Managing External API’s in Enterprise Systems
summary
A real-world Bloomberg rewrite encapsulates a sprawling external request/response API by wrapping it in a single service proxy that preserves the original send/get semantics while converting callbacks to future/promise returns and making responses move-only results (unique_ptr) to avoid hidden copies. A variant-of-futures approach lets the proxy route between old Service A and new Service B, enabling phased rollout and simple decommissioning; the migration spans months to years with unit tests using injected mocks to verify outbound request formation.
AI-written summary · about
chapters (23)
- 0:00Start
- 7:39System Pipeline
- 8:03Full System
- 8:46External API
- 10:05Wrapper Design
- 14:10Application
- 22:40Requirements Change
- 23:18System Slice
- 23:49Straight Swap
- 26:26Phased Approach
- 27:12Brute Force
- 30:23Prioritize Migration Path
- 33:55Proxy Design
- 36:01Future semantics
- 38:01Utility
- 40:06Result Semantics
- 41:38Point of use
- 44:24Result Analysis
- 45:23Testing
- 50:54Decommission
- 54:02Reality Check
- 55:29Change Priorities
- 57:32Endgame
related talks
50mA Guide to Managing External APIs in Enterprise Systems
16mBuilding Future-Proof API Platforms
32mMulti-form API Management
23mDecentralized API Management for Agile Enterprises
23mLeverage APIs to Transform Your Business Landscape
29mAdapting to an API Enabled World (Panel Debate)
38mAdvanced API Design for Scalable Distributed Systems
20mThe Dark Side of Managing APIs as a Product