Google Cloud Unveils M4N Instances for High-Memory Workloads
Google Cloud has introduced a new range of machine instances called M4N, designed for high-memory and I/O-intensive workloads. These machines are optimized for applications that require large memory pools and fast storage access without a matching increase in compute power.
The M4N series uses 5th Gen Intel Xeon Scalable processors and Google Cloud's Titanium offload architecture to provide memory-to-core ratios of up to 26GB per vCPU and memory sizes of up to nearly 6TB. This makes them ideal for workloads such as mission-critical enterprise databases, generative AI, and electronic design automation.
Google Cloud is pitching the M4N instances to customers that currently add extra virtual CPUs to gain more RAM capacity and storage bandwidth, which can lead to higher licensing bills due to core-based charging. The company estimates that using M4N instances could reduce total cost of ownership for Oracle database deployments by more than 20% compared with similar products from rival hyperscale cloud providers.