New Bigtable in-memory tier for sub-millisecond read latency

Google Cloud Bigtable has introduced an in-memory tier for sub-millisecond read latency, offering 10x higher point read throughput per dollar and hotspot resistance. This eliminates the need for separate caching layers and overprovisioning resources. The in-memory tier automatically moves hot rows into memory to handle traffic spikes, reducing TCO and improving performance. Engineers should consider migrating to Bigtable's in-memory tier for improved workload performance and operational efficiency. For more information, see Bigtable performance documentation.