An administrator is selecting hardware for a new server that must be able to recover from minor memory issues. Which of the following is the BEST selection to ensure continued memory reliability?

ECC memory uses additional data bits to detect single-bit errors and automatically correct them before they reach the CPU or cause a crash, and can detect double-bit errors. This hardware-level correction is transparent to the operating system and provides continuous protection against minor memory faults. It is the industry-standard feature specifically designed to ensure continued memory reliability in servers.

DDR3 (Double Data Rate 3) is a memory generation and signaling standard that defines speed and interface specifications, not error detection or correction capability.

Non-uniform memory access (NUMA) is a multi-processor memory architecture concerned with access latency across CPU sockets, not with detecting or correcting memory errors.

AMD virtualization is a processor feature enabling hardware-assisted virtual machine execution and has no relationship to memory error correction or reliability.