Class A is correct because it supports the largest number of host addresses under the original classful IPv4 addressing scheme.

Class A uses a default mask of /8 which leaves 24 bits for host addresses. That yields 2^24 minus 2 usable addresses per network which equals 16,777,214 hosts after removing the network and broadcast addresses.

Class B is incorrect because it uses a default /16 mask which leaves 16 bits for hosts. That gives 2^16 minus 2 usable addresses which equals 65,534 hosts per network and that is far fewer than a Class A network.

Class C is incorrect because it uses a default /24 mask which leaves only 8 bits for hosts. That yields 2^8 minus 2 usable addresses which equals 254 hosts per network and that is much smaller than Class A or Class B.

Class E is incorrect because those addresses are reserved for experimental or future use and are not assigned for general host addressing. Class E is not available for normal host networks which makes it irrelevant when choosing the class with the greatest number of hosts.

Full AWS Practitioner Certification Question