L2 switches forward frames based on MAC address.

Packets are Layer 3 datagrams, and while a switch forwards the packets encapsulated in layer 2 headers, it does not use anything at layer 3 to make the forwarding decision. You could say that packets are forwarded, and forwarding is based on the MAC address, but that is not the best description, and since packets could be fragmented, it is not correct.

RFC 1918 defines the following as private address ranges:

     10.0.0.0        -   10.255.255.255  (10/8 prefix)
     172.16.0.0      -   172.31.255.255  (172.16/12 prefix)
     192.168.0.0     -   192.168.255.255 (192.168/16 prefix)

10.12.256.6 is an invalid address, 256 is not a valid octet value.

EUI-64 addresses are created by inserting the Hex value FFFE between the first 24 bits and the last 24 bits in the interface MAC address. Since the burned in MAC address is guaranteed globally unique, the IP address is guaranteed to be unique on the network.

Technically a bridge and a switch provide the same function. A switch is just a bridge with more ports, but a switch is designed to connect end stations, where a bridge normally refers to a device that separates hubs.

Both a router and a Layer 3 switch device broadcast domains, but a layer 3 switch is designed to provide endpoint connection with collision domain separation, as well.

The correct answer is to replace your hubs with switches. Since switches are a shared bus technology, only one device can be transmitting on the entire Ethernet network, and devices cannot operate in full duplex. Going the other way definitely will not help you. Replacing hubs with routers is not likely to even allow your network to function. Placing bridges between your hubs will reduce the size of the collision domains, and should reduce collisions, but the devices connected to the hubs will still need to operate in half-duplex.

RFC 1918 provides 3 address ranges that can be assigned to endpoints in private networks.

“username ccna privilege 15 password cisco” is the correct command structure. 16 is not a valid privilege level, and if you enter it as “username ccna password cisco privilege 15,” you end up with a standard user with a password of “cisco privilege 15”

Access lists 1-99 are standard access lists, which only match on source IP address.

Cisco encodes a type in any command that stores a password. Type 0 is unencrypted, Type 7 is a reversible encryption, suitable for preventing a password from being seen “over the shoulder,” but easy to decrypt. Type 5 passwords use a non-reversible encryption, based on combining the password with a pseudo random value, and then computing an MD5 hash. Type 4 was introduced in some newer IOS trains, and uses a similar process to the type 5, but using a longer SHA-1 hash for increased security.

Collisions can occur on any shared media Ethernet network, i.e. hubs or coax. Late carrier detect should prevent late collisions. If there is a duplex mismatch, with one side of a cable running at full duplex, and the other side running half duplex, the full duplex side may attempt to transmit while the half duplex side is transmitting. The half duplex side will detect this as a collision.

A frame smaller than 64 octets is padded to that length. Any frame smaller than 64 octets is considered a runt, and should be discarded. While it is true that a frame under 32 octets is a runt, and a runt would not have a properly calculated FCS, the most accurate description is a frame under 64 octets.