<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Deep in the clouds, I know not where.</title><link>https://lihuiba.github.io/en/</link><description>Recent content on Deep in the clouds, I know not where.</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Mon, 24 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://lihuiba.github.io/en/index.xml" rel="self" type="application/rss+xml"/><item><title>The Microkernel VMM, Done Right</title><link>https://lihuiba.github.io/en/p/microkernel-vmm-done-right/</link><pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate><guid>https://lihuiba.github.io/en/p/microkernel-vmm-done-right/</guid><description>In 2005 the Xen team asked whether VMMs are microkernels done right; over twenty years isolation went to microVMs, the dataplane to silicon, discipline to DPUs, and only 'the service stack restarts painlessly' went unclaimed — this article traces that history and its four obstacles, then offers a pure-software assembly: a thin core defined as firmware, stateless service VMs, a dataplane delegated to silicon, holding with VM boundaries and the IOMMU the boundary the DPU holds with PCIe.</description></item><item><title>Time to End the DPU Myth</title><link>https://lihuiba.github.io/en/p/end-of-dpu-myth/</link><pubDate>Thu, 20 Aug 2026 00:00:00 +0000</pubDate><guid>https://lihuiba.github.io/en/p/end-of-dpu-myth/</guid><description>Every large-scale DPU success comes from a hyperscaler's vertically integrated, in-house program, while the merchant DPU market is a complete wreck. An attribution audit of its seven value propositions: five do not survive scrutiny, one is real but software-replaceable, and one is real but worth only half — half a claim left standing out of seven. Thin value wrapped in a thick narrative is the standard recipe for a myth.</description></item><item><title>The Myth of Fixed-Output Compression</title><link>https://lihuiba.github.io/en/p/against-fixed-output/</link><pubDate>Mon, 17 Aug 2026 00:00:00 +0000</pubDate><guid>https://lihuiba.github.io/en/p/against-fixed-output/</guid><description>Fixed-output compression — variable-length input stuffed into fixed-size output blocks — was credited with eliminating read amplification, aligning I/O, improving compression ratio, and saving memory. Term-by-term attribution finds otherwise: these benefits belong to granularity choices, to the index and the decompressor, hold only inside a restricted framing, or cannot be found in measurement at all — while the costs are format-level and permanent. The conclusion is polite but total: as a format design decision, fixed-output is not worth adopting.</description></item><item><title>Overlaybd — The Ultimate Image Format for Every Workload</title><link>https://lihuiba.github.io/en/p/overlaybd-the-ultimate-image-format/</link><pubDate>Mon, 10 Aug 2026 00:00:00 +0000</pubDate><guid>https://lihuiba.github.io/en/p/overlaybd-the-ultimate-image-format/</guid><description>OCI tar layers for containers, qcow2 for VMs — both carry structural costs. Overlaybd is a layered, lazily-loaded, seekably-compressed block-device image format for every workload: sub-second cold starts, proven in production across many companies.</description></item><item><title>Hello World</title><link>https://lihuiba.github.io/en/p/hello-world/</link><pubDate>Sun, 09 Aug 2026 00:00:00 +0000</pubDate><guid>https://lihuiba.github.io/en/p/hello-world/</guid><description>The first post of this blog</description></item><item><title>Stackful Coroutine Made Fast</title><link>https://lihuiba.github.io/en/p/stackful-coroutine-made-fast/</link><pubDate>Mon, 14 Oct 2024 00:00:00 +0000</pubDate><guid>https://lihuiba.github.io/en/p/stackful-coroutine-made-fast/</guid><description>Stackful coroutine is not intrinsically slow — current implementations are. With context-aware context switching (CACS), saving only the minimal registers per caller/callee context, stackful outperforms stackless in most cases and ties otherwise. Implemented in Photon.</description></item><item><title>Setting Up IPMI on a Shared LAN Port</title><link>https://lihuiba.github.io/en/p/ipmi-shared-lan-setup/</link><pubDate>Wed, 26 Mar 2014 22:51:49 +0800</pubDate><guid>https://lihuiba.github.io/en/p/ipmi-shared-lan-setup/</guid><description>Multiplexing the eth0 port with VLANs to run IPMI over the shared network interface</description></item></channel></rss>