Cheat Sheet for the hybrid infrastructure

🏗️ The Hybrid Infrastructure Stack 

1. Core Hardware (The "Garage" Servers)

• Primary Node: Dell Precision 7960 – High-end workstation with a 2200W power supply.

• Redundancy Node: Dell PowerEdge T640 – Dual Intel Xeon Platinum processors (48 cores) and 256GB RAM.

2. Virtualization & Management

• Proxmox: The open-source hypervisor used to manage Virtual Machines (VMs) and create an active-active cluster.

• Ubuntu & Windows VMs: Standardized templates used for rapid environment provisioning.

3. Data Engineering & Pipelines

• Apache Spark: Replaced Databricks for heavy data computation and processing real estate transaction data.

• Apache Airflow: Used for workflow orchestration (scheduling and monitoring data pipelines).

• DBeaver: The tool for SQL-based analysis and browsing Parquet files.

• Delta Lake / Parquet: The storage formats used for the data lake architecture.

4. Connectivity & Security

• Tailscale: A zero-config VPN that allowed the remote team to access garage-hosted VMs securely without public IPs or complex firewall rules.

5. Development AI (The "Architects")

• Claude Code & Gemini: Used as technical collaborators to design the network architecture, write Proxmox configurations, and migrate Spark code.

---

💡 Quick Comparison: Why the Switch?

Feature	Cloud (Azure/Databricks)	Hybrid (On-Prem)
Cost	Metered (Pay-per-run)	Fixed (Upfront hardware)
Innovation	Cautious (Fear of high bills)	Fearless (Unlimited iterations)
Access	Paid Public IPs	Tailscale (Secure/Private)
Performance	Scalable but expensive	High-density local compute

https://bezit.substack.com/p/the-16k-decision-that-saved-my-startup?r=7xveqk&utm_campaign=post&utm_medium=web&triedRedirect=true