Push Replication Enables Horizontal Scale
This excerpt is chapter 5 from Enea’s white paper «Scalable Database Design for 5G and Beyond», published in April 2026. The whole paper can be accessed below.
In many distributed deployments, data replication introduces an operational tax, including backlogs of write changes, periodic catch-up events, and persistent uncertainty around cross-site synchronization. Expanding to additional datacenters can increase this overhead, and in poorly designed systems, the cost of replication can grow unpredictably, creating bottlenecks and limiting scalability.
Stratum addresses these challenges with scalable push-based replication, in which updates are immediately sent (“pushed”) from each replica to its peers as soon as they are committed locally. Replication is “fire-and-forget”: messages are sent without maintaining a replay history, and updates are applied independently on each replica. This design keeps the common successful replication path simple and efficient. As the replication processing time on the receiving side is negligible in comparison to the network latency, the overall replication lag is effectively equal to that one-way network delay. When replication messages are lost due to a network or process failure, inconsistencies are resolved by the continuous anti-entropy process (covered in the next chapter) rather than by explicit replication mechanisms.
The predictable behavior of this replication approach ensures that horizontal scaling comes with a measurable and bounded replication cost. Operators can expand deployments with new sites and accurately estimate the capacity gains, while understanding how returns diminish as the system scales, all without encountering unexpected replication lag, processing stalls, or nonlinear scaling penalties.
Learn More About Enea Stratum
Enea’s Stratum is the purpose-built cloud-native distributed database that puts these principles into practice. Already deployed by Tier-1 operators in North America and Europe, it handles hundreds of millions of subscriber records while delivering millions of transactions per second with telco-grade low latency and high availability. The difference between good enough and truly exceptional network performance increasingly comes down to the database layer operating beneath the surface. Make sure your infrastructure is ready for what’s next.
If you’d like to explore how Stratum can support your 5G and future 6G network requirements, visit:
