Baidu Tera vs Elassandra

Baidu Tera is a free, self-hosted distributed NoSQL database for real-time applications. It stores sparse, multidimensional tables indexed by row key, column key and timestamp, with the project describing scale to petabytes across thousands of commodity servers. Tera supports strict consistency, MVCC, automatic failover, online schema changes and snapshots. Configurable automatic sharding can use ranged or hashed strategies, with linear and modular scalability. Storage options include RAMDISK, SSD and disk tables, plus tiered and block caches and Bloom Filters for real-time queries. Applications can connect through C++, Java, Python and REST APIs. Its architecture includes an SDK, master and tablet servers using an enhanced LevelDB engine. Deployment can use raw machines or Docker, with pseudo-distributed operation on one host. The repository lists BSD-3-Clause and Apache-2.0 licenses.

Elassandra combines Apache Cassandra’s distributed storage with embedded OpenSearch search on each node. Its maintained line targets Apache Cassandra 4.0.x, OpenSearch 1.3.x, and Java 11. OpenSearch indices map to Cassandra keyspaces, documents to rows, and indexed fields to columns or cells; the current _doc model maps to a Cassandra table. Cassandra remains the durable layer, allowing restart recovery to replay Cassandra state and rebuild a local search view. Applications can search through the OpenSearch HTTP API or CQL, with HTTP responses following OpenSearch style. Data placement uses Cassandra token ownership, while replicas follow keyspace replication. Schemas can be created from OpenSearch indexes or discovered from existing Cassandra schemas through mappings. Integrations listed include OpenSearch plugins, OpenSearch Dashboards, and Spark. Deployment options include tarball, Debian package, RPM, Docker image, and Helm chart. Elassandra is free and open source, with a 30-day trial noted.