Design Considerations of Yelp
In this lesson3 sections
Design Considerations of Yelp
Introduction
The architecture leaves two separate problems: finding every place that could satisfy a geographic query, and ranking those candidates. This lesson compares fixed cells with a quadtree, then considers placement, updates, and failure recovery.
Summary of the Lesson
| Topic | Design question |
|---|---|
| Fixed segments | Which cells intersect a radius query, and how much candidate data must we read? |
| Dynamic segments | How should dense cells split, and when can the search safely stop? |
| Storage and placement | Where do spatial indexes and authoritative place records live? |
| Updates and recovery | How do new listings become searchable, and how can a failed index be rebuilt? |
| Ranking | Which scoring rule and freshness target define the final results? |
Searching
The Maps case study partitioned road data to reduce routing work. A proximity directory can also use geographic partitions, but it searches points within an area rather than shortest paths through roads. We do not need to connect every candidate place through a road graph for a straight-distance radius query.
The rest of this lesson is for Premium members
Premium opens every lesson in this course, along with all questions and solutions on PracHub.