<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Infrastructure on Rik Kisnah - Blog</title><link>https://www.rik-kisnah.ai/tags/infrastructure/</link><description>Recent content in Infrastructure on Rik Kisnah - Blog</description><generator>Hugo</generator><language>en</language><lastBuildDate>Fri, 06 Mar 2026 21:33:29 -0800</lastBuildDate><atom:link href="https://www.rik-kisnah.ai/tags/infrastructure/feed.xml" rel="self" type="application/rss+xml"/><item><title>The AI Pyramid: Five Layers Between Hardware and AGI</title><link>https://www.rik-kisnah.ai/posts/ai-pyramid-five-layers-between-hardware-and-agi/</link><pubDate>Fri, 06 Mar 2026 21:33:29 -0800</pubDate><guid>https://www.rik-kisnah.ai/posts/ai-pyramid-five-layers-between-hardware-and-agi/</guid><description>Disclaimer: The views expressed in this article are my own and do not represent those of my employer. This content was written with the assistance of AI language models but curated and controlled by me.
Most conversations about AI focus on the model. The chatbot, the assistant, the thing that answers your question in seconds. But the model is one layer in a larger stack. Understanding the full stack changes how you think about AI strategy, investment, and what comes next.</description></item><item><title>The Complete NCCL Reference Guide: Commands, Errors, and Troubleshooting for OCI GPU Infrastructure</title><link>https://www.rik-kisnah.ai/posts/nccl-complete-reference-guide/</link><pubDate>Sat, 15 Nov 2025 00:00:00 -0700</pubDate><guid>https://www.rik-kisnah.ai/posts/nccl-complete-reference-guide/</guid><description>Disclaimer: This article reflects my personal research and analysis based on publicly available information and is not representative of my employer&amp;rsquo;s official position.
Executive Summary NCCL (NVIDIA Collective Communication Library) is the cornerstone of distributed GPU computing, enabling efficient communication between GPUs in multi-node clusters. This comprehensive guide provides every NCCL command, parameter, error message, and troubleshooting technique you need for successful deployment on Oracle Cloud Infrastructure (OCI).
Table of Contents Why NCCL Exists Understanding Collective Communications NCCL Fundamentals Complete NCCL Commands Reference All NCCL Environment Variables NCCL Error Messages and Solutions OCI GPU-Specific Configurations Advanced Troubleshooting Scenarios Performance Tuning Reference Quick Reference Tables Why NCCL Exists The Distributed Training Challenge Modern AI models have grown exponentially in size and complexity.</description></item><item><title>From First Principles to Zettascale: How OCI's GPU/RDMA Architecture Redefines AI Infrastructure</title><link>https://www.rik-kisnah.ai/posts/summary-gpu-oci-first-principles-blog/</link><pubDate>Sun, 26 Oct 2025 22:43:15 -0400</pubDate><guid>https://www.rik-kisnah.ai/posts/summary-gpu-oci-first-principles-blog/</guid><description>Disclaimer: This article reflects my personal research and analysis based on publicly available information and is not representative of my employer&amp;rsquo;s official position.
In the rapidly evolving landscape of AI infrastructure, one company has quietly revolutionized how we think about GPU computing at scale. Through a series of &amp;ldquo;First Principles&amp;rdquo; engineering blogs and groundbreaking deployments, Oracle Cloud Infrastructure (OCI) has demonstrated that starting from fundamental physics and systems design—rather than following industry conventions—can yield extraordinary results.</description></item><item><title>Oracle AI World 2025: Enterprise AI from Promise to Production</title><link>https://www.rik-kisnah.ai/posts/oracle-ai-world-2025/</link><pubDate>Sat, 18 Oct 2025 09:00:00 -0700</pubDate><guid>https://www.rik-kisnah.ai/posts/oracle-ai-world-2025/</guid><description>The AI Inflection Point At Oracle AI World 2025, a clear message emerged: AI isn&amp;rsquo;t just another technology trend—it&amp;rsquo;s the bridge between decades of enterprise data and the future of business innovation.1
Unlocking Enterprise Data&amp;rsquo;s True Potential Oracle founder Larry Ellison framed the opportunity succinctly: &amp;ldquo;AI offers the highest-value technology we have ever seen so far.&amp;rdquo;2 The critical insight? Most AI models train on internet data, missing the real competitive advantage: enterprise databases containing mission-critical business information.</description></item><item><title>Three Weeks in Batam: Bringing NVIDIA GB200 to Life on the Data Plane</title><link>https://www.rik-kisnah.ai/posts/gb200-batam-data-plane-rollout/</link><pubDate>Sat, 15 Mar 2025 10:00:00 -0700</pubDate><guid>https://www.rik-kisnah.ai/posts/gb200-batam-data-plane-rollout/</guid><description>Three weeks in Batam, Indonesia in March 2025. Not a vacation - something far more meaningful. I was there to help bring the NVIDIA GB200 data plane to life, working alongside some of the brightest minds at OCI and NVIDIA. This was the moment where cutting-edge technology meets real-world infrastructure, and I got to be part of making it happen.
Related readings:
Behind the Scenes: NVIDIA GB200 NVL72 on OCI APIs - Technical deep dive on the NVIDIA GB200 and OCI integration Supercluster: NVIDIA Blackwell Dedicated Alloy - OCI&amp;rsquo;s Supercluster offering with Blackwell GPUs Nvidia GB200 NVL72 Now Available via Oracle Cloud - Data Center Dynamics coverage of the launch The Mission: Time to Market When I arrived in Batam, the pressure was real.</description></item><item><title>Wi-Fi Experiments: From Lab to Campus Networks</title><link>https://www.rik-kisnah.ai/posts/wifi-experiments-journey-1999-2000/</link><pubDate>Thu, 25 May 2000 00:00:00 +0000</pubDate><guid>https://www.rik-kisnah.ai/posts/wifi-experiments-journey-1999-2000/</guid><description>Wireless Networking (1999) IEEE 802.11 shipped in 1997. By 1999, Singapore was ahead of the curve—early 802.11b trials at universities. At NTU, we had access to experimental wireless infrastructure. The dream: no wires. Just laptops talking over the air.
Reality: signal was weak, interference constant, range 20 meters on a good day. Tropical humidity killed performance. Water vapor absorbs radio waves. Our tests worked in the lab, then failed in the hallway.</description></item></channel></rss>