Gravty is a scalable graph database designed to efficiently search relational information in large datasets using graph techniques. It uses a tall-narrow data model in HBase to improve query performance compared to a flat-wide model. Gravty also implements techniques like custom row salting and pre-splitting to avoid hot-spotting, uses asynchronous indexing with Kafka to scale indexing, and includes monitoring and recovery tools to improve reliability. Future plans include adding OLAP functionality and supporting multiple graph clusters.