arnaucube
arnaucube/arnaucube.comPublic Self-hosted repository
arnaucube.com/blog/kzg-batch-proof.html
blog/kzg-batch-proof.htmlRaw
  1<!DOCTYPE html>
  2<html lang="en">
  3
  4  <head>
  5    <meta name="description" content="The benefit of batch proof is that allows us to proof multiple points while the proof size remains constant to one G1 point." />
  6    <meta charset="utf-8">
  7    <title> Batch proof in KZG Commitments - arnaucube - blog</title>
  8    <meta name="title" content=" Batch proof in KZG Commitments - arnaucube - blog">
  9    <meta name="description" content="The benefit of batch proof is that allows us to proof multiple points while the proof size remains constant to one G1 point.">
 10
 11    <meta property="og:title" content=" Batch proof in KZG Commitments - arnaucube - blog" />
 12    <meta property="og:description" content="The benefit of batch proof is that allows us to proof multiple points while the proof size remains constant to one G1 point." />
 13    <meta property="og:url" content="https://arnaucube.com/blog/kzg-batch-proof.html" />
 14    <meta property="og:type" content="article" />
 15    <meta property="og:image" content="https://arnaucube.com/blog/" />
 16    <meta name="twitter:title" content=" Batch proof in KZG Commitments - arnaucube - blog">
 17    <meta name="twitter:description" content="The benefit of batch proof is that allows us to proof multiple points while the proof size remains constant to one G1 point.">
 18    <meta name="twitter:image" content="https://arnaucube.com/blog/">
 19    <meta name="twitter:card" content="summary_large_image">
 20    <meta name="author" content="arnaucube">
 21
 22    <link rel="icon" type="image/png" href="img/logoArnauCubeFavicon.png">
 23
 24    <meta name="viewport" content="width=device-width, initial-scale=1">
 25
 26    <link href="css/bootstrap.min.css" rel="stylesheet">
 27    <link rel="stylesheet" href="css/style.css">
 28
 29    <!-- highlightjs -->
 30    <!-- <link rel="stylesheet" href="js/highlightjs/atom-one-dark.css"> -->
 31    <link rel="stylesheet" href="js/highlightjs/atom-one-light.css">
 32    <!-- <link rel="stylesheet" href="js/highlightjs/gruvbox-dark.css"> -->
 33    <script src="js/highlightjs/highlight.pack.js"></script>
 34
 35    <!-- katex -->
 36    <link rel="stylesheet" href="js/katex/katex.min.css">
 37  </head>
 38
 39  <body>
 40
 41    <!-- o_gradient_background" -->
 42    <nav id="mainNav" class="navbar navbar-default navbar-fixed-top"
 43                      style="height:50px;font-size:100%;">
 44      <div class="container">
 45        <div style="float:left;">
 46        <a href="/blog" style="color:#000;display:inline-block;">blog</a>
 47        <span style="margin-right:20px; margin-left:20px;">|</span>
 48        <a href="/blog/notes.html" style="font-size:90%;color:#000;display:inline-block;">notes</a>
 49        </div>
 50        <div style="float:right;">
 51          <a href="/" style="color:#000;display:inline-block;margin-right:10px;">arnaucube.com</a>
 52          <div class="onoffswitch" style="margin-left:0px;display:inline-block;" title="change theme">
 53            <input onclick="switchTheme()" type="checkbox" name="onoffswitch" class="onoffswitch-checkbox"
 54              id="themeSwitcher">
 55            <label class="onoffswitch-label" for="themeSwitcher"></label>
 56          </div>
 57        </div>
 58      </div>
 59      <!-- <img style="height:5px; width:100%; margin-top:8px;" src="img/gradient-line.jpg" /> -->
 60    </nav>
 61
 62
 63
 64    <div class="container" style="margin-top:0px;max-width:800px;">
 65      <h2>Batch proof in KZG Commitments</h2>
 66
 67<p><em>2021-08-14</em></p>
 68
 69<blockquote>
 70<p><strong>Warning</strong>: I want to state clearly that I&rsquo;m not a mathematician, I&rsquo;m just an amateur on math studying in my free time, and this article is just an attempt to try to sort the notes that I took while reading about the KZG Commitments.</p>
 71</blockquote>
 72
 73<p>Last week I posted some <em><a href="https://arnaucube.com/blog/kzg-commitments.html">notes on KZG Commitmens</a></em>, which overviews the scheme for single proofs. The current notes, try to overview the <em>batch proof</em> iteraton on KZG Commitments, and the <em>vector commitment</em> usage of it. Again (as in the previous post), big thanks to <a href="https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html">Dankrad Feist</a> and <a href="https://alinush.github.io/2020/05/06/kzg-polynomial-commitments.html">Alin Tomescu</a> for their articles which helped me to follow a bit the reasoing behind this, and again, I recommend spending the time <strong>reading their articles (<a href="https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html">1</a>, <a href="https://alinush.github.io/2020/05/06/kzg-polynomial-commitments.html">2</a>) instead of this current notes</strong>.</p>
 74
 75<h4>Batch proof</h4>
 76
 77<p>The benefit of <em>batch proof</em> is that allows us to proof multiple points while the proof size remains constant to one <span class="math inline">\(\mathbb{G}_1\)</span> point.
 78Let <span class="math inline">\((z_0, y_0), (z_1, y_1), ..., (z_k, y_k)\)</span> be the points that we want to proof, and that we have a polynomial <span class="math inline">\(p(x)\)</span> that goes through the points.
 79The <em>commitment</em> to the polynomial stands the same than for single proofs: <span class="math inline">\(c=[p(\tau)]_1\)</span>.</p>
 80
 81<p>For the evaluation proof, while in the single proofs we compute <span class="math inline">\(q(x) = \frac{p(x)-y}{x-z}\)</span>, we will replace <span class="math inline">\(y\)</span> and <span class="math inline">\(x-z\)</span> by the following two polynomials.
 82The constant <span class="math inline">\(y\)</span> is replaced by a polynomial that has roots at all the points that we want to prove. This is achieved by computing the <a href="shamir-secret-sharing.html#lagrange-polynomial%20interpolation">Lagrange interpolation</a> for the given set of points:</p>
 83<p><span class="math display">\[
 84I(x) = \sum_{j=0}^k y_j l_j(x)\newline
 85where \space\space\space l_j(x) = \prod_{0\leq m \leq k} \frac{x-x_m}{x_j - x_m}
 86\]</span></p><p>And the <span class="math inline">\(x-z\)</span>, which was to ensure that <span class="math inline">\(q(x)\)</span> had a root at <span class="math inline">\(z\)</span>, now, as we want to ensure that <span class="math inline">\(q(x)\)</span> has roots at all the points of the commitment, we will use the <em>zero polynomial</em>:</p>
 87<p><span class="math display">\[
 88Z(x) = \prod_{i=0}^{k} x-z_i =\newline
 89=(x-z_0)(x-z_1)...(x-z_k)
 90\]</span></p><p>This polynomial ensures that when <span class="math inline">\(x=z_i\)</span> (<span class="math inline">\(z_i\)</span> being one of our points), the polynomial evaluation will be zero.</p>
 91
 92<p>Now we can put <span class="math inline">\(I(x)\)</span> and <span class="math inline">\(Z(x)\)</span> in place, obtaining <span class="math inline">\(q(x)=\frac{p(x)-I(x)}{Z(x)}\)</span>. And the batch proof evaluation is obtained by <span class="math inline">\(\pi=[q(\tau)]_1\)</span>.</p>
 93
 94<p>The verification is quite similar than what we did for single proofs, but using the mentioned <span class="math inline">\(z(x)\)</span> and <span class="math inline">\(I(x)\)</span>:</p>
 95<p><span class="math display">\[
 96\hat{e}(\pi, [Z(\tau)]_2) == \hat{e}(c - [I(\tau)]_1, H)
 97\]</span></p><p>Which, as we did with the single proofs in the previous post, we can unroll it and see that:</p>
 98<p><span class="math display">\[
 99\hat{e}(\pi, [Z(\tau)]_2) == \hat{e}(c - [I(\tau)]_1, H)\newline
100\Rightarrow \hat{e}([q(\tau)]_1, [Z(\tau)]_2) == \hat{e}([p(\tau)]_1 - [I(\tau)]_1, H)\newline
101\Rightarrow [q(\tau) \cdot Z(\tau)]_T == [p(\tau) - I(\tau)]_T
102\]</span></p><p>From where we see that is the equation <span class="math inline">\(q(x)\cdot Z(x)=p(x)-I(x)\)</span>, which can be expressed as <span class="math inline">\(q(x) = \frac{p(x) - I(x)}{Z(x)}\)</span>, evaluated at <span class="math inline">\(\tau\)</span> from the trusted setup, which is not known: <span class="math inline">\(q(\tau) = \frac{p(\tau) - I(\tau)}{Z(\tau)}\)</span>.</p>
103
104<h4>Vector commitments</h4>
105
106<p>As mentioned earlier, this scheme can be used as a <em>vector commitment</em> scheme.</p>
107
108<p>A vector commitment allows a prover to commit to a vector and later proof that a certain value belongs to that vector. As a traditional example, we can think of <em>Merkle Trees</em>, where we can commit to a vector of values, which are placed in the tree leafs. Then we compute the <em>root</em> which acts as the commitment. Then we can provide a proof that a certain leaf belongs to the vector for the commitment (<em>root</em>) that we showed earlier.
109The problem with Merkle Trees is that the proof size grows linearly with the size of the tree, as it contains the siblings from the leaf to the root. Here is where <em>KZG Commitments</em> can be benefitial, as the proof always stands with the same size, one <span class="math inline">\(\mathbb{G}_1\)</span> point, no matter of how many points we are batching.</p>
110
111<p>We can use KZG Commitments as a vector commitment scheme by mapping the <em>vector</em> as a batch of points that build the polynomial, so when commiting to the polynomial we are commiting to the vector. Then, it&rsquo;s a matter of using the <em>batch proof</em> approach explained above in order to proof multiple elements of the vector in a single proof that can be verified later.</p>
112
113<h4>Final note</h4>
114
115<p>I&rsquo;m fascinated by this scheme and its potential. One next rabbit hole (related to KZG Commitments) to look at, would be the <a href="https://vitalik.ca/general/2019/09/22/plonk.html">Plonk</a> and other similar constructions. Also, another use case that is getting attention in the Ethereum community is the <a href="https://vitalik.ca/general/2021/06/18/verkle.html"><em>Verkle Trees</em></a>.</p>
116
117<p>As in the previous notes, in order to try to put in practice the concepts, I&rsquo;ve added the implementation of the <em>batch proof</em> logic at the Go KZG implementation <a href="https://github.com/arnaucube/kzg-commitments-study">https://github.com/arnaucube/kzg-commitments-study</a>.</p>
118
119    </div>
120
121    <footer style="text-align:center; margin-top:100px;margin-bottom:50px;">
122      <div class="container">
123        <br>
124        <a href="/blog">Go to main</a>
125        <br><br>
126        <div class="row">
127          <ul class="list-inline">
128            <li><a href="https://twitter.com/arnaucube"
129                   style="color:gray;text-decoration:none;"
130                   target="_blank">twitter.com/arnaucube</a>
131            </li>
132            <li><a href="https://github.com/arnaucube"
133                   style="color:gray;text-decoration:none;"
134                   target="_blank">github.com/arnaucube</a>
135            </li>
136          </ul>
137        </div>
138        <div class="row" style="display:inline-block;">
139          Blog made with <a href="http://github.com/arnaucube/blogo/"
140                            target="_blank" style="color: gray;text-decoration:none;">Blogo</a>
141        </div>
142      </div>
143    </footer>
144
145    <script>
146    </script>
147    <script src="js/external-links.js"></script>
148    <script>hljs.initHighlightingOnLoad();</script>
149    <script defer src="js/katex/katex.min.js"></script>
150    <script defer src="js/katex/auto-render.min.js"></script>
151    <script>
152      document.addEventListener("DOMContentLoaded", function() {
153        renderMathInElement(document.body, {
154          displayMode: false,
155          // customised options
156          // • auto-render specific keys, e.g.:
157          delimiters: [
158            {left: '$$', right: '$$', display: true},
159            {left: '$', right: '$', display: false},
160            {left: "\\[", right: "\\]", display: true},
161            {left: "\\(", right: "\\)", display: false},
162            {left: "\\begin{equation}", right: "\\end{equation}", display: true}
163          ],
164          // • rendering keys, e.g.:
165          throwOnError : true
166        });
167      });
168
169      ///
170      let theme = localStorage.getItem("theme");
171      if ((theme === "light-theme")||(theme==null)) {
172          theme = "light-theme";
173          document.getElementById("themeSwitcher").checked = false;
174      } else if (theme === "dark-theme") {
175          theme = "dark-theme";
176          document.getElementById("themeSwitcher").checked = true;
177      }
178      document.body.className = theme;
179      localStorage.setItem("theme", theme);
180
181      function switchTheme() {
182        theme = localStorage.getItem("theme");
183        if (theme === "light-theme") {
184          theme = "dark-theme";
185          document.getElementById("themeSwitcher").checked = true;
186        } else {
187          theme = "light-theme";
188          document.getElementById("themeSwitcher").checked = false;
189        }
190        document.body.className = theme;
191        localStorage.setItem("theme", theme);
192
193        console.log(theme);
194      }
195    </script>
196    <script>
197    function tagLinks(tagName) {
198      var tags = document.getElementsByTagName(tagName);
199      for (var i=0, hElem; hElem =  tags[i]; i++) {
200        if (hElem.parentNode.className=="row postThumb") {
201          continue;
202        }
203        hElem.id = hElem.innerHTML.toLowerCase().replace(" ", "-");
204        hElem.innerHTML = "<a style='color:black!important;' href='#"+hElem.id+"'>"+hElem.innerHTML+"</a>";
205      }
206    }
207    tagLinks("h2");
208    tagLinks("h3");
209    tagLinks("h4");
210    tagLinks("h5");
211    </script>
212    <script>
213function generateTOC(tocContainerId = "toc") {
214    const tocContainer = document.getElementById(tocContainerId);
215    if (!tocContainer) return;
216
217    const headings = document.querySelectorAll("h2, h3, h4, h5, h6");
218    if (!headings.length) return;
219
220    // Generate IDs if missing
221    headings.forEach((heading, index) => {
222        if (!heading.id) {
223            heading.id =
224                heading.textContent
225                    .trim()
226                    .toLowerCase()
227                    .replace(/[^\w\s-]/g, "")
228                    .replace(/\s+/g, "-") +
229                "-" +
230                index;
231        }
232    });
233
234    const toc = document.createElement("ul");
235    let currentLevel = 1;
236    let currentList = toc;
237    const stack = [toc];
238
239    headings.forEach((heading) => {
240        const level = parseInt(heading.tagName.substring(1), 10);
241
242        const li = document.createElement("li");
243        const a = document.createElement("a");
244
245        a.href = "#" + heading.id;
246        a.textContent = heading.textContent;
247
248        li.appendChild(a);
249
250        if (level > currentLevel) {
251            for (let i = currentLevel; i < level; i++) {
252                const nestedList = document.createElement("ul");
253
254                const lastLi =
255                    currentList.lastElementChild ||
256                    currentList.appendChild(document.createElement("ul"));
257
258                lastLi.appendChild(nestedList);
259
260                stack.push(nestedList);
261                currentList = nestedList;
262            }
263        } else if (level < currentLevel) {
264            for (let i = currentLevel; i > level; i--) {
265                stack.pop();
266            }
267            currentList = stack[stack.length - 1];
268        }
269
270        currentList.appendChild(li);
271        currentLevel = level;
272    });
273
274    tocContainer.innerHTML = "";
275    tocContainer.appendChild(toc);
276}
277
278document.addEventListener("DOMContentLoaded", () => {
279    generateTOC();
280});
281    </script>
282    <script src="js/mermaid.min.js"></script>
283
284  </body>
285</html>
286