Outages, Attacks, and a Need for Resilience

This quarter has been a season of major cloud disruptions. They're a sobering reminder of how much entire economies depend on online services. Your business, or one you rely on, could go down from a single misconfiguration or software failure. At Fastly, we believe that complex systems fail in complex ways and that nobody is immune.  

On the morning of November 18th, concurrent with a broadly advertised outage, several advanced DDoS attacks unsuccessfully targeted Fastly's platform. At the same time, several major customers shifted tens of petabytes of traffic over to Fastly due to another provider's outage. 

You can see the attack event and unrelated parallel rapid traffic swing over the span of approximately an hour:

Along with the concurrent and targeted attack on the Fastly platform

These events – the advanced attack mitigation and the high-scale traffic failover – happened simultaneously, with no end-user disruption. When running a modern, global platform, security events and customer performance fuse together into a single resilience mission. 

We run one of the largest production platforms on the planet, so we can’t allow for a separation of performance and security in our focus. What we’ve seen these past few months has made it clear that resilience can no longer be treated as a backup plan. It must be a design principle.

Multi-Vendor for Resiliency

In light of these outages, a multi-vendor strategy is no longer optional. 

Whether you are evaluating an ISP, telco, CableCo, cloud provider, or CDN, a multi-vendor architecture yields better resilience: less downtime, fewer outages, and global scalability. It helps customers enjoy the best possible user experience. Multi-vendor strategies provide added redundancy, security, and performance headroom, effectively diversifying your risk and strengthening resilience to complex, multi-factor incident scenarios. With the right infrastructure providers in place, an organization can strategically route around outages, bypass bottlenecks, and deliver applications quickly and with the highest fidelity, reducing the associated risks and costs of downtime while maintaining a strong security posture.

Choosing Platforms that Earn Your Traffic - Every Day 

With the right multi-vendor solution in place, decisions become rooted in evidence. The fastest, most consistent providers who are continuously improving the details of their systems win out over time. Many of our customers run multi-vendor architectures, and that’s an environment that we know really well and thrive in.  

That means every provider has to earn traffic based on how they perform, not just what they promise. That's why when another platform runs into trouble and traffic shifts suddenly, our systems take on that extra load without hesitation. Being ready for rainy-day moments is simply part of operating modern infrastructure at scale.

Platforms must be engineered with the expectation that failures will occur. What matters is how they respond.

In practice, a platform designed with this perspective should have these critical requirements in mind:

• Treat availability and integrity as first-order design constraints rather than optional layers

• Maintain predictable performance under sudden shifts in demand or unexpected surges

• Mitigate high-volume attack traffic without sacrificing availability for legitimate users

• Adapt to internet-scale variability without introducing unnecessary latency

• Sustain elevated load during failover events or dependency disruptions

These capabilities aren’t defined by an isolated incident; they emerge from architectural choices, operational discipline, and continued refinement over time.

If recent disruptions have you reassessing your infrastructure approach, Fastly can help guide resilient design best practices. Our position as a trusted player and adviser in the space can help you feel secure in your journey to optimizing for high availability and resilience. Let’s build something great together.

Interested in learning more about what makes Fastly resilient? Check out these additional resources:

• "Fastly’s Pillars of Resilience: Building a More Robust Internet" (Part 1)

• "Design for Chaos: Fastly’s Principles of Fault Isolation and Graceful Degradation" (Part 2)

• "Build for Scale: Fastly’s Principles of Distributed Decision Making and Self-healing Systems" (Part 3)

• "Resilience by Design: Lessons in Multi-Cloud Readiness"