Switching is the base upon which all LANs are built. Understanding the concepts is required for the design and operation of networks.
- [Narrator] Switching operates at layer two, the data link layer. Here, the protocol data unit, which is a single unit of information transmitted, is called a frame. When frames are sent, the sending device puts its address as the sender and adds the destination address to the destination field inside of the frame's header. The header is a section of the PDU that contains information about how the frame should be transmitted. Ethernet addresses are known as Media Access Control addresses. MACs are six-byte long values represented as a 12-digit hex number. MACs are sometimes also called a burned-in address, as they are unique and assigned from the manufacturer. In addition to unique MAC addresses, there are also a few group addresses which are designed to send traffic to multiple hosts at once. Frames destined to the broadcast address, which is all Fs, will be delivered to all devices on the LAN. In essence, it should be delivered to all hosts on this subnet. Frames destined to the multicast address will be sent to a subset of devices that volunteer to receive the traffic. Ethernet contains a Frame Check Sequence field in the frame trailer. A complex formula is applied to the information being sent and the value is stored in the FCS field. If a switch checks the FCS of a frame and it's found not to match, then the frame is simply discarded with no further action. This is to ensure the packet isn't corrupted in transit. When a switch first powers up, it has no knowledge about what clients are connected where in the network. The switch will only learn about the network as hosts send frames. The switch learning process is as follows. First, host A in this diagram sends a frame destined for host B. The switch then references a special table known as the MAC address table, often referred to as address table or simply MAC table. This table maintains a list of all known MAC addresses and what ports they are available through. Since the MAC table is currently empty, the switch will employ a technique known as flooding. Flooding a frame means the switch will send it out every port save for the one it entered from. In this example, host B receives the frame and then responds back to the original sender. When host B's frame enters the switch the address of host B is added to the MAC address table for future forwarding. Since host A's address currently resides in a table, the frame is then forwarded directly out of the port that connects to host A. The same process repeats if multiple switches are connected together. Note that if a frame is flooded in a multi-switch environment and the destination is unknown on all of the switches, then the frame is flooded across all switches. Each entry in the MAC table consumes memory on the switch, thus each switch has a maximum number of entries it can store in its table. This maximum varies from switch to switch. To combat the table from filling up, switches have the concept of MAC address aging. The aging time, which usually defaults to around 300 seconds, is how long an entry will remain in the table if no further frame sourced from this host enter the switch. If a host ages out, then the standard flooding and learning process repeats for it. Understanding the MAC learning process can save you as you chase down hosts during troubleshooting.
Note: A complete overview of the exam and registration instructions can be found at https://www.cisco.com/c/en/us/training-events/training-certifications/exams/current-list/ccna-200-301.html.
- Network models
- Network components
- Troubleshooting the interface and cable issues
- Configuring IPv4 and IPv6 addressing
- Switching concepts
- Switching configuration
- Wireless architectures
- Wireless security