ShardingSphere-Proxy

ShardingSphere-Proxy is self-hosted middleware that presents distributed databases through a database-server protocol, letting applications in different languages connect without changing their basic client model. It is best for infrastructure teams coordinating several database backends, not for users seeking a managed database service. Its broad routing, sharding and transaction toolkit is compelling, provided you can operate the runtime and accommodate its system-database limitations.

Overview

Proxy sits between applications and underlying databases rather than replacing them. ShardingSphere, founded in 2018, uses the proxy to coordinate storage, queries and connections across a distributed environment. Deployment is available as a binary distribution, Docker image or Helm deployment, all under a self-hosted model. That gives Linux, macOS and Windows users control over placement, but also leaves installation, upgrades and availability operations to their team.

The binary requires Java JRE 8 or newer. Docker avoids an additional dependency. MySQL backends require adding mysql-connector-j-8.4.0.jar to the Proxy ext-lib directory, an extra setup step that matters in repeatable deployments.

Key features

Sharding and distributed transactions

Sharding distributes storage and computing across underlying databases, making Proxy a fit when horizontal growth is more important than keeping all data on one server. Its transaction engine combines XA and BASE approaches for work spanning data sources. That is useful for distributed workloads, but it is specialized infrastructure rather than a simpler single-database choice.

Routing, pooling and failover

Read/write splitting balances reads using SQL semantics and awareness of database topology. Query routing, connection pooling, load balancing and automatic failover give operators controls for directing traffic and maintaining service across backends. These features are most valuable when a team already manages replicas or multiple database nodes.

Migration, federation and protection

Migration tools support moving data in sharded environments, while federated queries can combine data from different sources. Transparent encryption is provided through an encryption module, and data masking is included as a protection area. Together, these capabilities extend Proxy beyond traffic forwarding into data-movement and governance work.

Protocols and compatibility

The quick-start documentation says client operations are theoretically supported through MySQL, PostgreSQL, openGauss and Firebird protocols. Compatibility is therefore broad on paper, but support for system databases and tables such as information_schema and pg_catalog is limited. Some graphical database clients may consequently display errors, an important drawback for administrators who rely on those tools.

Pricing

Apache ShardingSphere-Proxy — free. The free plan includes the open-source binary distribution plus Docker and Helm deployment. It has no paid tier, trial period or stated usage cap, and the deployment model is self-hosted. This suits teams able to provide their own compute, storage, operations and support. It is less suitable if you expect a hosted service or vendor-managed updates.

Release files should be checked with GPG or SHA-512 signatures before deployment. That verification step is sensible for production supply-chain hygiene, particularly when installing binaries directly.

Platforms

ShardingSphere-Proxy supports Linux, macOS, Windows and self-hosted installations. Docker removes local runtime dependencies, while the binary route requires Java JRE 8 or higher. Supported backend protocols are MySQL, PostgreSQL, openGauss and Firebird.

Who it's for

Choose Proxy if you need one protocol-facing layer for sharding, read/write routing, distributed transactions, migration or federation across several databases. It fits platform and database teams comfortable running Java-based or containerized infrastructure and tuning their own topology.

Look elsewhere if you need a managed offering, a straightforward single-database deployment, or dependable graphical-client access to system catalogs. The connector requirement and catalog limitations add operational friction for otherwise simple environments.

Pros and cons

• Pros: horizontal sharding spreads storage and compute; XA/BASE transactions address cross-database work; SQL-aware read/write splitting, pooling, load balancing and automatic failover cover core proxy duties; migration and federation support broader data operations; binary, Docker and Helm packaging gives deployment choice; free open-source distribution avoids license fees.
• Cons: self-hosting means your team operates the service; Java JRE 8+ is required for binary deployment; MySQL needs a separately placed connector JAR; limited system-database support can produce errors in graphical clients; protocol support is described as theoretical, so compatibility should be validated for your workload.

Alternatives

For a general database platform with SQL and NoSQL applications plus replication and high-availability components, consider MySQL; its Community Edition price is custom pricing and it also has a free trial.

OceanBase Database is the better direction when you want an open-source distributed database with full kernel capabilities, with a free Community Edition and a custom-priced Enterprise plan.

Pick PgDog when an AGPL-3.0 PostgreSQL proxy with an open-source tier is the priority; its Enterprise plan adds a production control plane and insights.

For a focused self-hosted PostgreSQL connection pool, choose pgagroal. PgBouncer is another free self-hosted option for PostgreSQL pooling. PgDoorman offers free self-hosted PostgreSQL proxying with no paid plans or stated usage limits.

TiCDC is a free open-source choice for change-data capture. TurbineProxy suits teams wanting an Apache-2.0 open-source proxy with no per-query fees.

Browse the Database Proxy Software list for more options.

Verdict

ShardingSphere-Proxy is a strong choice for teams that need self-hosted coordination of sharded, replicated or federated databases and can manage the operational complexity. Its main reason to choose it is the breadth of distributed routing and transaction capabilities at no license cost. Look elsewhere when managed hosting, uncomplicated setup or full system-catalog compatibility matters more than that flexibility.