<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Juncheng Yang</title>
    <link>https://blog.jasony.me/</link>
    <atom:link href="https://blog.jasony.me/feed.xml" rel="self" type="application/rss+xml" />
    <description>Notes on storage and machine-learning systems by Juncheng Yang — caching, systems research, and advising.</description>
    <language>en</language>
    <item>
      <title>Your LRU cache is working too hard</title>
      <link>https://blog.jasony.me/posts/2026-07-23-lazy-promotion-cache-eviction.html</link>
      <guid>https://blog.jasony.me/posts/2026-07-23-lazy-promotion-cache-eviction.html</guid>
      <pubDate>Thu, 23 Jul 2026 12:00:00 GMT</pubDate>
      <category>Research</category>
      <description>Across 6,357 production traces, most LRU promotions did little useful work. Delaying the decision made caches faster without making them less effective.</description>
    </item>
    <item>
      <title>Authorship inflation at OSDI: counting papers two ways</title>
      <link>https://blog.jasony.me/posts/2026-07-10-osdi-authorship-inflation.html</link>
      <guid>https://blog.jasony.me/posts/2026-07-10-osdi-authorship-inflation.html</guid>
      <pubDate>Fri, 10 Jul 2026 12:00:00 GMT</pubDate>
      <category>Academia</category>
      <description>OSDI author lists have more than doubled since 1994. Whether the field's most prolific authors are actually publishing more depends on whether a paper counts as one, or as one divided by its author count.</description>
    </item>
    <item>
      <title>Make model precision elastic</title>
      <link>https://blog.jasony.me/posts/2026-05-24-make-model-precision-elastic.html</link>
      <guid>https://blog.jasony.me/posts/2026-05-24-make-model-precision-elastic.html</guid>
      <pubDate>Sun, 24 May 2026 12:00:00 GMT</pubDate>
      <category>Systems</category>
      <description>MorphServe temporarily converts selected model layers into KV-cache headroom when traffic spikes, then restores full precision when the pressure passes.</description>
    </item>
    <item>
      <title>Why FIFO is (almost) all you need for cache eviction</title>
      <link>https://blog.jasony.me/posts/2026-05-20-fifo-is-all-you-need.html</link>
      <guid>https://blog.jasony.me/posts/2026-05-20-fifo-is-all-you-need.html</guid>
      <pubDate>Wed, 20 May 2026 12:00:00 GMT</pubDate>
      <category>Research</category>
      <description>A simple first-in-first-out queue, given one quick second chance, beats far more complex eviction policies on real workloads. Here is the intuition behind S3-FIFO and SIEVE.</description>
    </item>
    <item>
      <title>Measurement-first systems research</title>
      <link>https://blog.jasony.me/posts/2026-04-08-measurement-first-systems-research.html</link>
      <guid>https://blog.jasony.me/posts/2026-04-08-measurement-first-systems-research.html</guid>
      <pubDate>Wed, 08 Apr 2026 12:00:00 GMT</pubDate>
      <category>Research</category>
      <description>The most useful systems I have built started not with a design, but with a long, uncomfortable look at how the existing system actually behaved in production.</description>
    </item>
    <item>
      <title>Advice for prospective systems PhD students</title>
      <link>https://blog.jasony.me/posts/2026-03-02-advice-for-prospective-students.html</link>
      <guid>https://blog.jasony.me/posts/2026-03-02-advice-for-prospective-students.html</guid>
      <pubDate>Mon, 02 Mar 2026 12:00:00 GMT</pubDate>
      <category>Academia</category>
      <description>I am recruiting students who want to build real systems for storage, databases, and machine learning. Here is how I think about the work and what I look for.</description>
    </item>
  </channel>
</rss>
