diff --git a/.gitignore b/.gitignore
index 7cc94e0..ab952ef 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,7 +1,8 @@
-_site/
-.swp
-.sass-cache/
-.jekyll-metadata
 .bundle/
-vendor/
+.ipynb_checkpoints/
+.jekyll-metadata
+.sass-cache/
+.swp
 .vscode/
+_site/
+vendor/
diff --git a/_notebooks/2022-10-15-multicast/Multicast.ipynb b/_notebooks/2022-10-15-multicast/Multicast.ipynb
new file mode 100644
index 0000000..77496ef
--- /dev/null
+++ b/_notebooks/2022-10-15-multicast/Multicast.ipynb
@@ -0,0 +1,69 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "2bfdffd8",
+   "metadata": {},
+   "source": [
+    "# No More Magic: A Practical Introduction to Multicast\n",
+    "\n",
+    "Recent interaction - had a networking issue, likely required configuration tweaking, but:\n",
+    "\n",
+    "- Not a lot of people who knew how to diagnose\n",
+    "- Training people to diagnose is challenging\n",
+    "\n",
+    "I've said that [High Performance Systems aren't magical](/2019/07/high-performance-systems.html), but realization that for many topics, accessibility is problematic. If I _wanted_ to know more, I don't have any meaningful way to investigate beyond asking for someone's time. While manager's judgment call is fine (\"this isn't worth investigating further\"), I've had a lot of career \"upskilling\" learned from open source. Want to make sure that if other people are interested, there's a way to meaningfully engage with the topic."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d8cee425",
+   "metadata": {},
+   "source": [
+    "Add links for:\n",
+    "- Notebook source\n",
+    "- Script for running the notebook"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "2edfbd79-d246-4b9a-a4a3-8e383c6b22a1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#include <cstdlib>\n",
+    "#include <thread>\n",
+    "#include <vector>\n",
+    "\n",
+    "auto run_parallel(std::initializer_list<const char*> commands) {\n",
+    "    \n",
+    "}\n",
+    "\n",
+    "std::vector<std::thread> threads{};\n",
+    "\n",
+    "for (auto command : commands)\n",
+    "    threads.emplace_back([=]() { std::system(command); });\n",
+    "\n",
+    "for (auto& thread : threads)\n",
+    "    thread.join();"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "C++17",
+   "language": "C++17",
+   "name": "xcpp17"
+  },
+  "language_info": {
+   "codemirror_mode": "text/x-c++src",
+   "file_extension": ".cpp",
+   "mimetype": "text/x-c++src",
+   "name": "c++",
+   "version": "17"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/_notebooks/2022-10-15-multicast/run.sh b/_notebooks/2022-10-15-multicast/run.sh
new file mode 100755
index 0000000..bff1c80
--- /dev/null
+++ b/_notebooks/2022-10-15-multicast/run.sh
@@ -0,0 +1,24 @@
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+mkdir -p "$SCRIPT_DIR"/notebook
+
+# Package pinning because of https://github.com/jupyter-xeus/xeus-cling/issues/415
+read -r -d '' RUN <<EOF
+    apt update \
+ && apt install -y \
+        iperf3 \
+        sudo \
+        strace \
+ && chown \$MAMBA_USER:\$MAMBA_USER /notebook \
+ && sudo -u \$MAMBA_USER micromamba -y -n base install -c conda-forge \
+        jupyterlab \
+        xeus-cling==0.13.0 \
+        libstdcxx-devel_linux-64==9.4.0 \
+        libgcc-devel_linux-64==9.4.0 \
+ && sudo -u \$MAMBA_USER micromamba run jupyter-lab /notebook
+EOF
+
+# --net=host: Docker networking doesn't do multicast, use host networking so the
+#   local router handles it instead.
+# --user=root: Install iperf3/strace
+podman run --net=host --user=root --rm --volume "$SCRIPT_DIR"/notebook:/notebook:Z -it mambaorg/micromamba bash -c "$RUN"
+