Records are identified by 32 byte SHA-2 value. Using BINARY is more compact (8 bits per byte vs 4 bits per byte) but that is offset by the increased difficulty in using binary value in adhoc queries (troubleshooting, etc). I am leaning towards storing as hex strings in 64 byte CHAR field. Are there any other issues or concerns beyond the Records are identified by 32 byte SHA-2 value.