crossDC

CrossDC, short for cross data center, refers to the architectural approach of deploying and operating services across multiple geographically separated data centers. The goal is to increase availability, resilience, and user experience by reducing latency and providing disaster recovery options in the event of a regional outage. CrossDC deployments typically include geo-distributed storage and compute with data replicated between data centers under defined consistency and latency constraints.

Implementation choices center on replication modality and topology. Replication can be synchronous, where writes are acknowledged only after being durably stored in a majority of participating data centers, offering stronger consistency at the cost of higher latency. Alternatively, asynchronous replication prioritizes local latency, with data propagated to other centers later, which can permit temporary divergence. Topologies vary from active-active, which routes traffic to multiple centers for load balancing and fast local reads, to active-passive or hybrid setups that designate one center as primary for writes.

Architectures often rely on conflict-resolution mechanisms such as CRDTs or distributed consensus protocols to manage concurrent updates across centers. Systems may employ a global namespace, regional partitions, and routing policies to direct requests to the best-performing center. Operational considerations include network bandwidth, failure detection and automation of failover, schema evolution across centers, security, and regulatory constraints related to data residency.

Common use cases for crossDC include global web applications seeking low latency for users around the world, multi-region disaster recovery planning, compliance with data residency requirements, and real-time analytics that span multiple regions. Challenges include managing latency trade-offs, cost, complexity, and ensuring data consistency across locations.