Day 18 — Machine Learning System Design: Estimate Food Delivery System Design
Discover the Design of the Estimated Food Delivery System for Delivery Apps.
Computation & Approximation
Assumptions
To simplify our discussion, let’s assume the following:
- The platform engages 2 million active users monthly, with a user base of 20 million in total.
- There are 300,000 participating restaurants and 200,000 drivers facilitating the deliveries.
- Annually, the service handles approximately 20 million deliveries.
Data Volume
- Over one month, we amassed data from two million deliveries. Each delivery is associated with approximately 500 bytes of feature-related data. Consequently, the total data size is calculated as follows:
- 1 Gigabyte from 500 * 2 * 10⁶ = 10⁹
System Design
Feature Store: This component ensures rapid lookup times for low latency, ideal for a feature store utilizing key-value storage that boasts high availability — Amazon DynamoDB…