The latest articles on DEV Community by Sebastian Mejia (@sdmejia). https://dev.to/sdmejia https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3278279%2Fd98ed14e-d150-41e0-a0bc-038d1e19813d.png https://dev.to/sdmejia en Sebastian Mejia Fri, 20 Jun 2025 03:25:12 +0000 https://dev.to/sdmejia/highly-available-office-network-design-5ci1 https://dev.to/sdmejia/highly-available-office-network-design-5ci1 <p>In this post, I’ll walk through a high availability (HA) office network architecture I’ve implemented across several branch environments. The goal was to create a resilient, fault-tolerant design that ensures connectivity, security, and uptime — even during ISP outages, hardware failures, or power events.</p> <p>The design uses FortiGate firewalls in HA mode, Meraki switches, SD-WAN with SLA rules, and redundant power and network paths.</p> <h2> 🧱 Network Design Overview </h2> <p>Here’s the logical architecture:</p> <ul> <li> <strong>Dual ISPs</strong>: Two independent internet circuits for redundancy.</li> <li> <strong>WAN Layer</strong>: Two WAN switches (one per ISP).</li> <li> <strong>Firewalls</strong>: Two FortiGate appliances in HA (active/passive) mode.</li> <li> <strong>LAN Layer</strong>: Two internal Meraki switches — each connected to a different FortiGate unit.</li> <li> <strong>Switch Interconnect</strong>: Internal switches connected via a port-channel (LAG) for redundancy.</li> <li> <strong>UPS Redundancy</strong>: Power is fed through separate UPS circuits (A/B) to isolate failures.</li> </ul> <h2> 🖥️ Diagram </h2> <p><a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fp626mcaz5riso7kj1mbi.jpg" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fp626mcaz5riso7kj1mbi.jpg" alt="Image description" width="558" height="641"></a></p> <h2> 🔧 Key Design Highlights </h2> <h3> 🔁 FortiGate HA Configuration </h3> <ul> <li>Active/passive mode with heartbeat and session sync</li> <li>Dual WAN interfaces on each firewall, connecting to both WAN switches</li> <li>SD-WAN configured with SLA tracking for packet loss, jitter, and latency</li> </ul> <h3> 🌐 SD-WAN Failover Logic </h3> <ul> <li>If ISP 1 experiences packet loss or latency beyond SLA thresholds, traffic auto-fails over to ISP 2</li> <li>No user interaction required; seamless failover from the user perspective</li> </ul> <h3> 🔌 Power Resilience </h3> <ul> <li>UPS Circuit A powers one side (WAN switch + FortiGate + Internal switch)</li> <li>UPS Circuit B powers the other side</li> <li>A failure on one power path doesn’t take down the core network infrastructure</li> </ul> <h3> 🧩 Switch Redundancy </h3> <ul> <li>Internal Meraki switches are interconnected using a LAG (Port-Channel) to provide redundancy and maintain internal traffic paths</li> <li>Each Meraki switch also connects to the FortiGate cluster via a dedicated LAG, ensuring link-level redundancy between LAN and firewall layers</li> </ul> <h2> 🔄 Real-World Results </h2> <ul> <li>Survived ISP outages with zero downtime thanks to SD-WAN</li> <li>No business disruption during firewall firmware upgrades (failover was seamless)</li> <li>Local power failures affected one side only — only endpoints connected to the affected internal switch lost connectivity, while core infrastructure remained operational</li> <li>Easy to replicate at other offices due to modular and well-documented design</li> </ul> <h2> 🧠 Lessons Learned </h2> <ul> <li>SD-WAN SLA tuning is critical — overly sensitive thresholds can cause flapping</li> <li>Always validate HA failover under load — test VPNs, NAT, and logging</li> <li>Label power and data paths clearly — it saves time in emergencies</li> <li>Documentation is as important as design — it enabled repeatable deployments</li> </ul> <h2> 📌 Final Thoughts </h2> <p>This design has proven highly effective across multiple office deployments. It offers true high availability across networking, power, and internet access layers — while staying relatively simple to manage.</p> <p>If you’re designing or upgrading a branch office or small datacenter, consider using this architecture as a blueprint. Feel free to adapt it to your own stack — whether you're using FortiGate, Palo Alto, Meraki, or other gear.</p> networking highavailability networkdesign sdwan