.png)
How It Works
AWS organizes its EC2 compute instances into families, where each family is designed for a specific workload type. The general-purpose family (M-series) balances compute, memory, and networking. The compute-optimized family (C-series) prioritizes raw CPU performance. The memory-optimized family (R-series) suits workloads that need large amounts of RAM. Within each family, instance sizes scale predictably, from small to many-times-extra-large, sharing the same architecture but varying in vCPU count, memory, and network throughput. The family label always appears as the first letter or letters in the instance type name. For example, in “m5.xlarge,” the “m” identifies the general-purpose family and the “5” identifies the generation.
Why It Matters for Cloud Cost
Instance family selection directly determines which commitment-based discounts apply and how flexible those discounts are. AWS Reserved Instances (RIs) are tied to a specific instance family within a region. If your team purchases an RI for the M family and later migrates workloads to a C-family instance, that RI may not apply, leaving you paying on-demand rates while also carrying an unused commitment. AWS Compute Savings Plans offer more flexibility by applying discounts across families, but EC2 Instance Savings Plans are still scoped to a specific family. Getting the family selection wrong at purchasing time is one of the most common sources of commitment waste in growing engineering organizations.
Usage AI: Usage AI’s Autopilot mode monitors your active instance families daily and adjusts commitment coverage automatically, so a shift in workload type does not leave your organization holding underutilized reservations.