Building Reusable 3DS Authentication Infrastructure

Card authentication sits directly on the conversion path: every handoff must preserve state, security, latency expectations, and an explainable result for downstream systems.

Air India provided the first high-volume production use case, while the integration itself needed to remain extensible for additional authentication providers and payment experiences.

The platform needed to introduce a shared authentication service into an established transaction path without weakening existing payment behavior.

A merchant-specific implementation would have duplicated contracts and data movement each time a new authentication product was onboarded.

Payment correctness and secure data boundaries were non-negotiable.

Authentication calls needed explicit timeout, retry, and status-mapping behavior.

The contract had to remain observable and reusable without exposing sensitive card data.

Production readiness required realistic load and failure-path validation.

Led development across the shared API contract, transaction integration, authentication orchestration, data-pipeline touchpoints, load testing, and rollout readiness.

Worked across system boundaries so the first merchant launch also established a reusable integration pattern for subsequent authentication initiatives.

Modeled authentication as explicit transaction stages, separating validation, provider interaction, gateway handoff, and final status mapping.

Defined stable contracts between checkout-facing systems, the core transaction layer, and authentication services, including failure semantics and auditable state transitions.

Integrated a shared data pipeline for downstream analytics and operational monitoring while keeping sensitive card data outside observable payloads.

Validated the path with load tests shaped around realistic traffic, latency, and provider-failure scenarios.

A reusable platform boundary required more contract design than a merchant-only path, but reduced the cost and risk of future authentication onboarding.

Richer status modeling added implementation work while making production debugging and incident response safer.

The integration supports 250K+ monthly Air India card transactions.

Other teams can build on the same data and orchestration boundaries for additional authentication products; the pattern is already being extended toward flows such as Mastercard Click to Pay.

The durable output of a large integration is not only the launch; it is the contract and operating model that make the next integration safer.

Authentication quality is defined as much by failure semantics and observability as by happy-path latency.