Application Portfolio Management (APM) Best Practices

Overview

Organizations migrating applications to cloud rarely migrate the entire portfolio simultaneously, and some applications may never be practical cloud migration candidates. The result is a persistent hybrid architecture in which some applications run on-premises and others run in cloud, and many critical integration relationships span the boundary between the two environments. Hybrid integrations - data flows, API calls, event streams, and batch processes crossing the on-premises to cloud boundary - introduce latency characteristics, security complexity, network cost structures, and operational monitoring challenges that do not exist in purely on-premises or purely cloud architectures. The hybrid integration layer is frequently the most complex, most poorly documented, and most operationally fragile aspect of the modern enterprise application portfolio.

Best Practice

Map and actively govern the hybrid integration topology of the application portfolio as a portfolio management discipline, not only as an architectural concern managed by individual application teams. For every integration that crosses the on-premises to cloud boundary, document the integration mechanism, the data flows and their business criticality, the latency and bandwidth requirements and how they compare to actual hybrid network performance, the security controls in place for data in transit across the boundary, and the operational monitoring arrangements providing visibility into hybrid integration health. Assess hybrid integrations as a mandatory input to every migration decision - migrating an application to cloud without accounting for its hybrid integration dependencies frequently produces integration failures that are more expensive to remediate than the migration cost they were intended to avoid.

Benefit(s)

Governing hybrid integration complexity at the portfolio level prevents the most common failure mode of cloud migration programs: discovering that on-premises-to-cloud integrations are significantly more complex, costly, and operationally fragile than anticipated because they were not assessed as part of migration planning. Organizations with a current, governed hybrid integration map make more accurate migration cost estimates, sequence migrations to manage integration dependencies intelligently, and operate their hybrid architectures with the visibility needed to detect and address integration performance and reliability issues before they affect dependent business operations.