When Your Backup Plan Has Its Own Failure Points, You Do Not Have a Backup Plan

Network redundancy has become a standard line item for IT teams. Most businesses today have some form of failover connectivity. What far fewer have is a network designed to keep running when the firewall fails, when a switch dies, when a power surge takes out hardware mid-shift. Internet redundancy and network redundancy are not the same thing, and the gap between them is where most outages actually happen.

In this episode of Go Beyond the Connection, Hamed Mazrouei, Founder and CEO of Milagro Corporation, shares how he built the HAFT platform, a Highly Available, Fault Tolerant network architecture, and what it takes to design a system that covers every failure point from the ISP down to the access point. His approach is not theoretical. It has been stress-tested across healthcare offices, restaurant chains, legal firms, and multi-tenant office buildings, with real hardware pulled mid-operation to prove it works.

Designing for Every Layer, Not Just the Edge

Most redundancy deployments start and stop at the internet connection. Mazrouei’s team started there too, offering failover connectivity to their managed service clients beginning in 2019. But one large customer changed the design brief entirely. They wanted protection not just from ISP failures, but from hardware failures inside the building.

That conversation prompted Milagro to rethink the full architecture. What if the firewall failed? What if the switch failed? What if an access point went down? Each of those questions led to a design decision: double the critical equipment, put every device in an active-active configuration, and build redundancy from the edge of the network to the last hop inside the building.

Mazrouei is direct about why active-active matters more than active-passive. If a device is deployed and paid for, it should be working, not sitting dormant as a backup that may never get tested until something goes wrong. That philosophy drove the HAFT platform’s core architecture and later became the basis for scaling it down to SMB deployments at a fraction of the original cost.

Scaling Fault Tolerance Without Pricing Out the Customer

The original HAFT implementation was built for enterprise clients with the budget to match. The challenge was reproducing that level of resilience for restaurant operators who needed fault tolerance but could not spend several thousand dollars per site.

Mazrouei’s team made targeted changes to reduce complexity and cost. They shifted from two broadband connections with LTE backup to a primary and failover setup. They also moved from dual active firewalls to an active and passive pair at the edge, while keeping internal systems fully redundant.

The result cut monthly costs from thousands to a few hundred dollars, making reliable, enterprise-level uptime realistic for restaurants, where even a short outage can disrupt orders, payments, and delivery platforms all at once.

What SD-WAN Actually Changed in the Field

To confirm the result was not coincidental, the team removed the device, replicated the congestion, and watched the problems return. They reinstalled it. The problems disappeared. The static IP challenge followed the same logic: by moving IP management from the ISP layer to the Bigleaf SD-WAN layer, Mazrouei’s team kept multi-tenant office clients online regardless of which ISP was active.

Episode Highlights

• Why internet redundancy and network redundancy solve different problems, and how treating them as the same can leave businesses exposed
• How the HAFT platform removes single points of failure from the ISP connection all the way to the access point
• The design decisions that made enterprise-level fault tolerance affordable for SMBs
• How Bigleaf’s dynamic QoS fixed a bandwidth spike issue that firewall-based QoS could not resolve
• A proven approach to maintaining static IP availability in multi-tenant buildings during ISP outages
• Why Mazrouei rates his partnership with Bigleaf a ten, and what that standard says about how Milagro chooses its technology partners

“At the heart of every business today is connectivity to the network, not just the internet anymore. Yes, internet redundancy, but also network redundancy. The moment you experience an outage, it is catastrophic for most businesses.”

— Hamed Mazrouei, Founder and CEO, Milagro Corporation

The Bigleaf Connection

Bigleaf’s SD-WAN platform is built on two capabilities that show up repeatedly in Mazrouei’s account of what made the difference: continuous circuit health monitoring and dynamic traffic prioritization. The system does not wait for a link to go down before responding. It detects degradation in real time and reroutes traffic automatically, without user intervention and without dropping active calls or sessions.

For MSPs like Milagro, that behavior matters at two levels. At the client level, it means the network adapts to conditions the client never sees. At the business level, it means the MSP can deliver a reliability promise without depending on manual monitoring or reactive support calls. Mazrouei describes Bigleaf not as a vendor, but as a partner, a distinction he treats as operational, not semantic. If a partner cannot support at the level the MSP delivers, the entire promise to the end client falls apart.

The plug-and-play deployment model reinforces that position. Milagro pre-configures devices before deployment, building on Bigleaf’s robust defaults, so that installation at the client site is a matter of plugging in, not spending hours in a CLI. That efficiency compounds across a managed services practice responsible for network reliability at scale.