September 2018 - tor-commits - lists.torproject.org

[metrics-web/release] Add reproducible-metrics page.
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 9957b04b768176be58380d6d688023007dcd8696 Author: Karsten Loesing <karsten.loesing(a)gmx.net> Date: Fri Mar 30 16:16:48 2018 +0200 Add reproducible-metrics page. Implements the first part of #26857. --- .../metrics/web/ReproducibleMetricsServlet.java | 30 + src/main/resources/web.xml | 11 + .../resources/web/jsps/reproducible-metrics.jsp | 906 +++++++++++++++++++++ 3 files changed, 947 insertions(+) diff --git a/src/main/java/org/torproject/metrics/web/ReproducibleMetricsServlet.java b/src/main/java/org/torproject/metrics/web/ReproducibleMetricsServlet.java new file mode 100644 index 0000000..3299cd2 --- /dev/null +++ b/src/main/java/org/torproject/metrics/web/ReproducibleMetricsServlet.java @@ -0,0 +1,30 @@ +/* Copyright 2018 The Tor Project + * See LICENSE for licensing information */ + +package org.torproject.metrics.web; + +import java.io.IOException; + +import javax.servlet.ServletException; +import javax.servlet.http.HttpServletRequest; +import javax.servlet.http.HttpServletResponse; + +public class ReproducibleMetricsServlet extends AnyServlet { + + private static final long serialVersionUID = 6099009779662419291L; + + @Override + public void init() throws ServletException { + super.init(); + } + + @Override + public void doGet(HttpServletRequest request, + HttpServletResponse response) throws IOException, ServletException { + + request.setAttribute("categories", this.categories); + request.getRequestDispatcher("WEB-INF/reproducible-metrics.jsp") + .forward(request, response); + } +} + diff --git a/src/main/resources/web.xml b/src/main/resources/web.xml index 1fe51b9..154a64e 100644 --- a/src/main/resources/web.xml +++ b/src/main/resources/web.xml @@ -273,6 +273,17 @@ </servlet-mapping> <servlet> + <servlet-name>ReproducibleMetricsServlet</servlet-name> + <servlet-class> + org.torproject.metrics.web.ReproducibleMetricsServlet + </servlet-class> + </servlet> + <servlet-mapping> + <servlet-name>ReproducibleMetricsServlet</servlet-name> + <url-pattern>/reproducible-metrics.html</url-pattern> + </servlet-mapping> + + <servlet> <servlet-name>OperationServlet</servlet-name> <servlet-class> org.torproject.metrics.web.OperationServlet diff --git a/src/main/resources/web/jsps/reproducible-metrics.jsp b/src/main/resources/web/jsps/reproducible-metrics.jsp new file mode 100644 index 0000000..f1b97f6 --- /dev/null +++ b/src/main/resources/web/jsps/reproducible-metrics.jsp @@ -0,0 +1,906 @@ +<%@ taglib prefix="c" uri="http://java.sun.com/jsp/jstl/core" %> +<%@ taglib prefix="fn" uri="http://java.sun.com/jsp/jstl/functions" %> +<jsp:include page="top.jsp"> + <jsp:param name="pageTitle" value="Sources – Tor Metrics"/> + <jsp:param name="navActive" value="Sources"/> +</jsp:include> + +<div class="container"> +<ul class="breadcrumb"> +<li><a href="/">Home</a></li> +<li><a href="sources.html">Sources</a></li> +<li class="active">Reproducible Metrics</li> +</ul> +</div> + +<div class="container"> + +<div class="panel panel-danger"> +<div class="panel-heading"> +<h5 class="panel-title">Work in progress notice</h5> +</div> +<div class="panel-body"> +<p>As of July 2018, this page is still a work in progress. Handle with care!</p> +</div> +</div> + +<h1>Reproducible Metrics +<a href="#reproducible-metrics" name="reproducible-metrics" class="anchor">#</a></h1> + +<p>The graphs and tables on Tor Metrics are the result of aggregating data obtained from several points in the Tor network. +Some of these aggregations are straightforward, but some are not.</p> + +<p>We want to make the graphs and tables on this site easier to access and reproduce, so on this page, we specify how you can reproduce the data behind them to create your own. +We also provide background for some of the design decisions behind our aggregations and link to <a href="https://research.torproject.org/techreports.html">technical reports</a> and other additional information.</p> + +<p>This page is a living document that reflects the latest changes to graphs and tables on Tor Metrics. +Whenever we create new aggregations or visualizations, we may write down our thoughts in technical reports; but, if we later expand or change a statistic, we don't update the original technical reports. +Instead, we update the specification here.</p> + +<p>While we may refer to technical reports for additional details, we do not assume their knowledge in order to make sense of the specifications here. +Knowledge of our source code is not needed, either.</p> +</div> + +<div class="container"> + +<h2><i class="fa fa-users fa-fw" aria-hidden="true"></i> +Users <a href="#users" name="users" class="anchor">#</a></h2> + +<p>The number of Tor users is one of our most important statistics. It is vital for us to know how many people use the Tor network on a daily basis, whether they connect via <a href="/glossary.html#relay">relays</a> or <a href="/glossary.html#bridge">bridges</a>, from which countries they connect, what <a href="/glossary.html#pluggable-transport">transports</a> they use, and whether they connect via IPv4 or IPv6.</p> + +<p>Due to the nature of Tor being an anonymity network, we cannot collect identifying data to learn the number of users. That is why we actually don't count users, but we count requests to the directories or bridges that <a href="/glossary.html#client">clients</a> clients make periodically to update their list of relays and estimate user numbers indirectly from there.</p> + +<p>The result is an average number of concurrent users, estimated from data collected over a day. +We can't say how many distinct users there are. +That is, we can't say whether the same set of users stays connected over the whole day, or whether that set leaves after a few hours and a new set of users arrives. +However, the main interest is finding out if usage changes, for which it is not critical to estimate exact absolute user numbers.</p> + + +</div> + +<div class="container"> + +<h3 id="relay-users" class="hover">Relay users +<a href="#relay-users" class="anchor">#</a> +</h3> + +<p>Relay users are users that connect directly to a relay in order to connect to the Tor network—as opposed to bridge users that connect to a bridge as entry point into the Tor network. +Many steps here are similar to the steps for estimating bridge users, which are specified further down <a href="#bridge-users">below</a>.</p> + +<p>The following description applies to the following graph and tables:</p> + +<ul> +<li>Relay users <a href="/userstats-relay-country.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Top-10 countries by relay users <a href="/userstats-relay-table.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> table</a></li> +<li>Top-10 countries by possible censorship events <a href="/userstats-censorship-events.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> table</a></li> +</ul> + +<h4>Step 1: Parse consensuses to learn which relays have been running</h4> + +<p>Obtain consensuses from <a href="/collector.html#type-network-status-consensus-3">CollecTor</a>. +Refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>From each consensus, parse the <code>"valid-after"</code> and <code>"fresh-until"</code> times from the header section.</p> + +<p>From each consensus entry, parse the base64-encoded relay fingerprint from the <code>"r"</code> line. Also parse the <a href="/glossary.html#relay-flag">relay flags</a> from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this entry. +(Consensuses with consensus method 4, introduced in 2008, or later do not list non-running relays, so that checking relay flags in recent consensuses is mostly done as a precaution without actual effect on the parsed data.)</p> + +<h4>Step 2: Parse relay extra-info descriptors to learn relevant statistics reported by relays</h4> + +<p>Also obtain relay extra-info descriptors from <a href="/collector.html#type-extra-info">CollecTor</a>. +As above, refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>Parse the relay fingerprint from the <code>"extra-info"</code> line and the descriptor publication time from the <code>"published"</code> line.</p> + +<p>Parse the <code>"dirreq-write-history"</code> line containing written bytes spent on answering directory requests. If the contained statistics end time is more than 1 week older than the descriptor publication time in the <code>"published"</code> line, skip this line to avoid including statistics in the aggregation that have very likely been reported in earlier descriptors and processed before. If a statistics interval spans more than 1 UTC date, split observations to the covered UTC dates by assuming a linear distribution of observations.</p> + +<p>Parse the <code>"dirreq-stats-end"</code> and <code>"dirreq-v3-reqs"</code> lines containing directory-request statistics. +If the statistics end time in the <code>"dirreq-stats-end"</code> line is more than 1 week older than the descriptor publication time in the <code>"published"</code> line, skip these directory request statistics for the same reason as given above: to avoid including statistics in the aggregation that have very likely been reported in earlier descriptors and processed before. +Also skip statistics with an interval length other than 1 day. +Parse successful requests by country from the <code>"dirreq-v3-reqs"</code> line. From each number, subtract <code>4</code> to undo the binning operation that has been applied by the relay. Discard the resulting number if it's zero or negative. +Split observations to the covered UTC dates by assuming a linear distribution of observations.</p> + +<h4>Step 3: Estimate fraction of reported directory-request statistics</h4> + +<p>The next step after parsing descriptors is to estimate the fraction of reported directory-request statistics on a given day. +This fraction, a value between <var>0%</var> and <var>100%</var>, will be used in the next step to extrapolate observed request numbers to expected network totals. +For further background on the following calculation method, refer to the technical report titled <a href="https://research.torproject.org/techreports/counting-daily-bridge-users-201…">"Counting daily bridge users"</a> which also applies to relay users. +In the following, we're using the term server instead of relay or bridge, because the estimation method is exactly the same for relays and bridges.</p> + +<p>For each day in the time period, compute five variables:</p> + +<ul> +<li>Compute <var>n(N)</var> as the total server uptime in hours on a given day, that is, the sum of all server uptime hours on that day. This is the sum of all intervals between <code>"valid-after"</code> and <code>"fresh-until"</code>, multiplied by the contained running servers, for all consensuses with a valid-after time on a given day. +A more intuitive interpretation of this variable is the average number of running servers—however, that interpretation only works as long as fresh consensuses are present for all hours of a day.</li> +<li>Compute <var>n(H)</var> as the total number of hours for which servers have reported written directory bytes on a given day.</li> +<li>Compute <var>n(R\H)</var> as the number of hours for which responses have been reported but no written directory bytes. This fraction is determined by summing up all interval lengths and then subtracting the written directory bytes interval length from the directory response interval length. Negative results are discarded.</li> +<li>Compute <var>h(H)</var> as the total number of written directory bytes on a given day.</li> +<li>Compute <var>h(R^H)</var> as the number of written directory bytes for the fraction of time when a server was reporting both written directory bytes and directory responses. As above, this fraction is determined by first summing up all interval lengths and then computing the minimum of both sums divided by the sum of reported written directory bytes.</li> +</ul> + +<p>From these variables, compute the estimated fraction of reported directory-request statistics using the following formula:</p> + +<pre> + h(R^H) * n(H) + h(H) * n(R\H) +frac = ----------------------------- + h(H) * n(N) +</pre> + +<h4>Step 4: Compute estimated relay users per country</h4> + +<p>With the estimated fraction of reported directory-request statistics from the previous step it is now possible to compute estimates for relay users. +Similar to the previous step, the same approach described here also applies to estimating bridge users by country, transport, or IP version as described further down below.</p> + +<p>First compute <var>r(R)</var> as the sum of reported successful directory requests from a given country on a given day. +This approach also works with <var>r(R)</var> being the sum of requests from <em>all</em> countries or from any other subset of countries, if this is of interest.</p> + +<p>Estimate the number of clients per country and day using the following formula:</p> + +<pre>r(N) = r(R) / frac / 10</pre> + +<p>A client that is connected 24/7 makes about 15 requests per day, but not all clients are connected 24/7, so we picked the number 10 for the average client. We simply divide directory requests by 10 and consider the result as the number of users. Another way of looking at it, is that we assume that each request represents a client that stays online for one tenth of a day, so 2 hours and 24 minutes.</p> +<p>Skip dates where <var>frac</var> is smaller than 10% and hence too low for a robust estimate, or where <var>frac</var> is greater than 100%, which would indicate an issue in the previous step.</p> + +<h4>Step 5: Compute ranges of expected clients per day to detect potential censorship events</h4> + +<p>As last step in reproducing relay user numbers, compute ranges of expected clients per day to detect potential censorship events. +For further details on the detection method, refer to the technical report titled <a href="https://research.torproject.org/techreports/detector-2011-09-09.pdf">"An anomaly-based censorship-detection system for Tor"</a>. +Unlike the previous two steps, this step only applies to relay users, not to bridge users.</p> + +<ol> +<li>Start by finding the 50 largest countries by estimated relay users, excluding <code>"??"</code>, on the last day in the data set. (Note that, as new data gets available, the set of 50 largest countries may change, too, affecting ranges for the entire data set.)</li> +<li>For each of these largest countries and for each date in the data set, compute the ratio R_ij = C_ij / C_(i-7)j of estimated users on any given date divided by estimated users 1 week earlier. Exclude ratios for which there are no user estimates from 1 week earlier, or where that estimate is 0.</li> +<li>For the computed ratios on each date, remove outliers that fall outside four interquartile ranges of the median, and remove dates with less than 8 ratios remaining.</li> +<li>For each date, compute mean and standard variation in order to use these as parameters of a normal distribution.</li> +<li>For each date and country, compute a range of expected user numbers by using a normal distribution with parameters from the previous step and the ratio of estimated users divided by estimated users 1 week earlier as input. As previously, exclude ratios for which there are no user estimates from 1 week earlier, or where that estimate is 0. Compute the low and high ranges as: +<ul> +<li>low = max(0, NormalDistribution(mu, standard deviation, P = 0.0001) * PoissonDistribution(lambda = C_(i-7)j, P = 0.0001))</li> +<li>high = NormalDistribution(mu, standard deviation, P = 0.9999) * PoissonDistribution(lambda = C_(i-7)j, P = 0.9999)</li> +</ul> +</ol> + +</div> + +<div class="container"> +<h3 id="bridge-users" class="hover">Bridge users +<a href="#bridge-users" class="anchor">#</a> +</h3> + +<p>Bridge users are users that connect to a <a href="/glossary.html#bridge">bridge</a> as entry point into the Tor network as opposed to relay users that connect directly to a relay. +Many steps here are similar to the steps for estimating relay users, which are specified above.</p> + +<ul> +<li>Bridge users by country <a href="/userstats-bridge-country.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Bridge users by transport <a href="/userstats-bridge-transport.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Bridge users by country and transport <a href="/userstats-bridge-combined.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Bridge users by IP version <a href="/userstats-bridge-version.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Top-10 countries by bridge users <a href="/userstats-bridge-table.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> table</a></li> +</ul> + +<h4>Step 1: Parse bridge network statuses to learn which bridges have been running</h4> + +<p>Obtain bridge network statuses from <a href="/collector.html#type-bridge-network-status">CollecTor</a>. +Refer to the <a href="/bridge-descriptors.html">Tor bridge descriptors page</a> for details on the descriptor format.</p> + +<p>From each status, parse the <code>"published"</code> time from the header section.</p> + +<p>From each status entry, parse the base64-encoded hashed bridge fingerprint from the <code>"r"</code> line. Also parse the <a href="/glossary.html#relay-flag">relay flags</a> from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this entry.</p> + +<p>As opposed to relay consensuses, there are no <code>"valid-after"</code> or <code>"fresh-until"</code> times in the header of bridge network statuses. +To unify processing, we use the publication hour as valid-after time and one hour later as fresh-until time. +If we process multiple statuses published in the same hour, we take the union of contained running bridges as running bridges in that hour.</p> + +<h4>Step 2: Parse bridge extra-info descriptors to learn relevant statistics reported by bridges</h4> + +<p>Also obtain bridge extra-info descriptors from <a href="/collector.html#type-bridge-extra-info">CollecTor</a>. +As above, refer to the <a href="/bridge-descriptors.html">Tor bridge descriptors page</a> for details on the descriptor format.</p> + +<p>Parse the hashed bridge fingerprint from the <code>"extra-info"</code> line and the descriptor publication time from the <code>"published"</code> line.</p> + +<p>Parse the <code>"dirreq-write-history"</code> line containing written bytes spent on answering directory requests. If the contained statistics end time is more than 1 week older than the descriptor publication time in the <code>"published"</code> line, skip this line to avoid including statistics in the aggregation that have very likely been reported in earlier descriptors and processed before. If a statistics interval spans more than 1 UTC date, split observations to the covered UTC dates by assuming a linear distribution of observations.</p> + +<p>Parse the <code>"dirreq-stats-end"</code> and <code>"dirreq-v3-resp"</code> lines containing directory-request statistics. +If the statistics end time in the <code>"dirreq-stats-end"</code> line is more than 1 week older than the descriptor publication time in the <code>"published"</code> line, skip these directory request statistics for the same reason as given above: to avoid including statistics in the aggregation that have very likely been reported in earlier descriptors and processed before. +Also skip statistics with an interval length other than 1 day. +Parse successful requests from the <code>"ok"</code> part of the <code>"dirreq-v3-resp"</code> line. Subtract <code>4</code> to undo the binning operation that has been applied by the bridge. Discard the resulting number if it's zero or negative. +Split observations to the covered UTC dates by assuming a linear distribution of observations.</p> + +<p>Parse the <code>"bridge-ips"</code>, <code>"bridge-ip-versions"</code>, and <code>"bridge-ip-transports"</code> lines containing unique connecting IP addresses by country, IP version, and transport. From each number of unique IP addresses, subtract 4 to undo the binning operation that has been applied by the bridge. Discard the resulting number if it's zero or negative.</p> + + + +<h4>Step 3: Approximate directory requests by country, transport, and IP version</h4> + +<p>Bridges, unlike relays, do not report directory request numbers by country, transport, or IP version. +However, bridges do report unique IP address counts by country, by transport, and by IP version. +We approximate directory request numbers by multiplying the fraction of unique IP addresses from a given country, transport, or IP version with the total number of successful requests.</p> + +<p>As a special case, we also approximate lower and upper bounds for directory requests by country <em>and</em> transport. +This approximation is based on the fact that most bridges only provide a small number of transports. +This allows us to combine unique IP address sets by country and by transport and obtain lower and upper bounds:</p> + +<ul> +<li>We calculate the lower bound as <code>max(0, C(b) + T(b) - S(b))</code> using the following definitions: <code>C(b)</code> is the number of requests from a given country reported by bridge <code>b</code>; <code>T(b)</code> is the number of requests using a given transport reported by bridge <code>b</code>; and <code>S(b)</code> is the total numbers of requests reported by bridge <code>b</code>. Reasoning: If the sum <code>C(b) + T(b)</code> exceeds the total number of requests from all countries and transports <code>S(b)</code>, there must be requests from that country and transport. And if that is not the case, <code>0</code> is the lower limit.</li> +<li>We calculate the upper bound as <code>min(C(b), T(b))</code> with the definitions from above. Reasoning: There cannot be more requests by country and transport than there are requests by either of the two numbers. +</ul> + +<p>If a bridge does not report unique IP addresses by country, transport, or IP version, we attribute all requests to "??" which stands for Unknown Country, to the default onion-routing protocol, or to IPv4.</p> + +<h4>Step 4: Estimate fraction of reported directory-request statistics</h4> + +<p>The step for estimating the fraction of reported directory-request statistics is pretty much the same for bridges and for relays. +This is why we refer to Step 3 of the <a href="#relay-users">Relay users</a> description for this estimation.</p> + +<h4>Step 5: Compute estimated bridge users per country, transport, or IP version</h4> + +<p>Similar to the previous step, this step is equivalent for bridge users and relay users. +We therefore refer to Step 4 of the <a href="#relay-users">Relay users</a> description for transforming directory request numbers to user numbers.</p> + +</div> + +<div class="container"> +<h2><i class="fa fa-server fa-fw" aria-hidden="true"></i> +Servers <a href="#servers" name="servers" class="anchor">#</a></h2> + +<p>The following description applies to the following graphs and table:</p> + +<p>Statistics on the number of servers—<a href="/glossary.html#relay">relays</a> and <a href="/glossary.html#bridge">bridges</a>—were among the first to appear on Tor Metrics. +These statistics have one thing in common: they use the number of running servers as their metric. +Possible alternatives would be to use <a href="/glossary.html#consensus-weight">consensus weight</a> fractions or guard/middle/exit probabilities as metrics, but we're not doing that yet. +In the following, we describe how exactly we count servers.</p> + +<h3 id="running-relays" class="hover">Running relays +<a href="#running-relays" class="anchor">#</a> +</h3> + +<p>We start with statistics on the number of running relays in the network, broken down by criteria like assigned <a href="/glossary.html#relay-flag">relay flag</a>, self-reported tor version and operating system, or IPv6 capabilities.</p> + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Relays and bridges (just the relays part; for the bridges part <a href="#running-bridges">see below</a>) <a href="/networksize.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Relays by relay flag <a href="/relayflags.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Relays by tor version <a href="/versions.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Relays by platform <a href="/platforms.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Relays by IP version <a href="/relays-ipv6.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse consensuses</h4> + +<p>Obtain consensuses from <a href="/collector.html#type-network-status-consensus-3">CollecTor</a>. +Refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>Parse and memorize the <code>"valid-after"</code> time from the consensus header. We use this UTC timestamp to uniquely identify the consensus while processing, and to later aggregate by the UTC date of this UTC timestamp.</p> + +<p>Repeat the following steps for each consensus entry:</p> + +<ul> +<li>Server descriptor digest: Parse the server descriptor digest from the <code>"r"</code> line. This is only needed for statistics based on relay server descriptor contents.</li> +<li>IPv6 reachable OR: Parse any <code>"a"</code> lines, if present, and memorize whether at least one of them contains an IPv6 address. This indicates that at least one of the relay's IPv6 OR addresses is reachable.</li> +<li>Relay flags: Parse relay flags from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this consensus entry. This ensures that we only consider running relays. Also parse any other relay flags from the <code>"s"</code> line that the relay had assigned.</li> +</ul> + +<p>If a consensus contains zero running relays, we skip it in the <a href="/relays-ipv6.html">Relays by IP version</a> graph, but not in the other graphs (simply because we didn't get around to changing those graphs). +This is mostly to rule out a rare edge case when only a minority of <a href="/glossary.html#directory-authority">directory authorities</a> voted on the <code>"Running"</code> flag. +In those cases, such a consensus would skew the average, even though relays were likely running.</p> + +<h4>Step 2: Parse relay server descriptors</h4> + +<p>Parsing relay server descriptors is an optional step. You only need to do this if you want to break down the number of running relays by something that relays report in their server descriptors. This includes, among other things, the relay's platform string containing tor software version and operating system and whether the relay announced an IPv6 OR address or permitted exiting to IPv6 targets.</p> + +<p>Obtain relay server descriptors from <a href="/collector.html#type-server-descriptor">CollecTor</a>. +Again, refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>Parse any or all of the following parts from each server descriptor:</p> + +<ul> +<li>Tor version: Parse the tor software version from the <code>"platform"</code> line and memorize the first three dotted numbers from it. +If the <code>"platform"</code> line does not begin with "Tor" followed by a space character and a dotted version number, memorize the version as "Other". + +If the platform line is missing, we skip this descriptor, which later leads to not counting this relay at all rather than including it in the "Other" group, which is slightly wrong. +Note that consensus entries also contain a <code>"v"</code> line with the tor software version from the referenced descriptor, which we do not use, because it was not present in very old consensuses, but which should work just as well for recent consensus.</li> +<li>Operating system: Parse the <code>"platform"</code> line and memorize whether it contains either one of the substrings "Linux", "Darwin" (macOS), "BSD", or "Windows". +If the <code>"platform"</code> line contains neither of these substrings, memorize the platform as "Other". +If the platform line is missing, we skip this descriptor, which later leads to not counting this relay at all rather than including it in the "Other" group, which is slightly wrong.</li> +<li>IPv6 announced OR: Parse any <code>"or-address"</code> lines and memorize whether at least one of them contains an IPv6 address. This indicates that the relay announced an IPv6 address.</li> +<li>IPv6 exiting: Parse the <code>"ipv6-policy"</code> line, if present, and memorize whether it's different from "reject 1-65535". This indicates whether the relay permitted exiting to IPv6 targets. If the line is not present, memorize that the relay does not permit exiting to IPv6 targets.</li> +<li>Server descriptor digest: Compute the SHA-1 digest, or determine it from the file name in case of archived descriptor tarballs.</p> +</ul> + +<h4>Step 3: Compute daily averages</h4> + +<p>Optionally, match consensus entries with server descriptors by SHA-1 digest. +Every consensus entry references exactly one server descriptor, and a server descriptor may be referenced from an arbitrary number of consensus entries. +We handle missing server descriptors differently in the graphs covered in this section:</p> + +<ul> +<li><a href="/versions.html">Relays by tor version</a> and <a href="/platforms.html">Relays by platform</a>: If a referenced server descriptor is missing, we also skip the consensus entry. We are aware that this is slightly wrong, because we should either exclude a consensus with too few matching server descriptors from the overall result, or at least count these relays as unknown tor version or unknown platform.</li> +<li><a href="/relays-ipv6.html">Relays by IP version</a>: If at least 0.1% of referenced server descriptors are missing, we skip the consensus. We chose this threshold as low, because missing server descriptors may easily skew the results. However, a small number of missing server descriptors per consensus is acceptable and also unavoidable.</li> +</ul> + +<p>Go through all previously processed consensuses by valid-after UTC date. +Compute the arithmetic mean of running relays, possibly broken down by relay flag, tor version, platform, or IPv6 capabilities, as the sum of all running relays divided by the number of consensuses. +Round down to the next integer number.</p> + +<p>Skip the last day of the results if it matches the current UTC date, because those averages may still change throughout the day. +For the <a href="/relays-ipv6.html">Relays by IP version</a> graph we further skip days for which fewer than 12 consensuses are known. The goal is to avoid over-representing a few consensuses during periods when the directory authorities had trouble producing a consensus for at least half of the day.</p> + +<h3 id="running-bridges" class="hover">Running bridges +<a href="#running-bridges" class="anchor">#</a> +</h3> + +<p>After explaining our running <a href="/glossary.html#relay">relays</a> statistics we continue with our running <a href="/glossary.html#bridge">bridges</a> statistics. +The steps are quite similar, except for a couple differences in data formats that justify explaining these statistics in a separate subsection.</p> + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Relays and bridges (just the bridges part; for the relays part <a href="#running-relays">see above</a>) <a href="/networksize.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Bridges by IP version <a href="/bridges-ipv6.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse bridge network statuses</h4> + +<p>Obtain bridge network statuses from <a href="/collector.html#type-bridge-network-status">CollecTor</a>. +Refer to the <a href="/bridge-descriptors.html">Tor bridge descriptors page</a> for details on the descriptor format.</p> + +<p>Parse the <a href="/glossary.html#bridge-authority">bridge authority</a> identify from the file name and memorize it. +This is only relevant for times when more than 1 bridge authority was running. +In those cases, bridges typically register at a single bridge authority only, so that taking the average of running bridges over all statuses on those day would be misleading.</p> + +<p>Parse and memorize the <code>"published"</code> time either from the file name or from the status header. +This timestamp is used to uniquely identify the status while processing, and the UTC date of this timestamp is later used to aggregate by UTC date.</p> + +<p>Repeat the following steps for each status entry:</p> + +<ul> +<li>Server descriptor digest: Parse the server descriptor digest from the <code>"r"</code> line and memorize it.</li> +<li>Relay flags: Parse <a href="/glossary.html#relay-flag">relay flag</a> from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this entry. This ensures that we only consider running bridges.</li> +</ul> + +<p>If a status contains zero running bridges, skip it. This may happen when there is a temporary issue with the bridge authority.</p> + +<h4>Step 2: Parse bridge server descriptors.</h4> + +<p>Parsing bridge server descriptors is an optional step. You only need to do this if you want to break down the number of running bridges by something that bridges report in their server descriptors. +This includes, among other things, whether the bridge announced an IPv6 OR address.</p> + +<p>Obtain bridge server descriptors from <a href="/collector.html#type-bridge-server-descriptor">CollecTor</a>. +As above, refer to the <a href="/bridge-descriptors.html">Tor bridge descriptors page</a> for details on the descriptor format.</p> + +<p>Parse the following parts from each server descriptor:</p> + +<ul> +<li>IPv6 announced OR: Parse any <code>"or-address"</code> lines and memorize whether at least one of them contains an IPv6 address. This indicates that the bridge announced an IPv6 address.</li> +<li>Server descriptor digest: Parse the SHA-1 digest from the <code>"router-digest"</code> line, or determine it from the file name in case of archived descriptor tarballs.</li> +</ul> + +<h4>Step 3: Compute daily averages</h4> + +<p>Optionally, match status entries with server descriptors by SHA-1 digest. +Every status entry references exactly one server descriptor, and a server descriptor may be referenced from an arbitrary number of status entries. +If at least 0.1% of referenced server descriptors are missing, we skip the status. +We chose this threshold as low, because missing server descriptors may easily skew the results. +However, a small number of missing server descriptors per status is acceptable and also unavoidable.</p> + +<p>We compute averages differently in the graphs covered in this section:</p> + +<ul> +<li><a href="/networksize.html">Relays and bridges</a>: For each bridge authority, compute the arithmetic mean of running bridges as the sum of all running bridges divided by the number of statuses; sum up averages for all bridge authorities per day and round down to the next integer number.</li> +<li><a href="/bridges-ipv6.html">Bridges by IP version</a>: Compute the arithmetic mean of running bridges as the sum of all running bridges divided by the number of statuses and round down to the next integer number. We are aware that this approach does not correctly reflect that bridges typically register at a single bridge authority only.</li> +</ul> + +<p>Skip the last day of the results if it matches the current UTC date, because those averages may still change throughout the day. +For the <a href="/bridges-ipv6.html">Bridges by IP version</a> graph we further skip days for which fewer than 12 statuses are known. +The goal is to avoid over-representing a few statuses during periods when the bridge directory authority had trouble producing a status for at least half of the day.</p> + +</div> + +<div class="container"> +<h2><i class="fa fa-road fa-fw" aria-hidden="true"></i> +Traffic <a href="#traffic" name="traffic" class="anchor">#</a></h2> + +<p>Our traffic statistics have in common that their metrics are based on user-generated traffic. +This includes advertised and consumed bandwidth and connection usage statistics.</p> + +<h3 id="advertised-bandwidth" class="hover">Advertised bandwidth +<a href="#advertised-bandwidth" class="anchor">#</a> +</h3> + +<p><a href="/glossary.html#advertised-bandwidth">Advertised bandwidth</a> is the volume of traffic, both incoming and outgoing, that a relay is willing to sustain, as configured by the operator and claimed to be observed from recent data transfers. +Relays self-report their advertised bandwidth in their server descriptors which we evaluate together with consensuses.</p> + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Total relay bandwidth (just the advertised bandwidth part; for the consumed bandwidth part <a href="#consumed-bandwidth">see below</a>) <a href="/bandwidth.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Advertised and consumed bandwidth by relay flag (just the advertised bandwidth part; for the consumed bandwidth part <a href="#consumed-bandwidth">see below</a>) <a href="/bandwidth-flags.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Advertised bandwidth by IP version <a href="/advbw-ipv6.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Advertised bandwidth distribution <a href="/advbwdist-perc.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Advertised bandwidth of n-th fastest relays <a href="/advbwdist-relay.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse relay server descriptors</h4> + +<p>Obtain relay server descriptors from <a href="/collector.html#type-server-descriptor">CollecTor</a>. +Refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>Parse the following parts from each server descriptor:</p> + +<ul> +<li>Advertised bandwidth: Parse the three values (or just two in very old descriptors) from the <code>"bandwidth"</code> line. These values stand for the average bandwidth, burst bandwidth, and observed bandwidth. The advertised bandwidth is the minimum of these values.</li> +<li>Server descriptor digest: Compute the SHA-1 digest, or determine it from the file name in case of archived descriptor tarballs.</p> +</ul> + + + +<h4>Step 2: Parse consensuses</h4> + +<p>Obtain consensuses from <a href="/collector.html#type-network-status-consensus-3">CollecTor</a>. +Refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>From each consensus, parse the <code>"valid-after"</code> time from the header section.</p> + +<p>From each consensus entry, parse the base64-encoded server descriptor digest from the <code>"r"</code> line. +We are going to use this digest to match the entry with the advertised bandwidth value from server descriptors later on.</p> + +<p>Also parse the <a href="/glossary.html#relay-flag">relay flags</a> from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this entry. +(Consensuses with consensus method 4, introduced in 2008, or later do not list non-running relays, so that checking relay flags in recent consensuses is mostly done as a precaution without actual effect on the parsed data.) +Further parse the <code>"Guard"</code>, <code>"Exit"</code>, and <code>"BadExit"</code> relay flags from this line. +We consider a relay with the <code>"Guard"</code> flag as guard and a relay with the <code>"Exit"</code> and without the <code>"BadExit"</code> flag as exit.</p> + +<h4>Step 3: Compute daily averages</h4> + +<p>The first three graphs described here, namely <a href="/bandwidth.html">Total relay bandwidth</a>, <a href="/bandwidth-flags.html">Advertised and consumed bandwidth by relay flag</a>, and <a href="/advbw-ipv6.html">Advertised bandwidth by IP version</a>, have in common that they show daily averages of advertised bandwidth.</p> + +<p>In order to compute these averages, first match consensus entries with server descriptors by SHA-1 digest. +Every consensus entry references exactly one server descriptor, and a server descriptor may be referenced from an arbitrary number of consensus entries. +We handle missing server descriptors differently in the graphs covered in this section:</p> + +<ul> +<li><a href="/bandwidth.html">Total relay bandwidth</a> and <a href="/bandwidth-flags.html">Advertised and consumed bandwidth by relay flag</a>: If a referenced server descriptor is missing, we also skip the consensus entry. We are aware that this is slightly wrong, because we should rather exclude a consensus with too few matching server descriptors from the overall result than including it with an advertised bandwidth sum that is too low.</li> +<li><a href="/advbw-ipv6.html">Advertised bandwidth by IP version</a>: If at least 0.1% of referenced server descriptors are missing, we skip the consensus. We chose this threshold as low, because missing server descriptors may easily skew the results. However, a small number of missing server descriptors per consensus is acceptable and also unavoidable.</li> +</ul> + +<p>Go through all previously processed consensuses by valid-after UTC date. +Compute the arithmetic mean of advertised bandwidth as the sum of all advertised bandwidth values divided by the number of consensuses. +Round down to the next integer number.</p> + +<p>Break down numbers by guards and/or exits by taking into account which <a href="/glossary.html#relay-flag">relay flags</a> a consensus entry had that referenced a server descriptor.</p> + +<p>Skip the last day of the results if it matches the current UTC date, because those averages may still change throughout the day. +For the <a href="/advbw-ipv6.html">Advertised bandwidth by IP version</a> graph we further skip days for which fewer than 12 consensuses are known. +The goal is to avoid over-representing a few consensuses during periods when the directory authorities had trouble producing a consensus for at least half of the day.</p> + +<h4>Step 4: Compute ranks and percentiles</h4> + +<p>The remaining two graphs described here, namely <a href="/advbwdist-perc.html">Advertised bandwidth distribution</a> and <a href="/advbwdist-relay.html">Advertised bandwidth of n-th fastest relays</a>, display advertised bandwidth ranks or percentiles.</p> + +<p>Similar to the previous step, match consensus entries with server descriptors by SHA-1 digest. +We handle missing server descriptors by simply skipping the consensus entry, at the risk of over-representing available server descriptors in consensuses where most server descriptors are missing.</p> + +<p>For the <a href="/advbwdist-perc.html">Advertised bandwidth distribution</a> graph, determine the i-th percentile value for each consensus. +We use a non-standard percentile definition that is loosely based on the nearest-rank method: +the P-th percentile (<code>0 ≤ P ≤ 100</code>) of a list of <code>N</code> ordered values (sorted from least to greatest) is the <em>largest</em> value in the list such that no more than <code>P</code> percent of the data is strictly less than the value and at least <code>P</code> percent of the data is less than or equal to that value. +We calculate the ordinal rank <code>n</code> using the following formula: <code>floor((P / 100) * (N - 1)) + 1</code></p> + +<p>Calculate the median value over all consensus from a given day for each percentile value.</p> + +<p>Consider the set of all running relays as well as the set of exit relays.</p> + +<p>For the <a href="/advbwdist-relay.html">Advertised bandwidth of n-th fastest relays</a> graph, determine the n-th highest advertised bandwidth value for each consensus, and then calculate the median value over all consensus from a given day. +Again consider the set of all running relays as well as the set of exit relays.</p> + +<h3 id="consumed-bandwidth" class="hover">Consumed bandwidth +<a href="#consumed-bandwidth" class="anchor">#</a> +</h3> + +<p>Consumed bandwidth, or <a href="/glossary.html#bandwidth-history">bandwidth history</a>, is the volume of incoming and/or outgoing traffic that a relay claims to have handled on behalf of clients. +Relays self-report bandwidth histories as part of their extra-info descriptors, which we evaluate in combination with consensuses.</p> + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Total relay bandwidth (just the consumed bandwidth part; for the advertised bandwidth part <a href="#advertised-bandwidth">see above</a>) <a href="/bandwidth.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Advertised and consumed bandwidth by relay flag (just the consumed bandwidth part; for the advertised bandwidth part <a href="#advertised-bandwidth">see above</a>) <a href="/bandwidth-flags.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Consumed bandwidth by Exit/Guard flag combination <a href="/bwhist-flags.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Bandwidth spent on answering directory requests <a href="/dirbytes.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse extra-info descriptors</h4> + +<p>Obtain extra-info descriptors from <a href="/collector.html#type-extra-info">CollecTor</a>. +Refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>Parse the fingerprint from the <code>"extra-info"</code> line. +We will use this fingerprint to deduplicate statistics included in other extra-info descriptor published by the same relay. +We may also use this fingerprint to attribute statistics to relays with the <code>"Exit"</code> and/or <code>"Guard"</code> flag.</p> + +<p>Parse the <code>"write-history"</code>, <code>"read-history"</code>, <code>"dirreq-write-history"</code>, and <code>"dirreq-read-history</code> lines containing consumed bandwidth statistics. +The first two histories include all bytes written or read by the relay, whereas the last two include only bytes spent on answering directory requests. +If a statistics interval spans more than 1 UTC date, split observations to the covered UTC dates by assuming a linear distribution of observations. +As a simplification, we shift reported statistics intervals forward to fully align with multiples of 15 minutes since midnight. +We also discard reported statistics with intervals that are not multiples of 15 minutes.</p> + +<h4>Step 2: Parse consensuses</h4> + +<p>Obtain consensuses from <a href="/collector.html#type-network-status-consensus-3">CollecTor</a>. +Refer to the <a href="https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt">Tor directory protocol, version 3</a> for details on the descriptor format.</p> + +<p>From each consensus, parse the <code>"valid-after"</code> time from the header section.</p> + +<p>From each consensus entry, parse the base64-encoded relay fingerprint from the <code>"r"</code> line. +We are going to use this fingerprint to match the entry with statistics from extra-info descriptors later on.</p> + +<p>Also parse the <a href="/glossary.html#relay-flag">relay flag</a> from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this entry. +(Consensuses with consensus method 4, introduced in 2008, or later do not list non-running relays, so that checking relay flags in recent consensuses is mostly done as a precaution without actual effect on the parsed data.) +Further parse the <code>"Guard"</code>, <code>"Exit"</code>, and <code>"BadExit"</code> relay flags from this line. +We consider a relay with the <code>"Guard"</code> flag as guard and a relay with the <code>"Exit"</code> and without the <code>"BadExit"</code> flag as exit.</p> + +<h4>Step 3: Compute daily totals</h4> + +<p>The first three graphs described here, namely <a href="/bandwidth.html">Total relay bandwidth</a>, <a href="/bandwidth-flags.html">Advertised and consumed bandwidth by relay flag</a>, and <a href="/bwhist-flags.html">Consumed bandwidth by Exit/Guard flag combination</a>, show daily totals of all bytes written or read by relays. +For all three graphs, we sum up all read and written bytes on a given day and divide the result by 2. +However, we only include bandwidth histories for a given day if a relay was listed as running in a consensus at least once on that day. +We attribute bandwidth to guards and/or exits if a relay was a guard and/or exit at least in one consensus on a day.</p> + +<p>The fourth graph, <a href="/dirbytes.html">Bandwidth spent on answering directory requests</a>, shows bytes spent by <a href="/glossary.html#directory-mirror">directory mirrors</a> on answering directory requests. +As opposed to the first three graphs, all bandwidth histories are included, regardless of whether a relay was listed as running in a consensus. +Also, we compute total read directory and total written directory bytes for this fourth graph, not an average of the two.</p> + +<h3 id="connbidirect" class="hover">Connection usage +<a href="#connbidirect" class="anchor">#</a> +</h3> + +<p>The last category of traffic statistics concerns statistics on the fraction of connections used uni- or bidirectionally. +A subset of relays reports these highly aggregated statistics in their extra-info descriptors.</p> + +<p>The following description applies to the following graph:</p> + +<ul> +<li>Fraction of connections used uni-/bidirectionally <a href="/connbidirect.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse relay extra-info descriptors</h4> + +<p>Obtain relay extra-info descriptors from <a href="/collector.html#type-extra-info">CollecTor</a>.</p> + +<p>Parse the relay fingerprint from the <code>"extra-info"</code> line. +We deduplicate reported statistics by UTC date of reported statistics and relay fingerprint. +If a given relay publishes different statistics on a given UTC day, we pick the first encountered statistics and discard all subsequent statistics by that relay on that UTC day.</p> + +<p>Parse the UTC date from the <code>"conn-bi-direct"</code> line. +We use this date to aggregate statistics, regardless of what period of time fell on the statistics end date as opposed to the previous date.</p> + +<p>From the same line, parse the three counts <code>READ</code>, <code>WRITE</code>, and <code>BOTH</code>, but disregard the <code>BELOW</code> value. +Discard any statistics where the sum of these three values is 0. +Compute three fractions by dividing each of the three values <code>READ</code>, <code>WRITE</code>, and <code>BOTH</code> by the sum of all three. +Multiply results with 100 and truncate any decimal places, keeping a fraction value between 0 and 100.</p> + +<h4>Step 2: Aggregate statistics</h4> + +<p>For each date, compute the 25th, 50th, and 75th percentile of fractions computed in the previous step.</p> + +<p>We use a non-standard percentile definition that is similar to the nearest-rank method: +the P-th percentile (<code>0 < P ≤ 100</code>) of a list of <code>N</code> ordered values (sorted from least to greatest) is the <em>largest</em> value in the list such that no more than <code>P</code> percent of the data is strictly less than the value and at least <code>P</code> percent of the data is less than or equal to that value. +We calculate the ordinal rank <code>n</code> using the following formula: <code>floor((P / 100) * N) + 1</code></p> + +</div> + +<div class="container"> +<h2><i class="fa fa-dashboard fa-fw" aria-hidden="true"></i> +Performance <a href="#performance" name="performance" class="anchor">#</a></h2> + +<p>We perform active measurements of Tor network performance by running several <a href="https://github.com/robgjansen/onionperf">OnionPerf</a> (previously: <a href="https://gitweb.torproject.org/torperf.git">Torperf</a>) instances from different vantage points. +Here we explain how we evaluate Torperf/OnionPerf measurement to obtain the same results as on Tor Metrics.</p> + + + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Time to download files over Tor <a href="/torperf.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Timeouts and failures of downloading files over Tor <a href="/torperf-failures.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Circuit build times <a href="/onionperf-buildtimes.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Circuit round-trip latencies <a href="/onionperf-latencies.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse OnionPerf and/or Torperf measurement results</h4> + +<p>Obtain OnionPerf/Torperf measurement results from <a href="/collector.html#type-torperf">CollecTor</a>.</p> + +<p>From each measurement result, parse the following keys:</p> + +<ul> +<li><code>SOURCE</code>: Configured name of the data source.</li> +<li><code>FILESIZE</code>: Configured file size in bytes.</li> +<li><code>START</code>: Download start time that we use for two purposes: to determine how long a request took and to aggregate measurements by date.</li> +<li><code>DATAREQUEST</code>: Time when the HTTP request was sent.</li> +<li><code>DATARESPONSE</code>: Time when the HTTP response header was received.</li> +<li><code>DATACOMPLETE</code>: Download end time that is only set if the request succeeded.</li> +<li><code>READBYTES</code>: Total number of bytes read, which indicates whether this request succeeded (if ≥ <code>FILESIZE</code>) or failed.</li> +<li><code>DIDTIMEOUT</code>: 1 if the request timed out, 0 otherwise.</li> +<li><code>BUILDTIMES</code>: Comma-separated list of times when circuit hops were built, which includes all circuits used for making measurement requests, successful or not.</li> +<li><code>ENDPOINTREMOTE</code>: Hostname, IP address, and port that was used to connect to the remote server; we use this to distinguish a request to a public server (if <code>ENDPOINTREMOTE</code> is not present or does not contain <code>".onion"</code> as substring) or to an onion server.</li> +</ul> + +<h4>Step 2: Aggregate measurement results</h4> + +<p>Each of the measurement results parsed in the previous steps constitutes a single measurement. +We're first interested in statistics on download times for the <a href="/torperf.html">Time to download files over Tor</a> graph. +Therefore we consider only measurements with <code>DATACOMPLETE > START</code>, for which we calculate the download time as: <code>DATACOMPLETE - START</code>. +We then compute the 25th, 50th, and 75th percentile of download times by sorting download times, determining the percentile rank, and using linear interpolation between adjacent ranks.</p> + +<p>We're also interested in circuit round-trip latencies for the <a href="/onionperf-latencies.html">Circuit round-trip latencies</a> graph. +We measure circuit latency as the time between sending the HTTP request and receiving the HTTP response header. +We calculate latencies as <code>DATARESPONSE - DATAREQUEST</code> for measurements with non-zero values for both timestamps. +We then compute 25th, 50th, and 75th percentiles in the same way as for download times above.</p> + +<p>Ideally, all measurements would succeed. +But it's also possible that some measurements did not complete within a pre-defined timeout or failed for some other reason. +We distinguish three cases for the <a href="/torperf-failures.html">Timeouts and failures of downloading files over Tor</a> graph and provide counts of each case per day:</p> + +<ul> +<li>Timeouts: measurements that either timed out (with <code>DIDTIMEOUT = 1</code>) or that have an invalid measurement end time (<code>DATACOMPLETE ≤ START</code>),</li> +<li>Failures: measurements that did not time out (with <code>DIDTIMEOUT = 0</code>), that had a valid measurement end time (<code>DATACOMPLETE > START</code>), and that had fewer bytes read than expected (<code>READBYTES < FILESIZE</code>), and</li> +<li>Requests: all measurements, including successes, timeouts, and failures.</li> +</ul> + +<p>The fourth metric that we obtain from OnionPerf/Torperf measurements is <a href="/glossary.html#circuit">circuit</a> build time, which is shown in the <a href="/onionperf-buildtimes.html">Circuit build times</a> graph. +We extract circuit build times from the <code>BUILDTIMES</code> field included in measurement results. +We use the first value as build time for the first hop and deltas between subsequent values as build times for the second and third hop. +Again, we compute the 25th, 50th, and 75th percentiles of these build times in the same way as for download times and circuit round-trip latencies.</p> + +</div> + +<div class="container"> +<h2><i class="fa fa-map-signs fa-fw" aria-hidden="true"></i> +Onion Services <a href="#onion-services" name="servers" class="anchor">#</a></h2> + +<p>Our <a href="/glossary.html#onion-service">onion services</a> statistics are based on two statistics reported by relays that have been added in 2014 to give some first insights into onion-service usage. +For further background on the following steps, refer to the technical report titled <a href="https://research.torproject.org/techreports/extrapolating-hidserv-stats-201…">"Extrapolating network totals from hidden-service statistics"</a> that this description is based on (which was written before hidden services were renamed to onion services).</p> + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Unique .onion addresses (version 2 only) <a href="/hidserv-dir-onions-seen.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Onion-service traffic (versions 2 and 3) <a href="/hidserv-rend-relayed-cells.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Fraction of relays reporting onion-service statistics <a href="/hidserv-frac-reporting.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse reported statistics from extra-info descriptors</h4> + +<p>Obtain relay extra-info descriptors from <a href="/collector.html#type-extra-info">CollecTor</a>.</p> + +<p>Parse the following parts from each extra-info descriptor:</p> + +<ul> +<li>Relay fingerprint: The <code>"extra-info"</code> line tells us which relay reported these statistics, which we need to know to match them with the expected fraction of onion-service activity throughout the statistics interval.</li> +<li>Onion service statistics interval end: The <code>"hidserv-stats-end"</code> line tells us when the statistics interval ended, and, together with the interval length, when it started.</li> +<li>Cells relayed as rendezvous point: The <code>"hidserv-rend-relayed-cells"</code> line tells us the number of cells that the relay handled on rendezvous circuits, and it tells us how this number has been obfuscated by the relay. +The value for <code>"bin_size"</code> is the bin size used for rounding up the originally observed cell number, and the values for <code>"delta_f"</code> and <code>"epsilon"</code> are inputs for the additive noise following a Laplace distribution.</li> +<li>.onion addresses observed as directory: Finally, the <code>"hidserv-dir-onions-seen"</code> line tells us the number of .onion addresses that the relay observed in published onion-service descriptors in its role as onion-service directory.</li> +</ul> + +<p>Note: Unlike other statistics, we're not splitting statistics by UTC date. Instead, we're only accepting statistics intervals that are exactly 1 day long, and we're counting all reported values for the UTC date of the statistics end time.</p> + +<h4>Step 2: Remove previously added noise</h4> + +<p>When processing onion-service statistics, we need to handle the fact that they have been obfuscated by relays. +As first step, we're attempting to remove the additive Laplace-distributed noise by rounding up to the nearest multiple of <code>bin_size</code>. +The idea is that it's most likely that noise was added to the closest right side of a bin than to the right side of another bin. +In step two, we're subtracting half of <code>bin_size</code>, because the relay added between 0 and <code>bin_size - 1</code> to the originally observed value. +All in all, we're using the following formula to remove previously added noise:</p> + +<pre> +halfBin = binSize / 2 +floor((reported + halfBin) / binSize) * binSize - halfBin +</pre> + +<h4>Step 3: Parse consensuses</h4> + +<p>Obtain consensuses from <a href="/collector.html#type-network-status-consensus-3">CollecTor</a>.</p> + +<p>From each consensus, parse the <code>"valid-after"</code> time from the header section.</p> + +<p>From each consensus entry, parse the base64-encoded relay fingerprint from the <code>"r"</code> line.</p> + +<p>Also parse the <a href="/glossary.html#relay-flag">relay flags</a> from the <code>"s"</code> line. If there is no <code>"Running"</code> flag, skip this entry. +(Consensuses with consensus method 4, introduced in 2008, or later do not list non-running relays, so that checking relay flags in recent consensuses is mostly done as a precaution without actual effect on the parsed data.) +Parse the remaining relay flags from this line.</p> + +<p>Finally, parse the weights contained in the <code>"bandwidth-weights"</code> line from the footer section of the consensus.</p> + +<h4>Step 4: Derive network fractions from consensuses</h4> + +<p>The probability of choosing a relay as rendezvous point varies a lot between relays, and not all onion-service directories handle the same number of onion-service descriptors. +Fortunately, we can derive what fraction of rendezvous circuits a relay has handled and what fraction of descriptors a directory was responsible for.</p> + +<p>The first fraction that we compute is the probability of a relay to be selected as rendezvous point. +<a href="/glossary.html#client">Clients</a> only select relays as rendezvous point that have the <code>"Fast"</code> flag. +They weight relays differently based on their bandwidth and depending on whether they have the <code>"Exit"</code> and/or <code>"Guard"</code> flags: +they weight the bandwidth value contained in the <code>"w"</code> line with the value of <code>"Wmg"</code>, <code>"Wme"</code>, <code>"Wmd"</code>, or <code>"Wmm"</code>, depending on whether the relay has only the <code>"Guard"</code> flag, only the <code>"Exit"</code> flag, both such flags, or neither of them.</p> + +<p>The second fraction that we can derive from this consensus entry is the fraction of descriptor space that this relay was responsible for in its role as onion-service directory. +The <a href="https://gitweb.torproject.org/torspec.git/tree/rend-spec-v2.txt">Tor Rendezvous Specification</a> contains the following definition: +<em>"A[n onion] service directory is deemed responsible for a descriptor ID if it has the HSDir flag and its identity digest is one of the first three identity digests of HSDir relays following the descriptor ID in a circular list."</em></p> + +<p>Based on the fraction of descriptor space that a directory was responsible for we can compute the fraction of descriptors that this directory has seen. +Intuitively, one might think that these fractions are the same. +However, this is not the case: each descriptor that is published to a directory is also published to two other directories. +As a result we need to divide the fraction of descriptor space by <em>three</em> to obtain the fraction of descriptors observed the directory. +Note that, without dividing by three, fractions of all directories would not add up to 100%.</p> + +<p>We calculate network fraction per consensus. +When we extrapolate reported statistics, we compute the average (arithmetic mean) of all such network fractions with consensus valid-after times falling into the statistics interval. +In particular, we're <em>not</em> computing the average of network fractions from the UTC day when the statistics interval ends; even though we're attributing extrapolated statistics to the UTC date of the statistics interval end in the next step.</p> + +<h4>Step 5: Extrapolate network totals</h4> + +<p>We are now ready to extrapolate network totals from reported statistics. +We do this by dividing reported statistics by the calculated fraction of observations made by the reporting relay. +The underlying assumption is that statistics grow linearly with calculated fractions. +We only exclude relays from this step that have a calculated fraction of exactly zero, to avoid dividing by zero.</p> + +<p>While we can expect this method to work as described for extrapolating cells on rendezvous circuits, we need to take another step for estimating the number of unique .onion addresses in the network. +The reason is that a .onion address is not only known to a single relay, but to a couple of relays, all of which include that .onion address in their statistics. +We need to subtract out the multiple counting of .onion addresses to come up with a network-wide number of unique .onion addresses.</p> + +<p>As an approximation, we assume that an <a href="/glossary.html#onion-service">onion service</a> publishes its descriptor to <em>twelve</em> directories over a 24-hour period: +the service stores <em>two</em> replicas per descriptor using different descriptor identifiers, both descriptor replicas get stored to <em>three</em> different onion-service directories each, and the service changes descriptor identifiers once every 24 hours which leads to <em>two</em> different descriptor identifiers per replica.</p> + +<p>To be clear, this approximation is not entirely accurate. +For example, the two replicas or the descriptors with changed descriptor identifiers could have been stored to the same directory. +As another example, onion-service directories might have joined or left the network and other directories might have become responsible for storing a descriptor which also include that .onion address in their statistics. +However, for the subsequent analysis, we assume that neither of these cases affects results substantially.</p> + +<h4>Step 6: Compute daily averages</h4> + +<p>As last step in the analysis, we aggregate extrapolated network totals from all reporting relays to obtain a daily average. +We're using the weighted interquartile mean as metric, because it is robust against noisy statistics and potentially lying relays and considers half of the reported statistics. +For this metric we order extrapolated network totals by their value, discard the lower and the upper quartile by weight, and compute the weighted mean of the remaining values. +</p> + +<p>We further define a threshold of 1% for the total fraction of relays reporting statistics. +If less than these 1% of relays report statistics on a given day, we don't include that day in the end results.</p> + +</div> + +<div class="container"> +<h2><i class="fa fa-download fa-fw" aria-hidden="true"></i> +Applications <a href="#applications" name="applications" class="anchor">#</a></h2> + +<p>Our applications statistics are based on Tor web server requests where our users download applications initially and where they ask for updates.</p> + +<p>The following description applies to the following graphs:</p> + +<ul> +<li>Tor Browser downloads and updates <a href="/webstats-tb.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Tor Browser downloads by platform <a href="/webstats-tb-platform.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Tor Browser downloads by locale <a href="/webstats-tb-locale.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +<li>Tor Messenger downloads and updates <a href="/webstats-tm.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> +</ul> + +<h4>Step 1: Parse Tor web server logs</h4> + +<p>Obtain Tor web server logs from <a href="/collector.html#type-webstats">CollecTor</a>. +Refer to the separate <a href="/web-server-logs.html">specification page</a> for details on the data format.</p> + +<p>Each log file contains relevant meta data in its file name, including the site name, the server name, and the log date. +The log file itself contains sanitized requests to Tor web servers.</p> + +<p>All patterns mentioned in the following are understood by PostgreSQL's <code>LIKE</code> operator. +An underscore (<code>_</code>) matches any single character; a percent sign (<code>%</code>) matches any string of zero or more characters.</p> + +<h4>Step 2: Count Tor Browser initial downloads</h4> + +<p>We count a request as Tor Browser initial download if it matches the following criteria:</p> + +<ul> +<li>Request method: GET</li> +<li>Resource string: <code>'%/torbrowser/%.exe'</code>, <code>'%/torbrowser/%.dmg'</code>, or <code>'%/torbrowser/%.tar.xz'</code></li> +<li>Response code: 200</li> +</ul> + +<p>We distinguish platforms based on the resource string: <code>'%.exe%'</code> for Windows, <code>'%.dmg%'</code> for macOS, and <code>'%.tar.xz%'</code> for Linux.</p> + +<p>We distinguish release channels based on the resource string: <code>'%-hardened%'</code> for hardened releases, <code>'%/%.%a%/%'</code> for alpha releases, and stable releases otherwise.</p> + +<p>We extract the locale (for example, <code>'en-US'</code> for English as used in the United States or <code>'de'</code> for German) from the resource string using regular expression <code>'.*_([a-zA-Z]{2}|[a-zA-Z]{2}-[a-zA-Z]{2})[\._-].*'</code>, falling back to <code>'??'</code> for unrecognized locales if the regular expression does not match.</p> + +<h4>Step 3: Count Tor Browser signature downloads</h4> + +<p>We count a request as Tor Browser signature download if it matches the following criteria:</p> + +<ul> +<li>Request method: GET</li> +<li>Resource string: <code>'%/torbrowser/%.exe.asc'</code>, <code>'%/torbrowser/%.dmg.asc'</code>, or <code>'%/torbrowser/%.tar.xz.asc'</code></li> +<li>Response code: 200</li> +</ul> + +<p>We break down requests by platform, channel, and locale in the exact same way as for Tor Browser initial downloads (see above).</p> + +<h4>Step 4: Count Tor Browser update pings</h4> + +<p>We count a request as Tor Browser update ping if it matches the following criteria:</p> + +<ul> +<li>Request method: GET</li> +<li>Resource string: <code>'%/torbrowser/update\__/%'</code> but not <code>'%.xml'</code></li> +<li>Response code: 200</li> +</ul> + +<p>We distinguish platforms based on the resource string: <code>'%/WINNT%'</code> for Windows, <code>'%/Darwin%'</code> for macOS, and Linux otherwise.</p> + +<p>We distinguish release channels based on the resource string: <code>'%/hardened/%'</code> for hardened releases, <code>'%/alpha/%'</code> for alpha releases, and <code>'%/release/%'</code> for stable releases.</p> + +<p>We extract the locale (for example, <code>'en-US'</code> for English as used in the United States or <code>'de'</code> for German) from the resource string using regular expression <code>'.*/([a-zA-Z]{2}|[a-zA-Z]{2}-[a-zA-Z]{2})\??$'</code>, falling back to <code>'??'</code> for unrecognized locales if the regular expression does not match.</p> + +<h4>Step 5: Count Tor Browser update requests</h4> + +<p>We count a request as Tor Browser update request if it matches the following criteria:</p> + +<ul> +<li>Request method: GET</li> +<li>Resource string: <code>'%/torbrowser/%.mar'</code></li> +<li>Response code: 302</li> +</ul> + +<p>We distinguish platforms based on the resource string: <code>'%-win32-%'</code> for Windows, <code>'%-osx%'</code> for macOS, and Linux otherwise.</p> + +<p>We distinguish release channels based on the resource string: <code>'%-hardened%'</code> for hardened releases, <code>'%/%.%a%/%'</code> for alpha releases, and stable releases otherwise.</p> + +<p>We extract the locale (for example, <code>'en-US'</code> for English as used in the United States or <code>'de'</code> for German) from the resource string using regular expression <code>'.*_([a-zA-Z]{2}|[a-zA-Z]{2}-[a-zA-Z]{2})[\._-].*'</code>, falling back to <code>'??'</code> for unrecognized locales if the regular expression does not match.</p> + +<p>We distinguish incremental updates having <code>'%.incremental.%'</code> in the resource string from non-incremental (full) updates that don't contain this pattern in their resource string.</p> + +<h4>Step 6: Count Tor Messenger initial downloads</h4> + +<p>We count a request as Tor Messenger initial download if it matches the following criteria:</p> + +<ul> +<li>Request method: GET</li> +<li>Resource string: <code>'%/tormessenger/%.exe'</code>, <code>'%/tormessenger/%.dmg'</code>, and <code>'%/tormessenger/%.tar.xz'</code></li> +<li>Response code: 200</li> +</ul> + +<p>We distinguish platforms based on the resource string: <code>'%.exe'</code> for Windows, <code>'%.dmg'</code> for macOS, and <code>'%.tar.xz'</code> for Linux.</p> + +<p>We extract the locale (for example, <code>'en-US'</code> for English as used in the United States or <code>'de'</code> for German) from the resource string using regular expression <code>'.*_([a-zA-Z]{2}|[a-zA-Z]{2}-[a-zA-Z]{2})[\._-].*'</code>, falling back to <code>'??'</code> for unrecognized locales if the regular expression does not match.</p> + +<h4>Step 7: Count Tor Messenger update pings</h4> + +<p>We count a request as Tor Messenger update ping if it matches the following criteria:</p> + +<ul> +<li>Request method: GET</li> +<li>Resource string: <code>'%/tormessenger/update\__/%'</code> but none of <code>'%.xml'</code>, <code>'%/'</code> or <code>'%/?'</code></li> +<li>Response code: 200</li> +</ul> + +<p>We distinguish platforms based on the resource string: <code>'%/WINNT%'</code> for Windows, <code>'%/Darwin%'</code> for macOS, and <code>'%/Linux%'</code> for Linux.</p> + +<p>We extract the locale (for example, <code>'en-US'</code> for English as used in the United States or <code>'de'</code> for German) from the resource string using regular expression <code>'.*/([a-zA-Z]{2}|[a-zA-Z]{2}-[a-zA-Z]{2})\??$'</code>, falling back to <code>'??'</code> for unrecognized locales if the regular expression does not match.</p> + +</div> + + + + + +<jsp:include page="bottom.jsp"/> +

1 0

[metrics-web/release] Update to latest metrics-base.
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 963da56fa8db87a993561bc187e63b64d4881b29 Author: Karsten Loesing <karsten.loesing(a)gmx.net> Date: Tue Jul 31 13:32:57 2018 +0200 Update to latest metrics-base. --- src/build | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/build b/src/build index 23c6e0b..08514a3 160000 --- a/src/build +++ b/src/build @@ -1 +1 @@ -Subproject commit 23c6e0be5fab9463f137615053ef412e4da2315e +Subproject commit 08514a32afefbeef848b80f9a338ee840c282604

1 0

[metrics-web/release] Fix bug in generating Tor Messenger CSV file.
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit afc2fc4fa5f19b592dd8f09d6397fd244adfbe91 Author: Karsten Loesing <karsten.loesing(a)gmx.net> Date: Fri Aug 3 10:32:57 2018 +0200 Fix bug in generating Tor Messenger CSV file. --- src/main/R/rserver/graphs.R | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/main/R/rserver/graphs.R b/src/main/R/rserver/graphs.R index 56c8a55..12a80e9 100644 --- a/src/main/R/rserver/graphs.R +++ b/src/main/R/rserver/graphs.R @@ -1489,7 +1489,8 @@ plot_webstats_tm <- function(start_p, end_p, path_p) { write_webstats_tm <- function(start_p = NULL, end_p = NULL, path_p) { prepare_webstats_tm(start_p, end_p) %>% rename(date = log_date) %>% - spread(request_type, count) %>% + mutate(request_type = factor(request_type, levels = c("tmid", "tmup"))) %>% + spread(request_type, count, drop = FALSE) %>% rename(initial_downloads = tmid, update_pings = tmup) %>% write.csv(path_p, quote = FALSE, row.names = FALSE, na = "") }

1 0

[metrics-web/release] Document Onionoo versions 6.2 and 7.0.
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 5618a8ff342f06a00ab2282b7e823993f099ad11 Author: Karsten Loesing <karsten.loesing(a)gmx.net> Date: Thu Aug 2 22:26:48 2018 +0200 Document Onionoo versions 6.2 and 7.0. --- src/main/resources/web/jsps/onionoo.jsp | 93 ++++++++++++++++++++++++++++----- 1 file changed, 81 insertions(+), 12 deletions(-) diff --git a/src/main/resources/web/jsps/onionoo.jsp b/src/main/resources/web/jsps/onionoo.jsp index 84bc87c..7c78c7c 100644 --- a/src/main/resources/web/jsps/onionoo.jsp +++ b/src/main/resources/web/jsps/onionoo.jsp @@ -316,6 +316,19 @@ documents in favor of the new "verified_host_names" and "unverified_host_names" fields for more accurate DNS results on July 16, 2018. <a href="#versions_6_1" class="anchor">#</a></li> +<li><a id="versions_6_2"></a><strong>6.2</strong>: +Added an "as" field to details document, deprecated the "as_number" field, added +an "as_name" parameter to search for relays by AS name, and added support for +comma-separated lists in the "as" and the "lookup" parameters on August 3, 2018. +<a href="#versions_6_2" class="anchor">#</a></li> +<li><a id="versions_7_0"></a><strong>7.0</strong>: +(scheduled, but not deployed yet!): Extend the "version" parameter to support +lists and ranges, remove redundant "1_week" and "1_month" graphs from clients +documents, change "3_months" graphs to "6_months" graphs in all documents +containing history objects, remove the "fingerprint" parameter, and remove the +previously deprecated "as_number" field from details documents, to be deployed +after September 3, 2018. +<a href="#versions_7_0" class="anchor">#</a></li> </ul> @@ -472,7 +485,7 @@ Lookups are case-insensitive. <li> <a id="parameters_fingerprint"></a> -<b>fingerprint</b> +<b>fingerprint <span class="label label-warning">deprecated</span></b> <a href="#parameters_fingerprint" class="anchor">#</a> <p> Return only the relay with the parameter value matching the fingerprint @@ -484,6 +497,7 @@ with two exceptions: not</i> be hashed (again) using SHA-1; (2) the response will contain any matching relay or bridge regardless of whether they have been running in the past week. +<span class="red">Scheduled to be removed after September 3, 2018.</span> </p> </li> @@ -507,12 +521,32 @@ the GeoIP database. <b>as <span class="label label-primary">updated</span></b> <a href="#parameters_as" class="anchor">#</a> <p> -Return only relays which are located in the given autonomous system (AS) as -identified by the AS number (with or without preceding "AS" part). Filtering -by AS number is case-insensitive. The special AS number <tt>0</tt> can be used -for relays that were not found in the GeoIP database. +Return only relays which are located in either one of the given autonomous +systems (AS) as identified by AS number (with or without preceding "AS" part). +Multiple AS numbers can be provided separated by commas. +Filtering by AS number is case-insensitive. +The special AS number <tt>0</tt> can be used for relays that were not found in +the GeoIP database. <span class="blue">Updated to recognize special AS number <tt>0</tt> -on July 16, 2018.</span> +on July 16, 2018 and to support comma-separated lists of AS numbers on August 3, +2018.</span> +</p> +</li> + +<li> +<a id="parameters_as_name"></a> +<b>as_name <span class="label label-primary">new</span></b> +<a href="#parameters_as_name" class="anchor">#</a> +<p> +Return only relays with the parameter value matching (part of) the autonomous +system (AS) name they are located in. +If the parameter value contains spaces, only relays are returned which +contain all space-separated parts in their AS name. +Only printable ASCII characters are permitted in the parameter value, +some of which need to be percent-encoded (# as %23, % as %25, & as +%26, + as %2B, and / as %2F). +Comparisons are case-insensitive. +<span class="blue">Added on August 3, 2018.</span> </p> </li> @@ -596,12 +630,22 @@ a family. <li> <a id="parameters_version"></a> -<b>version</b> +<b>version <span class="label label-primary">updated</span></b> <a href="#parameters_version" class="anchor">#</a> <p> Return only relays or bridges running a Tor version that starts with the parameter value <i>without</i> leading <code>"Tor"</code>. Searches are case-insensitive. +<span class="blue">Scheduled to support lists and ranges after September 3, +2018: +Return only relays or bridges running either Tor version from a list or range +given in the parameter value. +Tor versions must be provided <i>without</i> the leading <code>"Tor"</code> +part. +Multiple versions can either be provided as a comma-separated list (","), as a +range separated by two dots (".."), or as a list of ranges. +Provided versions are parsed and matched by parsed dotted numbers, rather than +by string prefix.</span> </p> </li> @@ -1271,8 +1315,24 @@ database. </li> <li> +<a id="details_relay_as"></a> +<b>as <span class="label label-primary">new</span></b> +<code class="typeof">string</code> +<span class="required-false">optional</span> +<a href="#details_relay_as_number" class="anchor">#</a> +<p> +AS number as found in an AS database by +resolving the relay's first onion-routing IP address. +AS number strings start with "AS", followed directly by the AS number. +Omitted if the relay IP address could not be found in the AS +database. +<span class="blue">Added on August 3, 2018.</span> +</p> +</li> + +<li> <a id="details_relay_as_number"></a> -<b>as_number</b> +<b>as_number</b> <span class="label label-warning">deprecated</span> <code class="typeof">string</code> <span class="required-false">optional</span> <a href="#details_relay_as_number" class="anchor">#</a> @@ -1282,6 +1342,7 @@ resolving the relay's first onion-routing IP address. AS number strings start with "AS", followed directly by the AS number. Omitted if the relay IP address could not be found in the AS database. +<span class="red">Scheduled to be removed after September 3, 2018.</span> </p> </li> @@ -2088,7 +2149,7 @@ hexadecimal characters. <li> <a id="bandwidth_relay_write_history"></a> -<b>write_history</b> +<b>write_history <span class="label label-primary">updated</span></b> <code class="typeof">object</code> <span class="required-false">optional</span> <a href="#bandwidth_relay_write_history" class="anchor">#</a> @@ -2112,6 +2173,8 @@ The unit is bytes per second. Contained graph history objects may contain null values if the relay did not provide any bandwidth data or only data for less than 20% of a given time period. +<span class="blue">Scheduled to no longer contain a "3_months" graph and instead +a "6_months" graph after September 3, 2018.</span> </p> </li> @@ -2222,7 +2285,7 @@ hexadecimal characters. <li> <a id="weights_relay_consensus_weight_fraction"></a> -<b>consensus_weight_fraction</b> +<b>consensus_weight_fraction <span class="label label-primary">updated</span></b> <code class="typeof">object</code> <span class="required-false">optional</span> <a href="#weights_relay_consensus_weight_fraction" class="anchor">#</a> @@ -2245,6 +2308,8 @@ period and higher data resolution. The unit is path-selection probability. Contained graph history objects may contain null values if the relay was running less than 20% of a given time period. +<span class="blue">Scheduled to no longer contain a "3_months" graph and instead +a "6_months" graph after September 3, 2018.</span> </p> </li> @@ -2357,7 +2422,7 @@ of 40 upper-case hexadecimal characters. <li> <a id="clients_bridge_average_clients"></a> -<b>average_clients</b> +<b>average_clients <span class="label label-primary">updated</span></b> <code class="typeof">object</code> <span class="required-false">optional</span> <a href="#clients_bridge_average_clients" class="anchor">#</a> @@ -2377,6 +2442,8 @@ time period and higher data resolution. The unit is number of clients. Contained graph history objects may contain null values if the bridge did not report client statistics for at least 50% of a given time period. +<span class="blue">Scheduled to no longer contain "1_week", "1_month", and +"3_months" graphs and instead a "6_months" graph after September 3, 2018.</span> </p> </li> @@ -2420,7 +2487,7 @@ hexadecimal characters. <li> <a id="uptime_relay_uptime"></a> -<b>uptime</b> +<b>uptime <span class="label label-primary">updated</span></b> <code class="typeof">object</code> <span class="required-false">optional</span> <a href="#uptime_relay_uptime" class="anchor">#</a> @@ -2440,6 +2507,8 @@ period and higher data resolution. The unit is fractional uptime from 0 to 1. Contained graph history objects may contain null values if less than 20% of network statuses have been processed for a given time period. +<span class="blue">Scheduled to no longer contain a "3_months" graph and instead +a "6_months" graph after September 3, 2018.</span> </p> </li>

1 0

[metrics-web/release] Remove stray console.log
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 07520b1c3e07e0968162134d6297442acd034513 Author: Iain R. Learmonth <irl(a)fsfe.org> Date: Tue Aug 7 12:20:31 2018 +0100 Remove stray console.log --- src/main/resources/web/js/rs/views/search/do.js | 1 - 1 file changed, 1 deletion(-) diff --git a/src/main/resources/web/js/rs/views/search/do.js b/src/main/resources/web/js/rs/views/search/do.js index 5383580..53c80e6 100644 --- a/src/main/resources/web/js/rs/views/search/do.js +++ b/src/main/resources/web/js/rs/views/search/do.js @@ -52,7 +52,6 @@ define([ $(".tip").tooltip({'html':true}); }, "footerCallback": function( tfoot, data, start, end, display ) { - console.log(relays); var sumAdvertisedBandwidths = 0; for (var i = 0; i < relays.length; i++) { sumAdvertisedBandwidths += relays[i].get("advertised_bandwidth");

1 0

[metrics-web/release] Update news.json to version 284 of doc/MetricsTimeline.
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 33921ae7e07ad5ac34e4a7ba40b626f3eddcc6a2 Author: Karsten Loesing <karsten.loesing(a)gmx.net> Date: Tue Aug 7 14:37:48 2018 +0200 Update news.json to version 284 of doc/MetricsTimeline. --- src/main/resources/web/json/news.json | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/src/main/resources/web/json/news.json b/src/main/resources/web/json/news.json index f0440a6..03d7f6a 100644 --- a/src/main/resources/web/json/news.json +++ b/src/main/resources/web/json/news.json @@ -3291,6 +3291,31 @@ "target" : "https://metrics.torproject.org/userstats-bridge-country.html?start=2018-06-…" } ] }, { + "start" : "2018-07-29", + "ongoing" : true, + "places" : [ "bd" ], + "short_description" : "Protests over road safety in Dhaka, Bangladesh", + "description" : "Protests over road safety in Dhaka, Bangladesh. Reports of mobile network throttling and blocking of Facebook.", + "links" : [ { + "label" : "New York Times article", + "target" : "https://www.nytimes.com/2018/08/05/world/asia/bangladesh-students-protests.…" + }, { + "label" : "Daily Star article on throttling", + "target" : "https://www.thedailystar.net/country/bangladesh-mobile-internet-speed-broug…" + } ] +}, { + "start" : "2018-08-04", + "end" : "2018-08-05", + "short_description" : "Collector was missing consensuses", + "description" : "Collector was missing consensuses. They were recovered via recent relay consensus diffs.", + "links" : [ { + "label" : "metrics-team thread", + "target" : "https://lists.torproject.org/pipermail/metrics-team/2018-August/000869.html" + }, { + "label" : "tor-relays thread", + "target" : "https://lists.torproject.org/pipermail/tor-relays/2018-August/015850.html" + } ] +}, { "start" : "2016-02-24", "places" : [ "tm" ], "protocols" : [ "<OR>" ],

1 0

[metrics-web/release] Fixes loading JSON for aggregated map plotting
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 629a828e5296a999270edcecd8e8b2d7f0c81051 Author: Iain R. Learmonth <irl(a)fsfe.org> Date: Tue Aug 7 14:57:01 2018 +0100 Fixes loading JSON for aggregated map plotting --- build.xml | 2 +- src/main/resources/web/js/rs/views/aggregate/map.js | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/build.xml b/build.xml index 77ab672..2f06209 100644 --- a/build.xml +++ b/build.xml @@ -182,7 +182,7 @@ includes="logback.xml" /> <zipfileset dir="${resources}/web" prefix="" - includes="css/* images/** js/** fonts/** templates/**" /> + includes="css/* images/** js/** fonts/** templates/** json/countries.topo.json" /> <zipfileset dir="${resources}/web/jsps" prefix="WEB-INF" includes="*.jsp"/> diff --git a/src/main/resources/web/js/rs/views/aggregate/map.js b/src/main/resources/web/js/rs/views/aggregate/map.js index c77ec10..fe69bc0 100644 --- a/src/main/resources/web/js/rs/views/aggregate/map.js +++ b/src/main/resources/web/js/rs/views/aggregate/map.js @@ -121,7 +121,7 @@ define([ return text; } - d3.json("/rs/json/countries.topo.json", function(error, us) { + d3.json("/json/countries.topo.json", function(error, us) { g.append("g") .attr("id", "countries") .style("fill","#fff")

1 0

[metrics-web/release] Adds the experimental and outdated flags
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit a1c82166a89391d0a2a5dbfc22e777243c08cd31 Author: Iain R. Learmonth <irl(a)fsfe.org> Date: Tue Aug 7 17:45:21 2018 +0100 Adds the experimental and outdated flags Fixes: #26667 --- src/main/resources/web/images/flags/experimental.png | Bin 0 -> 346 bytes src/main/resources/web/images/flags/outdated.png | Bin 0 -> 304 bytes src/main/resources/web/js/rs/models/relay.js | 13 +++++++++++++ 3 files changed, 13 insertions(+) diff --git a/src/main/resources/web/images/flags/experimental.png b/src/main/resources/web/images/flags/experimental.png new file mode 100644 index 0000000..cc21254 Binary files /dev/null and b/src/main/resources/web/images/flags/experimental.png differ diff --git a/src/main/resources/web/images/flags/outdated.png b/src/main/resources/web/images/flags/outdated.png new file mode 100644 index 0000000..8509108 Binary files /dev/null and b/src/main/resources/web/images/flags/outdated.png differ diff --git a/src/main/resources/web/js/rs/models/relay.js b/src/main/resources/web/js/rs/models/relay.js index cdbe3a3..b5a6b8e 100644 --- a/src/main/resources/web/js/rs/models/relay.js +++ b/src/main/resources/web/js/rs/models/relay.js @@ -62,6 +62,12 @@ define([ if (flag == "Not Recommended") { output.push([flag,"notrecommended", "This relay is running a Tor version that is not recommended by the directory authorities and may contain known issues. This includes both obsolete and experimental versions."]); } + if (flag == "Outdated") { + output.push([flag,"outdated", "This relay is running a Tor version that is considered obsolete. If this is your relay then you should upgrade at the earliest opportunity."]); + } + if (flag == "Experimental") { + output.push([flag,"experimental", "This relay is running a Tor version that is considered experimental. Please report any bugs found. If this is not intentional, you may consider switching to the most recent release instead."]); + } if (flag == "Unmeasured") { output.push([flag,"unmeasured", "This relay has not been measured by at least 3 bandwidth authorities and so its consensus weight is currently capped. This is expected for new relays."]); } @@ -236,10 +242,17 @@ define([ relay.flags = model.parseflags(relay.flags, relay.is_bridge); relay.version_consistent = relay.version == relay.platform.split(" ")[1]; + if (relay.version_consistent) { + relay.version_status = relay.version_status ? relay.version_status : "recommended"; + } else { + relay.version_status = "recommended"; + } /* Synthetic Additional Flags */ var additional_flags = [] if (!((typeof relay.recommended_version !== 'undefined') ? relay.recommended_version : true) && relay.version_consistent) additional_flags.push("Not Recommended"); + if (relay.version_status === 'obsolete') additional_flags.push("Obsolete"); + if (relay.version_status === 'experimental') additional_flags.push("Experimental"); if (!((typeof relay.measured !== 'undefined') ? relay.measured : true)) additional_flags.push("Unmeasured"); if (((typeof relay.hibernating !== 'undefined') ? relay.hibernating : false)) additional_flags.push("Hibernating"); if (IsFallbackDir(relay.fingerprint)) additional_flags.push("FallbackDir");

1 0

[metrics-web/release] Remove hidserv-frac-reporting graph and CSV file.
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit ec86ad02973963e9efad46b0ac21ae4f04c87fc6 Author: Karsten Loesing <karsten.loesing(a)gmx.net> Date: Thu Aug 9 16:13:45 2018 +0200 Remove hidserv-frac-reporting graph and CSV file. This was the graph showing the "Fraction of relays reporting onion-service statistics". Fraction values will still be available as part of the other onion-service statistics' per-graph CSV files. Implements #26950. --- src/main/R/rserver/graphs.R | 34 ---------------------- src/main/resources/web.xml | 4 --- src/main/resources/web/json/categories.json | 3 +- src/main/resources/web/json/metrics.json | 11 ------- .../resources/web/jsps/reproducible-metrics.jsp | 1 - src/main/resources/web/jsps/stats.jsp | 25 ---------------- 6 files changed, 1 insertion(+), 77 deletions(-) diff --git a/src/main/R/rserver/graphs.R b/src/main/R/rserver/graphs.R index ab37a32..66a1414 100644 --- a/src/main/R/rserver/graphs.R +++ b/src/main/R/rserver/graphs.R @@ -1287,40 +1287,6 @@ write_hidserv_rend_relayed_cells <- function(start_p = NULL, end_p = NULL, write.csv(path_p, quote = FALSE, row.names = FALSE, na = "") } -prepare_hidserv_frac_reporting <- function(start_p, end_p) { - read.csv(paste(stats_dir, "hidserv.csv", sep = ""), - colClasses = c("date" = "Date")) %>% - filter(if (!is.null(start_p)) date >= as.Date(start_p) else TRUE) %>% - filter(if (!is.null(end_p)) date <= as.Date(end_p) else TRUE) %>% - select(date, frac, type) -} - -plot_hidserv_frac_reporting <- function(start_p, end_p, path_p) { - prepare_hidserv_frac_reporting(start_p, end_p) %>% - ggplot(aes(x = date, y = frac, colour = type)) + - geom_line() + - geom_hline(yintercept = 0.01, linetype = 2) + - scale_x_date(name = "", breaks = custom_breaks, - labels = custom_labels, minor_breaks = custom_minor_breaks) + - scale_y_continuous(name = "", labels = percent, limits = c(0, NA)) + - scale_colour_hue(name = "", - breaks = c("rend-relayed-cells", "dir-onions-seen"), - labels = c("Onion-service traffic", - "Unique .onion addresses")) + - ggtitle(paste("Fraction of relays reporting onion-service", - "statistics")) + - labs(caption = copyright_notice) + - theme(legend.position = "top") - ggsave(filename = path_p, width = 8, height = 5, dpi = 150) -} - -write_hidserv_frac_reporting <- function(start_p = NULL, end_p = NULL, path_p) { - prepare_hidserv_frac_reporting(start_p, end_p) %>% - mutate(type = ifelse(type == "dir-onions-seen", "onions", "relayed")) %>% - spread(type, frac) %>% - write.csv(path_p, quote = FALSE, row.names = FALSE, na = "") -} - prepare_webstats_tb <- function(start_p, end_p) { load(paste(rdata_dir, "webstats-tb.RData", sep = "")) data %>% diff --git a/src/main/resources/web.xml b/src/main/resources/web.xml index ae252da..5c88564 100644 --- a/src/main/resources/web.xml +++ b/src/main/resources/web.xml @@ -48,7 +48,6 @@ <url-pattern>/connbidirect.html</url-pattern> <url-pattern>/hidserv-dir-onions-seen.html</url-pattern> <url-pattern>/hidserv-rend-relayed-cells.html</url-pattern> - <url-pattern>/hidserv-frac-reporting.html</url-pattern> <url-pattern>/webstats-tb.html</url-pattern> <url-pattern>/webstats-tb-platform.html</url-pattern> <url-pattern>/webstats-tb-locale.html</url-pattern> @@ -190,9 +189,6 @@ <url-pattern>/hidserv-rend-relayed-cells.png</url-pattern> <url-pattern>/hidserv-rend-relayed-cells.pdf</url-pattern> <url-pattern>/hidserv-rend-relayed-cells.csv</url-pattern> - <url-pattern>/hidserv-frac-reporting.png</url-pattern> - <url-pattern>/hidserv-frac-reporting.pdf</url-pattern> - <url-pattern>/hidserv-frac-reporting.csv</url-pattern> <url-pattern>/webstats-tb.png</url-pattern> <url-pattern>/webstats-tb.pdf</url-pattern> <url-pattern>/webstats-tb.csv</url-pattern> diff --git a/src/main/resources/web/json/categories.json b/src/main/resources/web/json/categories.json index 23e6ee4..a5a4fcf 100644 --- a/src/main/resources/web/json/categories.json +++ b/src/main/resources/web/json/categories.json @@ -74,8 +74,7 @@ "description": "Onion services are services that are only accessible via the Tor network.", "metrics": [ "hidserv-dir-onions-seen", - "hidserv-rend-relayed-cells", - "hidserv-frac-reporting" + "hidserv-rend-relayed-cells" ] }, { diff --git a/src/main/resources/web/json/metrics.json b/src/main/resources/web/json/metrics.json index 0f85a28..ab5d98d 100644 --- a/src/main/resources/web/json/metrics.json +++ b/src/main/resources/web/json/metrics.json @@ -367,17 +367,6 @@ ] }, { - "id": "hidserv-frac-reporting", - "title": "Fraction of relays reporting onion-service statistics", - "type": "Graph", - "description": "<p>This graph shows the fraction of <a href=\"glossary.html#relay\">relays</a> that report statistics on <a href=\"glossary.html#onion-service\">onion service</a> usage. If at least 1% of relays report a statistic, it gets extrapolated towards a network total, where higher fractions are produce more accurate results. For more details on the extrapolation algorithm, see <a href=\"https://blog.torproject.org/blog/some-statistics-about-onions\">this blog post</a> and <a href=\"https://research.torproject.org/techreports/extrapolating-hidserv-stats-201…">this technical report</a>.</p>", - "function": "hidserv_frac_reporting", - "parameters": [ - "start", - "end" - ] - }, - { "id": "uncharted-data-flow", "title": "“TorFlow”", "type": "Link", diff --git a/src/main/resources/web/jsps/reproducible-metrics.jsp b/src/main/resources/web/jsps/reproducible-metrics.jsp index dc00751..87f692e 100644 --- a/src/main/resources/web/jsps/reproducible-metrics.jsp +++ b/src/main/resources/web/jsps/reproducible-metrics.jsp @@ -671,7 +671,6 @@ For further background on the following steps, refer to the technical report tit <ul> <li>Unique .onion addresses (version 2 only) <a href="/hidserv-dir-onions-seen.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> <li>Onion-service traffic (versions 2 and 3) <a href="/hidserv-rend-relayed-cells.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> -<li>Fraction of relays reporting onion-service statistics <a href="/hidserv-frac-reporting.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a></li> </ul> <h4>Step 1: Parse reported statistics from extra-info descriptors</h4> diff --git a/src/main/resources/web/jsps/stats.jsp b/src/main/resources/web/jsps/stats.jsp index 719176c..565f235 100644 --- a/src/main/resources/web/jsps/stats.jsp +++ b/src/main/resources/web/jsps/stats.jsp @@ -633,31 +633,6 @@ Onion Services <a href="#onion-services" name="servers" class="anchor">#</a></h2 <li><b>frac:</b> Total network fraction of statistics reported by rendezvous points.</li> </ul> -<h3>Fraction of relays reporting onion-service statistics -<a href="/hidserv-frac-reporting.html" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> graph</a> -<a href="/hidserv-frac-reporting.csv" class="btn btn-primary btn-xs"><i class="fa fa-chevron-right" aria-hidden="true"></i> data</a> -<a href="/hidserv-frac-reporting">#</a></h3> - -<h4>Parameters</h4> - -<ul> -<li><b>start:</b> First UTC date (YYYY-MM-DD) to include in the file.</li> -<li><b>end:</b> Last UTC date (YYYY-MM-DD) to include in the file.</li> -</ul> - -<h4>Columns</h4> - -<ul> -<li><b>date:</b> UTC date (YYYY-MM-DD) when relays have been listed as running.</li> -<li><b>onions:</b> Total network fraction of statistics reported by onion-service directories.</li> -<li><b>relayed:</b> Total network fraction of statistics reported by rendezvous points.</li> -</ul> - -<div class="bs-callout bs-callout-warning"> -<h3>Suggested change</h3> -<p>Remove section and data as suggested on <a href="https://bugs.torproject.org/26950">#26950</a>.</p> -</div> - </div> <div class="container">

1 0

[metrics-web/release] Display Onionoo protocol version and build rev
by karsten＠torproject.org 26 Sep '18

26 Sep '18

commit 0ac652aec6c9d42ceb29867ed49f2820ec1f4667 Author: Iain R. Learmonth <irl(a)fsfe.org> Date: Tue Aug 7 14:57:43 2018 +0100 Display Onionoo protocol version and build rev Fixes: #26863 --- src/main/resources/web/js/rs/collections/aggregates.js | 5 ++++- src/main/resources/web/js/rs/collections/results.js | 5 ++++- src/main/resources/web/js/rs/models/relay.js | 2 ++ src/main/resources/web/js/rs/router.js | 17 +++++++++++++---- src/main/resources/web/js/rs/views/aggregate/map.js | 2 ++ src/main/resources/web/js/rs/views/aggregate/search.js | 2 ++ src/main/resources/web/js/rs/views/search/do.js | 2 ++ src/main/resources/web/templates/rs/aggregate/map.html | 2 ++ .../resources/web/templates/rs/aggregate/search.html | 1 + src/main/resources/web/templates/rs/details/bridge.html | 1 + src/main/resources/web/templates/rs/details/router.html | 1 + src/main/resources/web/templates/rs/search/do.html | 1 + 12 files changed, 35 insertions(+), 6 deletions(-) diff --git a/src/main/resources/web/js/rs/collections/aggregates.js b/src/main/resources/web/js/rs/collections/aggregates.js index dbf65ea..75c5e22 100644 --- a/src/main/resources/web/js/rs/collections/aggregates.js +++ b/src/main/resources/web/js/rs/collections/aggregates.js @@ -20,6 +20,8 @@ define([ this.fresh_until = response.fresh_until; this.valid_after = response.valid_after; var aggregates = {}; + var onionooVersion = response.version; + var buildRevision = response.build_revision; var relaysPublished = response.relays_published; var bridgesPublished = response.bridges_published; options.error = function(options) { @@ -122,7 +124,8 @@ define([ aggregatesArr.push(aggregate); }); collection[options.add ? 'add' : 'reset'](aggregatesArr, options); - success(err, relaysPublished, bridgesPublished); + success(err, onionooVersion, buildRevision, relaysPublished, + bridgesPublished); }).fail(function(jqXHR, textStatus, errorThrown) { if(jqXHR.statusText == "error") { error(2); diff --git a/src/main/resources/web/js/rs/collections/results.js b/src/main/resources/web/js/rs/collections/results.js index 73ef819..0110b7b 100644 --- a/src/main/resources/web/js/rs/collections/results.js +++ b/src/main/resources/web/js/rs/collections/results.js @@ -19,6 +19,8 @@ define([ this.fresh_until = response.fresh_until; this.valid_after = response.valid_after; var relays = []; + var onionooVersion = response.version; + var buildRevision = response.build_revision; var relaysPublished = response.relays_published; var bridgesPublished = response.bridges_published; options.error = function(options) { @@ -64,7 +66,8 @@ define([ $('.progress-bar').html("Rendering results..."); setTimeout(function() { collection[options.add ? 'add' : 'reset'](relays, options); - success(err, relaysPublished, bridgesPublished); + success(err, onionooVersion, buildRevision, + relaysPublished, bridgesPublished); }, 500); } else { $('.progress-bar').width((lookedUpRelays / relays.length * 100) + "%"); diff --git a/src/main/resources/web/js/rs/models/relay.js b/src/main/resources/web/js/rs/models/relay.js index a3b0d42..cdbe3a3 100644 --- a/src/main/resources/web/js/rs/models/relay.js +++ b/src/main/resources/web/js/rs/models/relay.js @@ -273,6 +273,8 @@ define([ relay.is_bridge = true; } if (relay) { + relay.onionooVersion = data.version; + relay.buildRevision = data.build_revision; relay.bridgesPublished = data.bridges_published; relay.relaysPublished = data.relays_published; model.processRelay(options, model, relay); diff --git a/src/main/resources/web/js/rs/router.js b/src/main/resources/web/js/rs/router.js index 1f65c9f..c67c4a1 100644 --- a/src/main/resources/web/js/rs/router.js +++ b/src/main/resources/web/js/rs/router.js @@ -96,8 +96,11 @@ define([ query = ""; } aggregateSearchView.collection.lookup({ - success: function(err, relaysPublished, bridgesPublished){ + success: function(err, onionooVersion, buildRevision, relaysPublished, + bridgesPublished){ aggregateSearchView.error = err; + aggregateSearchView.onionooVersion = onionooVersion; + aggregateSearchView.buildRevision = buildRevision; aggregateSearchView.relaysPublished = relaysPublished; aggregateSearchView.bridgesPublished = bridgesPublished; aggregateSearchView.render(query); @@ -137,8 +140,11 @@ define([ query = ""; } aggregateMapView.collection.lookup({ - success: function(err, relaysPublished, bridgesPublished){ + success: function(err, onionooVersion, buildRevision, relaysPublished, + bridgesPublished){ aggregateMapView.error = err; + aggregateMapView.onionooVersion = onionooVersion; + aggregateMapView.buildRevision = buildRevision; aggregateMapView.relaysPublished = relaysPublished; aggregateMapView.bridgesPublished = bridgesPublished; aggregateMapView.render(query); @@ -176,7 +182,8 @@ define([ doSearchView.collection.url = doSearchView.collection.baseurl + this.hashFingerprint(query); doSearchView.collection.lookup({ - success: function(err, relaysPublished, bridgesPublished){ + success: function(err, onionooVersion, buildRevision, + relaysPublished, bridgesPublished){ doSearchView.relays = doSearchView.collection.models; // Redirect to the details page when there is exactly one // search result. @@ -185,7 +192,9 @@ define([ doSearchView.relays[0].fingerprint); return; } - doSearchView.error = err; + doSearchView.error = err; + doSearchView.onionooVersion = onionooVersion; + doSearchView.buildRevision = buildRevision; doSearchView.relaysPublished = relaysPublished; doSearchView.bridgesPublished = bridgesPublished; doSearchView.render(query); diff --git a/src/main/resources/web/js/rs/views/aggregate/map.js b/src/main/resources/web/js/rs/views/aggregate/map.js index fe69bc0..a573e32 100644 --- a/src/main/resources/web/js/rs/views/aggregate/map.js +++ b/src/main/resources/web/js/rs/views/aggregate/map.js @@ -246,6 +246,8 @@ define([ aggregates: this.collection.models, countries: CountryCodes, error: this.error, + onionooVersion: this.onionooVersion, + buildRevision: this.buildRevision, relaysPublished: this.relaysPublished, bridgesPublished: this.bridgesPublished})); diff --git a/src/main/resources/web/js/rs/views/aggregate/search.js b/src/main/resources/web/js/rs/views/aggregate/search.js index 3c656fc..a259332 100644 --- a/src/main/resources/web/js/rs/views/aggregate/search.js +++ b/src/main/resources/web/js/rs/views/aggregate/search.js @@ -25,6 +25,8 @@ define([ aType: this.collection.aType, countries: CountryCodes, error: this.error, + onionooVersion: this.onionooVersion, + buildRevision: this.buildRevision, relaysPublished: this.relaysPublished, bridgesPublished: this.bridgesPublished})); diff --git a/src/main/resources/web/js/rs/views/search/do.js b/src/main/resources/web/js/rs/views/search/do.js index 53c80e6..16ee7a9 100644 --- a/src/main/resources/web/js/rs/views/search/do.js +++ b/src/main/resources/web/js/rs/views/search/do.js @@ -24,6 +24,8 @@ define([ relays: relays, countries: CountryCodes, error: this.error, + onionooVersion: this.onionooVersion, + buildRevision: this.buildRevision, relaysPublished: this.relaysPublished, bridgesPublished: this.bridgesPublished})); diff --git a/src/main/resources/web/templates/rs/aggregate/map.html b/src/main/resources/web/templates/rs/aggregate/map.html index 87ce0dc..b30e800 100644 --- a/src/main/resources/web/templates/rs/aggregate/map.html +++ b/src/main/resources/web/templates/rs/aggregate/map.html @@ -64,4 +64,6 @@ <a class="btn btn-secondary" id="permalink">Permalink</a> </div> </div> +<p>Information for relays was published: <%= relaysPublished %> UTC.<p> +<p><small>Onionoo version: <%= onionooVersion %>/<%= buildRevision %></small></p> <% } %> diff --git a/src/main/resources/web/templates/rs/aggregate/search.html b/src/main/resources/web/templates/rs/aggregate/search.html index 6679846..151afb6 100644 --- a/src/main/resources/web/templates/rs/aggregate/search.html +++ b/src/main/resources/web/templates/rs/aggregate/search.html @@ -146,5 +146,6 @@ probabilities of using relays in a particular country or AS as a guard, middle or exit relay. The results are restricted to only relays that were running at the last time the relays data was updated and do not include bridge data.</p> <p>Information for relays was published: <%= relaysPublished %> UTC.<p> +<p><small>Onionoo version: <%= onionooVersion %>/<%= buildRevision %></small></p> <% } %> </div> diff --git a/src/main/resources/web/templates/rs/details/bridge.html b/src/main/resources/web/templates/rs/details/bridge.html index 05f4d7b..69a1cd1 100644 --- a/src/main/resources/web/templates/rs/details/bridge.html +++ b/src/main/resources/web/templates/rs/details/bridge.html @@ -201,4 +201,5 @@ </div> </div> <p>Information for bridges was published: <%= relay.get('bridgesPublished') %> UTC.</p> +<p><small>Onionoo version: <%= relay.get('onionooVersion') %>/<%= relay.get('buildRevision') %></small></p> diff --git a/src/main/resources/web/templates/rs/details/router.html b/src/main/resources/web/templates/rs/details/router.html index 4ee4c88..db5ec03 100644 --- a/src/main/resources/web/templates/rs/details/router.html +++ b/src/main/resources/web/templates/rs/details/router.html @@ -305,3 +305,4 @@ should update it as soon as possible.</div> </div> <p>Information for relays was published: <%= relay.get('relaysPublished') %> UTC.</p> +<p><small>Onionoo version: <%= relay.get('onionooVersion') %>/<%= relay.get('buildRevision') %></small></p> diff --git a/src/main/resources/web/templates/rs/search/do.html b/src/main/resources/web/templates/rs/search/do.html index 4f48a50..40745a6 100644 --- a/src/main/resources/web/templates/rs/search/do.html +++ b/src/main/resources/web/templates/rs/search/do.html @@ -144,6 +144,7 @@ bridge relays do not advertise family members in their descriptors and so there is no value shown for bridge relays.</p> <p>Information for relays was published: <%= relaysPublished %> UTC.<br> Information for bridges was published: <%= bridgesPublished %> UTC.</p> +<p><small>Onionoo version: <%= onionooVersion %>/<%= buildRevision %></small></p> <% } %> </div>

1 0