Zeus Kerravala Resilience Architecture
6 steps1 pieces2 Sources
ZK Research Principal Analyst Zeus Kerravala argues that high availability metrics do not equate to business resilience, a distinction highlighted by six major IT outages in November 2026. In a SiliconANGLE guest column, Kerravala uses Zscaler’s Business Continuity Cloud as a case study to demonstrate that while multi-availability zone designs protect against component failures, they often share control planes and software dependencies that allow single points of failure to cascade globally. He contends that true resilience requires 'fate separation'—architectural conditions where the failure of one service does not halt the entire business operation. The disagreement lies in whether cloud providers’ shared responsibility models sufficiently address these systemic risks or if enterprises must build independent recovery layers. Kerravala suggests that treating disaster recovery as an afterthought rather than a core design principle is the primary vulnerability exposed by recent incidents involving Microsoft, Workday, and the FAA.
How it got here
Nov. 12: Microsoft 365 experiences a nearly 24-hour global outage affecting 74 countries due to a dormant software bug introduced in July.
siliconangle.com ↗Nov. 13: Workday suffers a major outage impacting over 10,000 companies after its disaster recovery environment inadvertently connected to production databases.
siliconangle.com ↗Nov. 13: The U.S. Federal Aviation Administration issues a nationwide ground stop following a telecommunications failure in air traffic control centers.