TBase

TBase is Tencent-published, open-source distributed HTAP database software that provides database and data warehouse services through one cluster. It supports an extended subset of SQL, including transactions, foreign keys, user-defined types, and functions, and has PostgreSQL compatibility. A cluster uses Coordinator Nodes to plan queries, DataNodes to store user data, and a Global Transaction Manager. Tables can be distributed or replicated; queries are sent to DataNodes and results collected. The self-hosted deployment guide describes compiling from source and running a cluster across one or more hosts, including basic x86 servers. Administrators can use pgxc_ctl to configure and manage clusters. Table options include shard, partitioned shard, hot-and-cold partitioned shard, and replication tables. Shard-table distribution-key values, lengths, and types cannot be changed. Replication tables keep full copies on every DataNode, which the guide says makes them unsuitable for large tables and low in update performance. Monitoring includes session memory, DataNode operator details, and cluster-wide active sessions. The listed plan is free under the BSD 3-Clause License.

TBase offers self-hosted distributed HTAP capabilities with several table arrangements and monitoring tools. Teams should account for the fixed shard distribution keys and replication-table constraints.

TBase suits teams seeking self-hosted distributed database and data-warehouse services in one cluster. It may fit workloads needing PostgreSQL compatibility, horizontal scaling, and HTAP support.

TBase is a free, open-source database that combines distributed transactional and analytical services in one self-hosted cluster. It best suits teams prepared to compile, deploy and administer database infrastructure themselves. Its broad table options and monitoring capabilities are balanced by schema commitments and replication limits that call for careful planning.

Overview

TBase brings database and warehouse workloads together in a cluster built from Coordinator Nodes, DataNodes and a Global Transaction Manager. Coordinators plan queries, DataNodes store user data, and the system routes queries to those nodes and gathers results. That architecture supports horizontal scaling, but also means teams take on the work of operating a distributed system rather than adopting a managed database service.

The database supports an extended SQL subset that includes transactions, foreign keys, user-defined types and functions. Its language interfaces are similar to PostgreSQL’s, a practical advantage for teams familiar with that ecosystem; this does not establish complete PostgreSQL feature compatibility.

Key features

Administrators can choose shard tables, partitioned shard tables, hot and cold partitioned shard tables, or replication tables. This range accommodates different data layouts, but shard tables impose a consequential design constraint: the distribution-key field’s value, length and type cannot be modified. Teams should settle that key before depending on the table structure.

Replication tables keep a full copy of each table on every DataNode. That can suit smaller tables where replicated data is useful, but the approach is unsuitable for large tables and has low update performance. It is not a general substitute for distributing data across nodes.

Monitoring features include session-memory views, execution details for DataNode operators, and a cluster-wide view of active sessions from a coordinator. These tools give operators visibility into resource use and query activity across the cluster. The pgxc_ctl tool helps configure and manage clusters, while the maker’s quick-start guide describes compiling from source and deploying across one or more hosts, including basic x86 servers.

TBase also includes security and audit functions. It is released under the BSD 3-Clause License, with a listed exception for a third-party component.

Pricing

TBase: free. This is a free, open-source database under the BSD 3-Clause License. There is no paid tier described here; the trade-off for the zero-price software is that deployment and cluster administration remain the operator’s responsibility. The available plan information gives no seat or usage quota to weigh against that price.

Platforms

TBase is self-hosted and supports Linux, macOS and Windows, as well as Android and iOS. Its mobile-development support includes C++, Java, Kotlin, Swift and Objective-C, with Unity and UE4 among the supported game engines. The product also supports hybrid test environments and automated test execution, but does not provide playtesting workflows.

Who it's for

TBase is a sensible fit for teams that need distributed SQL and HTAP workloads in one cluster, want to operate their own infrastructure, and can make distribution-key decisions early. PostgreSQL-oriented teams may find its similar language interfaces familiar. It is a weaker fit for organizations seeking a managed deployment, teams with large tables that would otherwise rely on replication tables, or projects that need to change shard distribution keys later.

For broader comparisons, see Distributed SQL Databases, Mobile Databases and Game Testing Software.

Pros and cons

Pros

• Free and open source: The free plan avoids a software subscription charge.
• Transactional and analytical workloads in one cluster: The HTAP approach can serve teams that want both workload types within one distributed system.
• Several table arrangements: Shard, partitioned, hot-and-cold partitioned and replication tables offer distinct ways to organize data.
• Operational visibility: Session-memory, operator execution and cluster-wide active-session views give administrators concrete monitoring tools.

Cons

• Self-hosted operations: Teams must compile and deploy the software and manage the cluster themselves.
• Fixed shard distribution keys: The key’s value, length and type cannot be changed, making an early schema choice difficult to reverse.
• Replication-table trade-offs: Full copies on every DataNode make this option unsuitable for large tables and slow for updates.

Alternatives

Choose MySQL instead if a freemium database with a free plan, free trial, and replication and high-availability components better matches your SQL or NoSQL application needs.

CrateDB is another freemium option; its Shared Free plan includes 2 vCPUs, 2 GiB RAM and 8 GiB storage. That makes it worth considering when those stated resource limits suit the workload.

TiDB Cloud may fit teams looking for a cloud option with a free Starter plan and trial. Its stated quotas are 25 GiB row storage, 25 GiB column storage and 250M R.

CockroachDB Cloud offers a paid Standard plan at $0.06 a month, with cloud usage metered by the minute and rates varying by region. It is an alternative for readers considering usage-metered cloud billing.

OceanBase Database has a free, open-source Community Edition with full kernel capabilities, and is a relevant alternative for teams comparing distributed database options.

SingleStore Helios offers a free Shared workspace for evaluation, development and non-production testing, with self-support. It is the clearer fit when a limited evaluation workspace is the priority.

VoltDB has a free Developer Edition for individual developers and early exploration, making it a better match for those narrower use cases.

YDB offers a free, Apache 2.0-licensed open-source edition that can be self-hosted, so it merits consideration for teams comparing self-managed open-source databases.

Verdict

TBase is best for teams that want a free, self-hosted distributed HTAP database and are equipped to manage its cluster. Its combination of table choices, PostgreSQL-like language interfaces and monitoring makes it a capable infrastructure option, but fixed shard keys and replication-table limits demand deliberate design. Look elsewhere if managed operations or schema flexibility outweigh self-hosting and control.