a block, she can bias the nonce used to generate entropy for the hash function of the next round, thus biasing the proposer elected in the next block. However, to mitigate the biasability of the randomness scheme, the same nonce is used in the Hash function to select proposers for r rounds rather than to merely select the next proposer. This makes it computationally harder for an adversary to brute force a nonce that will enable them to be the proposer for the next r rounds. While this strategy guarantees only a polynomial security loss, it comes with a predictability trade-off discussed later. This scheme results in being able to tolerate only a slow adaptive adversary</p><p>In the above algorithm, when the same nonce is reused to seed the Hash function it leads to an adversary being able to predict the proposers ahead of time. Since the same nonce is reused as entropy during an epoch, it leads to the randomness seed being leaked ahead of the start of a new round. This enables an adversary to corrupt or DoS proposers.</p><h3>Casper FFG:</h3><p>The randao scheme that Casper FFG plans to use is dependent on the algorithm below:</p><p>Every validator at the start of an epoch commits H (H (H (..Sv )))) where S is the seed the validator commits to. R := R⊕ Pre-image of the inner layer of hash. At a round, a validator can either make a block or abort.</p><p>If in the fall back mechanism no randomness is generated then that appears to be a bigger problem than the randomness being predictable or biasable because then you no longer need a 1/3rd malicious to abort liveness, 1 person is sufficient.</p><p>To the validator who is the current proposer the seed for the next round is known.</p><p>Randomness is a crucial part of Cryptography and Blockchains. Bad randomness schemes can break all notions of security in blockchains by i) halting the blockchain protocol ii) causing centralization. Thus, it is important to understand the role that randomness plays in blockchain protocols in order to understand how secure a blockchain truly is!</p><p><em>Thank you to </em><a href="https://twitter.com/AlexisGauba"><em>Alexis Gauba</em></a><em>, </em><a href="https://twitter.com/snarkyzk"><em>Zubin Koticha</em></a><em> and </em><a href="https://twitter.com/soonaorlater"><em>Soona Amhaz</em></a><em> for their valuable feedback. This post was originally</em><a href="https://www.tokendaily.co/blog/randomness-in-blockchains"><em> featured on Token Daily</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=79192b173816" width="1" height="1" alt=""><hr><p><a href="https://medium.com/mechanism-labs/randomness-in-blockchains-part-1-79192b173816">Randomness in Blockchains (Part 1)</a> was originally published in <a href="https://medium.com/mechanism-labs">Mechanism Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Blockchain Stack]]></title>
            <link>https://medium.com/mechanism-labs/the-blockchain-stack-ca953c088692?source=rss-d4f4e8e73570------2</link>
            <guid isPermaLink="false">https://medium.com/p/ca953c088692</guid>
            <category><![CDATA[privacy]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[scalability]]></category>
            <category><![CDATA[incentives]]></category>
            <category><![CDATA[blockchain]]></category>
            <dc:creator><![CDATA[Aparna Krishnan]]></dc:creator>
            <pubDate>Fri, 06 Jul 2018 16:27:44 GMT</pubDate>
            <atom:updated>2018-07-19T17:07:49.620Z</atom:updated>
            <content:encoded><![CDATA[<h4>Co-authored by <a href="https://twitter.com/AlexisGauba?lang=en">Alexis Gauba</a></h4><p>At a fundamental level, blockchains are composed of multiple distinct layers, similar to other technology protocols like the internet paradigm (Link, Network, Internet, Transport, Application). Here, we present a framework of the layers that compose blockchains. The layers are defined such that each layer depends on the one(s) below it. Here, we discuss what each layer <em>provides </em>as opposed to how each layer is <em>implemented</em>.</p><p>The following are specific parameters that these base layers support to allow for properties including privacy, scalability, interoperability, incentivization, governance, and energy efficiency.</p><p>To best achieve any of these properties (privacy, scalability, interoperability, incentivization, governance, energy efficiency), it is importance to optimize across all layers (hardware, network, consensus, and sybil control).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/866/1*k-otyNoqGsAd4I1Y4kX57g.jpeg" /></figure><h3><strong>Layers of the Blockchain Paradigm</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/158/1*xIV8YEr4y857bBHgVxnfhg.png" /></figure><p><strong>Hardware / Hosting: </strong>Layer of hardware used for mining / validating. This refers to the devices where the all data in the blockchain is stored</p><p><strong>Network:</strong> Layer of message passing between the different nodes that make up the network</p><p><strong>Consensus Protocol:</strong> Layer of finalization</p><p><strong>Sybil Control Mechanism: </strong>Layer to prevent generation of multiple fake identities</p><p><strong>Application Layer:</strong> Layer of applications built on top of blockchain platforms</p><h3><strong>Understanding the structure of the Bitcoin Blockchain through the above framework:</strong></h3><p><strong>Hardware / Hosting: </strong>The miners who make the most money in the Bitcoin network use ASICs as hardware to mine. The full nodes store all the data of the Bitcoin blockchain. Light nodes are a storage optimization on full nodes. Nodes talk to each other using the network layer.</p><p><strong>Network Topology: </strong>The Bitcoin gossip protocol relies on the speed of the internet across the world at the network level. Bitcoin’s gossip protocol works in a way such that every node informs all of its peers about the hash of any new object (block or transaction) that it receives. Each peer checks to see if it has received the object before by searching for the hash. If any node has not heard of the object before, the node makes a request for the object from the peer that gossiped the unfindable hash to it.</p><p><strong>Consensus Protocol: </strong>Bitcoin reaches finality using the longest chain rule i.e the chain with the most blocks gets added to the blockchain. In Bitcoin miners are incentivized to create a block on the longest chain in order to get block rewards if their block gets accepted.</p><p><strong>Sybil Control Mechanism: </strong>Bitcoin uses Proof-of-Work as a sybil control mechanism making it computationally expensive for anyone to create multiple identities. In Bitcoin people are incentivized to perform Proof-of-Work in the process of mining because of the possibility of receiving a block reward.</p><p>Currently blockchain infrastructure is not yet at the stage to support large scale applications, and while there have been waves of focus on different parts of the stack, blockchains won’t be ready until we can optimize across all layers.</p><p><em>This post was originally inspired by Emin Gun Sirer’s </em><a href="https://twitter.com/el33th4xor/status/1006931658338177024"><em>tweets</em></a><em>. Thanks to </em><a href="https://twitter.com/el33th4xor?lang=en"><em>Emin Gun Sirer</em></a><em>, </em><a href="https://twitter.com/soumyab8?lang=en"><em>Soumya Basu</em></a><em>, and </em><a href="https://twitter.com/snarkyzk?lang=en"><em>Zubin Koticha</em></a><em> for feedback.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ca953c088692" width="1" height="1" alt=""><hr><p><a href="https://medium.com/mechanism-labs/the-blockchain-stack-ca953c088692">The Blockchain Stack</a> was originally published in <a href="https://medium.com/mechanism-labs">Mechanism Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>