Building the future of cloud infrastructure—enabling businesses to seamlessly migrate, replicate, and optimize workloads across multi-cloud environments is the name of the job for Harshit.

Previously the first engineer at Accurics, where he led core development efforts on its policy engine and cloud security platform, Harshit brings expertise in Go, Kubernetes, Terraform, and cloud compliance. He has spent over a decade designing resilient systems across AWS, Azure, and GCP and now has a mission to eliminate cloud lock-in and make infrastructure as portable and resilient as code.

In our quick conversation, Harshit discusses how outages aren’t isolated events — they highlight a bigger structural issue in how enterprises depend on the cloud and are dramatically underestimating their risk exposure.

M.R. Rangaswami: Why are we seeing a rapid rise in large-scale cloud outages, and what do they reveal about systemic fragility in today’s cloud monoculture?

Harshit Omar: What we are seeing right now is not a series of isolated incidents. It is the breakdown of a long-held industry assumption that hyperscalers are effectively too big to fail. They are not. At this scale, even routine configuration issues ripple across millions of workloads instantly. AWS, Azure, and Cloudflare all experiencing issues within weeks is not a coincidence. It is a structural warning.

We called this trend years ago. The hyperscaler model has concentrated compute, storage, networking, AI workloads, and identity into centralized control planes with enormous blast radius. As the global workload footprint grows, the fragility grows with it. Outages are more common because complexity has exploded and enterprises have almost no buffer when the core infrastructure falters.

M.R.: Despite a decade of talk about multi-cloud, why do most enterprises still depend on a single cloud, and how does this amplify the impact of recent outages?

Harshit: Multi-cloud became an idea rather than an architecture. Many leaders said they wanted it, but the practical tooling to make it possible without rebuilding everything did not exist. Most organizations could not justify duplicating pipelines, infrastructure definitions, security policies, or observability frameworks in multiple places. Over time they drifted into single-cloud dependency without fully realizing it.

This is why outages feel so catastrophic. When AWS or Azure goes down, organizations do not simply wait. They are stuck. They cannot move workloads or fail over quickly because their environment truly exists in only one location. It is the digital version of building an entire city that depends on a single bridge.

This is the fragility we have been highlighting for years. The cloud became so successful that many teams forgot a basic rule of cloud-native architectures: resilience only exists when redundancy is created across boundaries rather than within them. 

M.R.: What does a more resilient cloud architecture look like, and how can approaches like Cloud Cloning help enterprises respond instantly when a provider goes down?

Harshit: True resilience means your entire environment can run somewhere else without rewriting it. That is the standard enterprises now need. Until recently, that standard was out of reach for most organizations.

Cloud Cloning was created because we knew this moment was coming. Enterprises should not have to refactor workloads, rebuild pipelines, or recreate identity and access systems just to gain optionality. A cloned environment should be able to run in another cloud the moment you need it. Same architecture. Same configuration. Same governance. Just operating on a different provider.

If outages are becoming routine, portability becomes the primary defense strategy. The next generation of cloud architecture will not ask which cloud you are on. It will ask how quickly you can move when the cloud you depend on stops working.

M.R. Rangaswami is the Co-Founder of Sandhill.com