Ethernet is an important topic in the Cisco CCNA because network administrators typically oversee LANs (local area networks), and pretty much all LANs today use some form of Ethernet, whether it be copper Fast Ethernet, or fiber optic Gigabit Ethernet, or wireless Ethernet. Ethernet became what it is today, because it was cheap and easy to install. It continued to improve its standards and hardware (eg. hubs to switches), also it has remained backwards compatible with the ability to change physical implementations from wireless, to fiber, to copper, as well as change speeds and standards all within the same functional network.
Ethernet and Collision Domains
Early versions of Ethernet used coaxial cable (10Base5 Thicknet and 10Base2 Thinnet). The physical topology could be described as a single cable that all users connected to or tapped into, this was known as a physical bus or multi-access network. Logically Ethernet was also a bus, or multi-access network, all hosts on the network could see each other, and all packets as well. All users were essentially on the same cable or same collision domain. What characterizes an Ethernet collision domain is in a collision domain, when two users send packets at the same time, the result is a collision or spike of voltage on the wire and all sending of packets must cease for a short period of time.
If you have ten hosts connected to a hub using regular Ethernet cables (10BaseT, twisted pair) then all hosts comprise a single collision domain. If you connect to many hosts to a hub or extend the network by connecting hubs to more hubs and more hosts then network performance will decrease and collisions will increase. In this way, if you have ten hosts connected to a hub and that hub is connected to another hub with another 10 hosts, then that network also comprises just a single collision domain.
Collisions were exacerbated because of the fact that Ethernet was designed as a multi-access network, where all hosts see all other hosts and all packets as well. The number of hosts in the network, and the presence a broadcast packets coming from multiple hosts, would increase the chances for collisions to occur.
The advent of switches was a significant improvement for Ethernet and local area networks. Switches provide many important improvements to a network, including collision free networking and better bandwidth utilization. Whereas a hub receives a frame on one port and automatically forward it out of all other ports, in contrast a switch maintains a table or map of MAC addresses to switchports, and is able to switch a frame to the destination port where the destination MAC address resides. Only when a switch does not have the MAC address in its table, or if it is a layer 2 broadcast, will a switch forward a frame out of all ports except the one it came in on. Thus less frames are traveling on the network unnecessarily. Since traffic is sent to only one port, each port or link on a switch is considered its own collision domain. Thus, switches break apart or create collision domains as opposed to hubs which extend or grow collision domains. With the advent of full duplex communications, hosts connected to switches could both send and receive frames at the same time without collisions.
Ethernet and ARP
ARP stands for address resolution protocol and its function is to resolve IP addresses to MAC addresses at Layer 2. When a frame or “packet” needs to be delivered to a host on a local area network it needs to delivered to the host’s MAC address. If the sending host does not have the destination host’s MAC address in its ARP cache it will send an ARP broadcast packet requesting the MAC address from the destination host’s IP address. So a MAC address needs to be resolved from an IP address before a packet can be delivered on a local network. In this way, ARP is plays an important role in the functioning of local area networks. In the video below I demonstrate the ARP process using a command prompt and Wireshark.
For more information on ARP: http://en.wikipedia.org/wiki/Address_Resolution_Protocol
For more information on Multicast addresses: http://en.wikipedia.org/wiki/Multicast_address
Hexadecimal Notation, Counting and Conversion
The ability to convert binary to decimal and vice versa is important to the Cisco CCNA, but you must also know how to convert hexadecimal. Hexadecimal is a shorthand notation that is used in computers all the time. MAC addresses are written in hexadecimal notation like this: B3:A2:77:00:F1:C9. Their are hexadecimal color charts for HTML and the web like 0xFF0000 which equals the color red, and hexadecimal is used in programming as well.
In the Cisco CCNA, hexadecimal notation is introduced when learning about layer 2 physical addressing, or MAC addresses. MAC addresses are 48 bits long and are typically written in 6 character pairs separated by a colon or a dash (eg. B3:A2:77:00:F1:C9 or B3-A2-77-00-F1-C9), but they can also be written in pairs of six or groups of four (eg. B3A277:00F1C9 or B3A2:7700:F1C9). You will also find hexadecimal numbers with a “0x” prefix or a “h” suffix to indicate that the number is in hexadecimal notation.
Hexadecimal is a Base16 counting system because there are 16 characters or numbers (0,1, 2, 3, 4, 5, 6, 7, 8, 9, a, b, c, d, e, f) with “a” through “f”equaling the numbers 10 through 15. Since a single hexadecimal digit or character has 16 possible values we can equate one hexadecimal character with 4 bits (24 equals 16). This creates an easy conversion between a binary 8 bit number to a 2 digit hexadecimal number:
10111000 in binary = 184 in decimal
1011 – 1000 (splits the 8 bits into two 4 bit nibbles)
1011 = 11 in decimal and B in hex
1000 = 8 in decimal and 8 in hex
0xB8 = 184 in decimal
Ethernet, Data Link, and Local Area Network Tips
- Ethernet functions on both layer 2 the Data Link layer and layer 1 the Physical layer. In the TCP/IP model layers one and two from the OSI model are combined into the Network Access layer.
- The Data Link layer has an upper and lower sublayer, the LLC and the MAC sub layers
- 802.2 is the the LLC, logical link control sublayer. Its role is to function in software and identify the network layer protocol above it.
- Ethernet at its core is CSMA/CD. Ethernet is collisions and collision detection.
- Hubs cause collisions. Switches cause no collisions because each port is its own collision domain.
- Source and destination MAC addresses change as a frame travels across networks. Source and destination IP addresses do not change.
- You only need to send packets/frames to the gateway/router when you are trying to contact a different network.
- Packets/frames destined for a host on the same network do not have to go through the router but are delivered directly to the destination host’s MAC address
Other Ethernet Topics
Ethernet as a WAN
MAC Address Structure
Ethernet Unicast, Multicast, and Broadcast
10Mbps 100Mbps, 1000Mbps Ethernet