Caching & Aggregation

Concept#

Request Aggregation#

When rendering a dashboard, a client might need:

1. User Profile (GET /users/me)
2. Notification Count (GET /notifications/unread)
3. Project List (GET /projects)

Executing these sequentially from a mobile browser over cellular networks adds multiple high-latency connection roundtrips.

In a BFF architecture, the client fires a single consolidated request to the BFF gateway. The BFF then executes the downstream requests concurrently over local networks, aggregates the payloads, and returns a single merged response:

// Client requests dashboard data from BFF:
GET https://api.app.com/bff/dashboard

Concurrent Fetching in Node.js#

To avoid blocking execution, always execute downstream requests concurrently using Promise.all() (or Promise.allSettled()) rather than executing them sequentially:

// ❌ WRONG: Sequential requests block execution
const user = await fetchUser();       // Takes 100ms
const orders = await fetchOrders();   // Takes 100ms (Total: 200ms)