SAP Beyond Uptime: Engineering Intelligent AMS with High Availability & DR through Pacemaker Automation

Abstract

For two decades, the enterprise reliability discipline has myopically prioritized the statistical artifact of "five nines" availability, treating uptime as the ultimate proxy for system health while neglecting the structural fragility of the underlying automation. This theoretical stagnation has resulted in a "Manual Intervention Paradox," wherein increasingly complex SAP architectures rely on brittle shell scripts and human intuition for critical failover decisions, reducing "Intelligent Application Management Services" (AMS) to little more than reactive ticket clustering. Addressing this methodological crisis, this study proposes a "Unified Control Theory" that reclaims the physicist’s definition of AM simultaneous sensing and manipulating by integrating the formal verification standards of medical device engineering into the SAP Pacemaker stack. Through the development of a custom State Verification Engine (SVE) grounded in timed-automata models, we subject a distributed SAP HANA landscape to stochastic fault injection, contrasting the behaviour of standard resource agents against this formally verified logic. Results indicate that while the proposed architecture introduces a requisite latency penalty, it successfully prevents fatal "split-brain" scenarios in complex failure modes were legacy automation invariably corrupted data. Ultimately, this work redefines Intelligent AMS from a bureaucratic support function into a rigorous control plane, demonstrating that the future of enterprise resilience lies not in the pursuit of raw speed, but in the engineering of automated survival through epistemological certainty.