Implementing a Bloom Filter in Spring Boot with Redis for Fast Membership Checks

Bloom filter, a space‑efficient probabilistic data structure proposed by Burton Bloom in 1970, can test set membership with a controllable false‑positive rate but no false negatives, making it suitable for scenarios that tolerate occasional errors.

In a Spring Boot 2.3.8.RELEASE project with Redis, you can integrate a Bloom filter using Guava and Redis commands. Add the necessary Maven dependencies (spring-boot-starter-data-redis, spring-boot-starter-web, commons-pool2, guava 30.1-jre).

Configure Redis connection and Bloom filter parameters (expected insertions, false‑positive probability) in

application.yml

and bind them to a

BloomFilterProperties

class via @ConfigurationProperties.

Implement the core

RedisBloomFilter

component: calculate optimal bit size and hash function count, hash values with Murmur3, set bits in Redis with

setBit

, and check membership with

getBit

. Helper methods convert byte arrays to long values.

Example REST endpoints demonstrate inserting 10,000 random UUIDs into the filter and measuring false‑positive rate by querying 100 new UUIDs; results show error rates around the configured 1%.

Images illustrate the Bloom filter concept and test outcomes.