Logical scalable units (LSU) can be used within a single network data center to provide stateful scalability. However, LSUs are not suitable for operating across multiple network data centers in order to provide geographical redundancy in active-active scenarios while providing high availability. This is because the latencies associated with replicating full session information are unacceptably slow. An additional component in the LSU (known as an LSU Frontend) can be used to replicate a small subset of session information between LSUs. This subset of session information may be enough to enable LSUs to process requests in an acceptable way rather than outright failing. This may be particularly advantageous in scenarios where the requests relate to Voice over LTE calls, because outright failure is very perceivable to subscribers, and it causes resource leakage within the telecommunications network.