Topic: Describe switching concepts

Figure 9.1 and the list that follows provide the context for the discussion of several fundamental Ethernet switching concepts that you must review for the CCNA 200-301 exam:

• ▸ MAC learning and MAC aging: One of the responsibilities of a switch is to learn MAC addresses. The switch transparently observes incoming frames. It records the source MAC addresses of these frames in its MAC address table. It also associates the specific port with a source MAC address. The source MAC address is unique for each connected device. Based on this information, the switch can make intelligent frame forwarding (switching) decisions. Notice that a host system could be turned off or moved at any point. As a result, a switch must also age MAC addresses and remove them from the table after they have not been seen for some duration.

  In Figure 9.1, it will not take long before all four systems send some traffic. Keep in mind that many protocols broadcast information periodically. When these frames enter the switch, it records the source MAC address and port information in its MAC address table.

• ▸ Frame switching: Along with building a MAC address table (mapping MAC addresses to ports), a switch also forwards (switches) frames intelligently from port to port. This is sort of the opposite of how a Layer 1 hub works. The hub takes in a frame and always forwards that frame out all other ports. In a hub-based network, every port is part of the same collision domain. A switch is too smart for that. If its MAC address table is fully populated for all ports, it filters the frame from being forwarded out ports unnecessarily. It forwards the frame to the correct port, based on the destination MAC address. Using Figure 9.1 as an example, if D1 sends a unicast frame destined for D4, the switch examines the MAC address table, finds the destination MAC address in that table, and forwards the frame out only the port connecting to D4.

• ▸ Frame flooding: What happens when a frame has a destination address that is not in the MAC address table? The frame is flooded out all ports other than the port on which the frame was received. This is called unknown unicast flooding, and it also happens when the destination MAC address in the frame is the broadcast address (ffff.ffff.ffff).

• ▸ MAC address table: Obviously, the MAC address table is a critical component in a switch today. It really is the brains of the operation. It contains the MAC address–to–port mappings that allow the switch to work its network magic. Example 9.1 shows how easy it is to examine the MAC address table of a Cisco switch.

Example 9.1 Examining a Real MAC Address Table

Click here to view code image

Switch# show mac address-table
          Mac Address Table
-------------------------------------------
Vlan   Mac Address     Type      Ports
----   -----------     --------  -----
 1     e213.5864.ab8f  DYNAMIC   Gi0/0
 1     fa16.3ee3.7d71  DYNAMIC   Gi1/0

Remember that a switch learns based on source MAC information but then switches frames from port to port, based on the destination MAC information. Switches typically use one of three switching methods to forward frames:

• ▸ Store-and-forward: This means the LAN switch copies each complete frame into the switch memory buffers and computes the cyclic redundancy check (CRC) checksum over the frame’s data and compares the calculated value with the value carried within the FCS field for detecting any errors before forwarding the frame. Store-and-forward adds some delay for each frame, but it guarantees that any frames forwarded out of the switch are error free.

• ▸ Cut-through: As soon as the LAN switch copies into its memory just the destination MAC address, which is located in the first 6 bytes of the frame following the preamble, the switch looks up the destination MAC address in its switching table, determines the outgoing interface port, and starts forwarding the frame on to its destination through the designated switch port. The key point here is that the switch starts forwarding the first bits of the frame before having received the entire frame on the inbound port. In reality, cut-through tries to minimize the number of received bits before starting to copy the frame onto the outgoing port.

• ▸ Fragment-free: This works like cut-through switching with the exception that a switch in fragment-free mode stores the first 64 bytes of the frame before forwarding. Fragment-free switching can be viewed as a compromise between store-and-forward switching and cut-through switching. Note that many errors (such as collisions and runts) can be detected in the first 64 bytes of the frame. Fragment-free works best when forwarding frames received on an Ethernet segment where collisions might occur. Unlike cut-through, it only adds delay in the network shown in Figure 9.1 but does not protect against transmission errors.

Finally, please remember that Port LEDs on a switch indicate the health status of ports (when in port status mode). Blinking green indicates that the port is sending and receiving data. Alternating green and amber indicates errors on the link. Amber indicates that a port is blocked by Spanning Tree Protocol.

Topic: Interpret Ethernet frame format

Figure 9.2 shows the most common Ethernet frame format.

Here is information you should know regarding this format:

• ▸ The fields before the Data and Pad field are collectively termed the header.

• ▸ The field after the Data and Pad field is known as the trailer.

• ▸ The Preamble field is 7 bytes in length. It is simply a pattern of alternating 1 and 0 bits that allows devices on the network to easily synchronize their receiver clocks.

• ▸ The Start Frame Delimiter (SFD) field is 1 byte in length. It also has a simple job: It marks the end of the Preamble field and indicates the beginning of the Destination MAC field.

• ▸ The Destination MAC field is 6 bytes in length and is used to store the frame’s destination MAC address.

• ▸ The Source MAC address field is also 6 bytes in length. It stores the appropriate source MAC address.

• ▸ The Length/Type field is 2 bytes in length and identifies the protocol carried in the frame. For example, this field might indicate an IPv4 or IPv6 payload in a network. You might often see this field described as Ether Type instead of just Type (as I prefer).

• ▸ The Data and Pad field ranges from 46 to 1500 bytes. The padding might exist so that the field can meet the minimum length requirement of 46 bytes. Of course, the data portion represents the actual data being sent from a higher layer of the OSI model. Some Cisco switches have the capability to support larger-than-default frames. These frame sizes include Baby Giants (from 1502 up to 1600 bytes) and Jumbo frames (from 1601 up to 9216 bytes), depending on the switch model.

• ▸ The Frame Check Sequence (FCS) field is 4 bytes in length and holds the result of the cyclic redundancy check calculation on the frame’s fields (except with the Preamble, SFD, and FCS fields). The purpose of this field is to determine whether the frame experienced transmission errors in its journey through the network.

Review Questions

1. When a frame enters a Cisco switch, what field does the switch learn from?

A. Preamble

B. FCS

C. Source MAC

D. Destination MAC

2. What is the name of the database that stores address information in a Cisco switch?

A. MAC address table

B. Routing table

C. Interface table

D. Buffer table

3. What command allows you to view the addresses learned by a Cisco switch?

A. show mac-address-table

B. show mac address-table

C. show addresses

D. show mac addresses

4. Why might padding be used in an Ethernet frame?

A. To bring the entire length of the frame to 1500 bytes

B. To bring the length of the Data field to 46 bytes

C. To bring the length of the FCS field to 64 bytes

D. To bring the length of the SFD field to 1 byte

5. If a switch has five workstations attached, how many collision domains are created?

A. 1

B. 0

C. 5

D. 6

Answers to Review Questions

1. C is correct. The switch examines and learns the source MAC addresses of incoming frames.

2. A is correct. The database is termed the MAC address table.

3. B is correct. The command is show mac address-table.

4. B is correct. Padding might be used in the Data and Pad field in order to bring it to the required minimum length of 46 bytes.

5. C is correct. Each of the five workstations connected to a switch is in its own collision domain.

Additional Resources

Layer 2 Addressing & Forwarding

https://www.youtube.com/watch?v=JogWpd2w1Mg&t=1053s

The Ethernet Frame Format - In Less than 60 Seconds

https://www.youtube.com/watch?v=clHBnmUckzg

Ethernet: The Definitive Guide, 2nd edition

by Joann Zimmerman, Charles E. Spurgeon, O’Reilly Media, Inc., 2014

Interconnections: Bridges, Routers, Switches, and Internetworking Protocols, 2nd edition

by Radia Perlman, Addison-Wesley Professional, 1999