Katkat
Back

Design 64-bit distributed IDs and a parking lot

FlipsterSoftware Engineer · Technical Screen
EasySoftware EngineerTechnical ScreenSystem Design

Last updated

The question evaluates distributed-systems and scalability competencies for globally unique 64-bit ID generation (including sharding, time-ordering, fault tolerance and clock issues) as well as object-oriented design, concurrency control, API design, and data modeling for a parking lot backend.

You have two system-design prompts.

1) 64-bit database ID generator

Design a service/library that generates globally unique 64-bit integer IDs for database records in a distributed system.

Requirements

  • IDs must be unique across all shards/regions.
  • Very high write rate (assume 100K–1M IDs/sec globally).
  • Prefer IDs that are roughly time-ordered to reduce index fragmentation.
  • Support sharding and growth in number of nodes.
  • Discuss how to avoid conflicts and what happens on failures (node restart, clock issues, network partitions).

Deliverables

  • Bit layout proposal (what each bit range means).
  • How nodes obtain any needed configuration (e.g., worker IDs).
  • How the system behaves under clock skew/rollback.
  • Operational considerations (monitoring, rotation, migrations).

2) Parking lot system (object-oriented + backend)

Design a parking lot management system.

Requirements (assume typical)

  • Multiple levels; each level has multiple spots of different types (compact/large/EV/handicapped).
  • Vehicles arrive/leave; system issues a ticket and calculates fees.
  • Track spot availability and prevent double-allocation under concurrency.
  • Provide APIs for: entering, exiting, querying availability, and (optionally) reserving a spot.

Deliverables

  • Core classes/entities and relationships.
  • Key APIs and data storage approach.
  • Concurrency/consistency approach for assigning/freeing spots.