Tarantool Clusters Federation

Manage switching between independent Tarantool clusters in the current architecture with no downtime using the module for creating a disaster-resilient infrastructure.

Difficulties in architecture solutions with two data centers

Classic clusters architecture

The Active — Passive — Replica scheme is distributed across two data centers. A backup data center is needed for minimizing the negative consequences in case of a human-caused or natural incident.

Scaling, migrating, and changing the data schema and model

Any change in architecture causes changes in software, code, schema, or the data itself.

Downtime for high-load systems in the classic architecture

The timing for maintenance and troubleshooting, and handling of personal information will be regulated by law starting from 2023.

Solve architecture problems with Tarantool Clusters Federation

Architecture layout of Clusters Federation in DCs

Tarantool Clusters Federation provides near real-time performance and fault tolerance up to 99.999% without any need to buy new servers even if one of them fails. Master-Replica clusters are located in two different data centers.

How the Master-Replica layout works: Tarantool Clusters Federation installs two Tarantool clusters at once in the current data centers.

The Master cluster is active in one data center. Its Replica, in turn, runs in the passive cluster of the second one.

The Clusters Federation module enables building a disaster-resilient configuration between two independent Tarantool clusters. In fact, an additional layer is built over the replication layer; it protects against any disasters.

scroll-scheme