Explain Snowflake IDs vs ULIDs vs UUIDs.

Snowflake IDs, ULIDs, and UUIDs are unique identifier strategies used in distributed systems, each balancing size, orderability, and simplicity differently.

When to Use

• Snowflake IDs: Great for distributed services needing time-ordered, compact IDs (e.g., social media posts).
• ULIDs: Useful when you need lexicographically sortable IDs without central coordination.
• UUIDs: Best for universally unique IDs in systems where order is irrelevant.

Example

Twitter assigns every tweet a Snowflake ID so they can be sorted chronologically; using random UUIDs would break this natural ordering.

Want to master these concepts?

Explore Grokking System Design Fundamentals, Grokking the System Design Interview, or try Mock Interviews with ex-FAANG engineers.

Why Is It Important

The choice of ID scheme affects database indexing, query performance, and scalability. Time-ordered IDs minimize fragmentation, while random ones may degrade performance.

Interview Tips

Be prepared to explain differences succinctly:

• Snowflake: time-ordered, efficient, but clock-dependent.
• ULID: sortable, simple, but larger (128-bit).
• UUID: universal, but random and index-heavy. Show how this impacts real-world system design trade-offs.

Trade-offs

• Snowflake: + small & ordered; – requires clock sync.
• ULID: + sortable & decentralized; – bigger size.
• UUID: + simple & supported; – random order slows inserts.

Pitfalls

• Don’t use random UUIDs as DB primary keys for high-write workloads.
• Ensure time sync for Snowflake generators.
• Remember ULIDs may collide if many generated in the same millisecond.