Archives for July 2013

CSMA/CD, or Carrier Sense Multiple Access with Collision Detection, is an access method used in Ethernet.

When a device is about to send data over the network, it will sense if another device is already transmitting. If there is indeed another device transmitting, the device checking will back off from communicating for a random amount of time before retrying.

If two devices check the network at the same time and don’t see any other devices transmitting data, they will both transmit at the same time. When this occurs there is a collision. When a collision is detected, both devices use a random time to back off before retrying.

The back off time is random since if both devices talk on the network at the same time and back off at the same time they will have another collision.

Two tables are used in switch operations, CAM and TCAM. We visited the CAM table in an earlier discussion. In this post we briefly talk about Ternary Content Addressable Memory, or TCAM. Why is it called Ternary? Because it provides three results: 0, 1, and “don’t care”.

The TCAM allows a packet to be evaluated against an entire ACL in a single table lookup. One swoop. This makes lookups ultra-fast. It is done in hardware.

There are two components of TCAM operation:

• Feature Manager (FM)
  • Which compiles the access controle entries into the TCAM table.
• Switch Database Manager (SDM) **Not to be confused with Security Device Manager
  • Which configures or tunes TCAM partitions, if needed.

Greg Ferro has some good information on TCAM and it’s impact with IPv6.

CEF, or Cisco Express Forwarding, is a layer 3 switching technology. It’s an improvement over fast switching and route processor switching.

The FIB, or Forwarding Information Base, is a direct copy of the routing table. If a change occurs on the routing table, it is also reflected in the FIB.

CEF and FIB are a team. Together they are fast. CEF is known as a topology-based switching operation, hence the FIB. Because the information CEF uses is stored in cache, it is able to switch subsequent traffic flows much faster.

References: Understanding Cisco Express Forwarding (CEF)

A unicast is a message being sent to a single destination. So what is unicast flooding then?

When a switch receives a frame destined to a host in which it does not have in it’s CAM table then the switch will flood the unicast frame to all ports with the same VLAN ID.

Basically, the unicast becomes a broadcast. The term for this is unknown unicast flooding.

In the example below, let’s assume Host A sends a frame to the switch destined to Host B. Because host B is not in the CAM table, the switch will flood the unicast frame to all ports with the same VLAN ID. This is what the CAM table looks like for the switch as Host A sends it’s frame:

[table id=3 /]

We dig in deeper into the operations of a switch in the CCNP SWITCH Official Certification Guide. The CAM table is one of the fundamental operations of a switch. It’s not only important for the 642-813 SWITCH exam but it’s important to know for working on the job.

The CAM table, or content addressable memory table, is present in all Cisco Catalysts for layer 2 switching. It is used to record a stations mac address and it’s corresponding switch port location. In addition, a timestamp for the entry is recorded and it’s VLAN assignment.

The CAM table is used in multilayer switching for the purpose of quickly switching frames to their destination. The switch looks at the incoming frame’s source MAC address and enters it into the CAM table and keeps it there for 300 seconds before aging out. This is the default value.

If the device connected to that switchport is moved to another port, the switch records the incoming source MAC address, updates the CAM table and removes it’s previous entry for the same MAC address.

Host A is connected to switch port 1 and Host B is connected to switch port 2.

1. Host A sends traffic to the switch.
2. The switch looks into the frame and records the source MAC address (of Host A) and places an entry into the CAM table. Host A is on switchport 1, has the MAC address of AAAA, VLAND ID of 1, and the timestamp.
3. Host B has not communicated with the switch yet.
4. Host A decides to communicate with Host B.
5. When Host A sends a frame to the switch destined to Host B, the switch notices the destination MAC address (for Host B) in the frame, queries the CAM table for that MAC address but doesn’t find it.
6. Because the destination MAC is unknown, the switch marks the frame for flooding and sends the unicast frame to all ports with the same VLAN association.
7. Host B responds to the unicast frame.
8. The switch records the incoming frame from Host B and records Host B’s MAC, switchport location, VLAN ID, and applies a timestamp.
9. The next time Host A sends a frame destined for Host B, the switch queries it’s CAM table, finds Host B in the table and sends the frame directly to Host B.

CAM Table Before Host B Communicates on the Network

[table id=1 /]

CAM Table After Host B Communicates on the Network

[table id=2 /]