The authors of the article conducted a benchmarking test comparing the performance of LevelDB, RocksDB, HyperLevelDB, and LMDB storage engines for InfluxDB. They tested these engines on a Digital Ocean droplet with 4GB RAM, 2 Cores, and 60GB SSD storage. The test involved writing, querying, deleting, and compacting data in batches of 1,000 key/value pairs, simulating a common load pattern for InfluxDB. The results showed that LevelDB excelled on disk space utilization, RocksDB was the winner on reads and deletes, HyperLevelDB outperformed LMDB on writes, and LMDB had issues with compaction due to its lack of compression and inability to reclaim disk space after deletions. Based on these findings, RocksDB appears to be the best choice for the use case, but it's essential to consider hardware configuration and settings when selecting a storage engine.