Digital technology and infrastructure
Digital technology and infrastructure tests the systems that carry orders, data, payments, and records: APIs, messaging, cloud, latency, resilience, and security. Questions focus on design trade-offs and failure modes, not vendor names.
Architecture and interfaces
- APIs should authenticate, authorize, validate input, and enforce idempotency for retried writes.
- Message queues decouple producers and consumers; exactly-once delivery is rare, so consumers must deduplicate.
- Cloud shared-responsibility models split physical security from customer configuration, identity, encryption, and data governance.
Performance and resilience
Latency is elapsed time; throughput is work per unit time. A latency budget allocates milliseconds across market data, strategy, risk checks, gateway, and network. Availability of 99.9% allows about 8.76 hours of downtime per year; 99.99% allows about 52.6 minutes.
Worked example
An end-to-end order path has a 12 ms budget: market data 2 ms, strategy 1 ms, pre-trade risk 3 ms, order gateway 2 ms, network 4 ms. A new risk check adds 2 ms, taking the total to 14 ms. To stay within budget, the team must remove 2 ms elsewhere or accept later execution. Adding hardware without changing the slowest stage may not help.
Traps
Confusing latency with throughput; retrying payments without an idempotency key; assuming cloud providers secure customer data by default; treating blockchain immutability as proof of accuracy; ignoring time synchronization across logs.