<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Storage on *scratch*</title>
    <link>https://www.scrivano.org/tags/storage/</link>
    <description>Recent content in Storage on *scratch*</description>
    <generator>Hugo</generator>
    <language>en</language>
    <lastBuildDate>Sat, 06 Jun 2026 10:03:54 +0000</lastBuildDate>
    <atom:link href="https://www.scrivano.org/tags/storage/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Composefs - a file system for container images</title>
      <link>https://www.scrivano.org/posts/2021-10-26-compose-fs/</link>
      <pubDate>Tue, 26 Oct 2021 16:02:56 +0200</pubDate>
      <guid>https://www.scrivano.org/posts/2021-10-26-compose-fs/</guid>
      <description>&lt;p&gt;For the last couple of weeks, I&amp;rsquo;ve been playing on a PoC implementation of a file system for the Linux kernel. The goal is to address a fundamental limitation in how container images are stored: the existing overlay model deduplicates at the layer level, but once you want per-file deduplication — so that identical files across different images share a single copy on disk and in the page cache — the current architecture gets in the way and requires awkward workarounds involving hard links or filesystem-specific reflinks.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
