<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Architecture on Skycoin Blog</title><link>https://blog.theskywirenetwork.net/tags/architecture/</link><description>Recent content in Architecture on Skycoin Blog</description><generator>Hugo</generator><language>en</language><lastBuildDate>Thu, 04 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://blog.theskywirenetwork.net/tags/architecture/index.xml" rel="self" type="application/rss+xml"/><item><title>Standalone Skywire Utilities: Encrypted Connection, Public Key as Identity</title><link>https://blog.theskywirenetwork.net/posts/standalone-skywire-pk-identity/</link><pubDate>Thu, 04 Jun 2026 00:00:00 +0000</pubDate><guid>https://blog.theskywirenetwork.net/posts/standalone-skywire-pk-identity/</guid><description>&lt;p&gt;If you keep peeling back skywire&amp;rsquo;s layers — the apps, the multihop router, the multiplexed routes, the discovery services, the dmsg relay — what&amp;rsquo;s at the very bottom? One primitive: an &lt;strong&gt;encrypted connection between two public keys&lt;/strong&gt;. The noise handshake provides the encryption; the public key &lt;em&gt;is&lt;/em&gt; the identity. No certificate authority vouches for it, no DNS name resolves to it, no allocated IP belongs to it. A party is its key, and two keys that have completed a handshake have, by construction, authenticated each other and established a confidential channel.&lt;/p&gt;</description></item><item><title>Skywire as a Personal Fleet Substrate</title><link>https://blog.theskywirenetwork.net/posts/skywire-as-a-personal-fleet-substrate/</link><pubDate>Mon, 25 May 2026 00:00:00 +0000</pubDate><guid>https://blog.theskywirenetwork.net/posts/skywire-as-a-personal-fleet-substrate/</guid><description>&lt;p&gt;Most discussions of &amp;ldquo;distributed computing&amp;rdquo; reach for the wrong frame when applied to a personal cluster of machines you actually own. The Hadoop / Spark / Kubernetes mental model assumes you have a homogeneous workload that splits cleanly across many workers — the &amp;ldquo;parallelize one big problem&amp;rdquo; shape. For most personal infrastructure, that workload doesn&amp;rsquo;t exist. You don&amp;rsquo;t have terabytes of data to process or a million-row sweep to chew through every night.&lt;/p&gt;</description></item></channel></rss>