<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Starrocks on Kevin wang's blog</title><link>https://kevinwang930.github.io/categories/starrocks/</link><description>Recent content in Starrocks on Kevin wang's blog</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Sat, 22 Aug 2026 15:00:00 +0200</lastBuildDate><atom:link href="https://kevinwang930.github.io/categories/starrocks/index.xml" rel="self" type="application/rss+xml"/><item><title>StarRocks: FE query planning and the MPP path</title><link>https://kevinwang930.github.io/post/data/db/starrocks/architecture/</link><pubDate>Sat, 22 Aug 2026 15:00:00 +0200</pubDate><guid>https://kevinwang930.github.io/post/data/db/starrocks/architecture/</guid><description>&lt;p&gt;&lt;strong&gt;StarRocks&lt;/strong&gt; is a MySQL-compatible MPP OLAP engine. A cluster is &lt;strong&gt;Frontend (FE)&lt;/strong&gt; nodes for metadata and query coordination plus &lt;strong&gt;Backend (BE)&lt;/strong&gt; or &lt;strong&gt;Compute Node (CN)&lt;/strong&gt; workers for execution. This post follows &lt;strong&gt;FE query planning&lt;/strong&gt;—logical plan, Cascades optimize, fragmentize—and the &lt;strong&gt;MPP&lt;/strong&gt; path through schedule, deploy, and result pull. Worker process internals are in &lt;a href="../backend/"&gt;Backend and Compute Node&lt;/a&gt;.&lt;/p&gt;</description></item></channel></rss>