<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Workloads on Harvard Systems Group</title><link>https://systems.seas.harvard.edu/tags/workloads/</link><description>Recent content in Workloads on Harvard Systems Group</description><generator>Hugo</generator><language>en</language><copyright>Copyright © 2026 The President and Fellows of Harvard College</copyright><lastBuildDate>Mon, 12 Oct 2026 00:00:00 -0500</lastBuildDate><atom:link href="https://systems.seas.harvard.edu/tags/workloads/index.xml" rel="self" type="application/rss+xml"/><item><title>A Year in LLM Serving: Workloads, Caching, and Routing</title><link>https://systems.seas.harvard.edu/blog/a-year-in-llm-serving/</link><pubDate>Mon, 12 Oct 2026 00:00:00 -0500</pubDate><guid>https://systems.seas.harvard.edu/blog/a-year-in-llm-serving/</guid><description>An analysis of 6.12 billion production LLM requests, examining workload evolution, traffic patterns, prefix reuse, cache eviction, and request routing.</description></item></channel></rss>