Local hosting, hybrid deployments and multi-cloud design: How to stay online when Amazon breaks

A major Amazon Web Services (AWS) outage on Monday (yesterday), disrupted some of the world’s most widely used websites and apps – including Slack, Snapchat, Fortnite, Alexa, Coinbase, and multiple banking platforms – exposing once again the fragility of digital dependence on a handful of global cloud providers.

According to AWS, the disruption was caused by “increased error rates and latencies” in its US-East-1 region, one of the company’s largest and most critical data centre hubs in Northern Virginia. The ripple effects quickly spread across continents, affecting consumer apps, financial institutions, and enterprise platforms that rely on AWS for backend infrastructure.

‘Even global providers are not immune to disruption’

“The AWS outage serves as a reminder that even the most established global cloud providers are not immune to disruption,” Robert Azzopardi, Chief Technology Officer at Smart Technologies tells MaltaCEOs.mt. “While such incidents are rare, they highlight the importance of resilience, redundancy, and local fail-over strategies.”

He notes that Smart Technologies continuously advises clients to adopt architectures that can withstand cloud-level failures. “At Smart Technologies, we continuously advise our clients to build architectures that can withstand cloud-level failures through local hosting, hybrid deployments, and multi-region designs. It is about business continuity, not just uptime,” Mr Azzopardi adds.

This approach ensures that essential operations can continue even if a global service provider experiences technical issues – something that is becoming increasingly vital as businesses migrate critical workloads to the cloud.

Building resilience through hybrid and multi-cloud strategies

For Sean Cohen, Chief Service Delivery Officer at BMIT, the incident highlights just how deeply interconnected modern digital services have become. “Today’s AWS disruption serves as a reminder of how critical digital services have become. Most organisations rely on globally hosted applications and SaaS platforms for core business functions, for example for identity management, collaboration, analytics and access to customer portals,” he explains.

He points out that the effects of such outages often extend far beyond a single provider or region. “Even though hyperscalers operate multiple regions for redundancy, these environments are deeply interconnected, and disruptions in one region can influence others, as happened in this case,” he says.

From the consumer side, the effects are felt almost instantly. “Many everyday digital experiences, such as communication, creative work, online shopping, media streaming, and mobile payments, ride on the same global backbones,” Mr Cohen continues. “When those platforms slow down or fail, people notice it as service not available, login errors, delayed notifications, stalled checkouts, or content that will not load.”

He emphasises that resilience “should not be assumed but must be architected and planned for.” Relying on a single provider introduces concentration risk – a lesson many businesses are learning the hard way.

“This is why we at BMIT advocate a hybrid and multi cloud approach,” he adds. “By combining local infrastructure and cloud platforms in Malta with global cloud services, businesses retain control over their most critical workloads and data while still benefiting from the scalability and innovation of hyperscale environments.”

According to Mr Cohen, hosting or replicating core systems locally also enhances sovereignty, compliance, and recovery capabilities. “True resilience comes from deliberate design: diverse connectivity paths, clear operational runbooks, regular failover testing, and applications built with retry logic, queues, and timeouts,” he explained.