Well, I’m still stuck on Spanning Tree. Sort of. I’ve read and understood most of the concepts. I just don’t feel any motivation to take notes on the last part, MST. If I do, I’m tempted to just write a high overview of the concepts. I started reading the next section of the books. That is, Inter-VLAN routing. Then I plan to move on to CEF concepts. I had planned to get these both covered this week, but unfortunately things have gotten much slower for me in the study front.

I’ve been dedicating about an hour and a half to working out everyday. Well, really its about 45 minutes of working out but theres 10 minutes of warm up here, and another 10 minutes of stretching there, then cooling down, then cleaning up. So that takes a good chunk of my evening already. To top that, my wife also has to take some online classes to keep her teaching credentials up to date. So the time that she has to study I have to help watch the kids. She’s been very generous about allowing me to study for hours at a time in the past, but now, I have return the favor.

Nowadays I have to make up for the lost study time by studying at work during break or slow periods. Even that doesn’t add up to a lot because I have been pretty busy at work.

Today though I had a little shot of motivation when a good buddy of mine, some french dude from france , passed his BCMSN exam today. Or yesterday, depending on what time zone you’re in. Head over to Nicolas Michel’s page and offer your congratulations when you get the chance. He’s a motorcycle racer so that makes him pretty cool

Got this from a friend (appropriately enough, today is my and wife’s 4th year anniversary ):

====================================================

Dear Tech Support,

Last year I upgraded from Boyfriend 5.0 to Husband 1.0 and noticed a distinct slow down in overall system performance, particularly in the flower and jewelry applications, which operated flawlessly under Boyfriend 5.0.

In addition, Husband 1.0 uninstalled many other valuable programs, such as Romance 9.5 and Personal Attention 6.5, and then installed undesirable programs such as NBA 5.0, NFL 3.0 and Golf Clubs 4.1.

Conversation 8.0 no longer runs, and Housecleaning 2.6 simply crashes the system.

Please note that I have tried running Nagging 5.3 to fix these problems, but to no avail.

What can I do?

Signed,
Desperate.

DEAR DESPERATE,

First, keep in mind, Boyfriend 5.0 is an Entertainment Package, while Husband 1.0 is an operating system.

Please enter command: ithoughtyoulovedme. html and try to download Tears 6.2 and do not forget to install the Guilt 3.0 update. If that application works as designed, Husband 1.0 should then automatically run the applicationsJewelry 2.0 and Flowers 3.5.

However, remember, overuse of the above application can cause Husband 1.0 to default to Grumpy Silence 2.5, Happy Hour 7.0 or Beer 6.1. Please note that Beer 6. 1 is a very bad program that will download the Farting and Snoring Loudly Beta.

Whatever you do, DO NOT under any circumstances install Mother-In-Law 1.0 (it runs a virus in the background that will eventually seize control of all your system resources.)

In addition, please do not attempt to reinstall the Boyfriend 5.0-program.These are unsupported applications and will crash Husband 1.0.

In summary, Husband 1.0 is a great program, but it does have limited memory and cannot learn new applications quickly. You might consider buying additional software to improve memory and performance. We recommend Cooking 3.0 and Hot Lingerie 7.7.
Good Luck Babe!

Tech Support

Configure RSTP

• In the traditional 802.1D spanning tree protocol, a switch operates in Per-Vlan Spanning Tree Plus (PVST+).
• In order to use RSTP, the spanning-tree mode must changed.
• The following interface configuration command makes sure that a port operates as an RSTP edge port:

Switch(config-if)#spanning-tree portfast

• This configuration is the same PortFast feature in 802.1D that ensures the port  transitions automatically from blocking to forwarding.

• Recall that a port that is operating in full-duplex mode is automatically considered by the switch point-to-point.
• To manually override the automatic determination, use the following interface config command:

Switch(config-if)#spanning-tree link-type point-to-point

Rapid Per-Vlan Spanning Tree Protocol (RPVST+)

• As we mentioned above, PVST+ is the default STP mode on Catalyst switches.
  • This mode allows one spanning tree instance for each VLAN active on the switch.
• To use the more efficient RSTP mode, configure the switch to begin using the Rapid PVST+ (RPVST+) mode using the following global configuration command:

Switch(config)#spanning-tree mode rapid-pvst

• When this configured on a production network, any STP process is restarted.

To configure the switch back to PVST+ mode, use:

Switch(config)#spanning-tree mode pvst

To verify the STP mode, use the following command:

Switch#show spanning-tree vlan vlan-id

This entry is not an authoritative guide. These are merely notes and rehash of the primary text materials and resources that I use. For a thorough guide of the BCMSN course, consider purchasing Building Cisco Multilayer Switched Networks (BCMSN) (Authorized Self-Study Guide) (4th Edition)by Richard Froom, Balaji Sivasubramanian, and Erum Frahim and CCNP BCMSN Official Exam Certification Guide (4th Edition) by Dave Hucaby ; as well as following the links on the reference section of this entry.

RSTP Convergence

• Convergence of RSTP is a sequence of handshakes (proposal and agreement)  propagated over point-to-point links.
  • When a switch needs to make an STP decision, a handshake is made with its nearest neighbor.
  • When that is successful, the handshake sequence is moved to the next switch and the next and moves on the rest of the spanning tree towards the edge.
• During the handshake sequence the switch takes proper steps to make sure loops are prevented before moving on to the next handshake sequence.

Example

Step 1

• Assume in figure 1 that the topology on the left is a non-redundant topology.
• An administrator decides to add a link between the Root and Switch A for redundancy.
• As soon as the links come up, the ports on the link between A and the root are put in blocking.

Figure 1a: Step 1 

Proposal and Agreement

• A negotiation (handshake) sequence is started between A and the root.
  • Both bridges exchange BPDUs.
• As soon as A receives the BPDU of the root, non-edge designated ports on A are blocked.
• This process is called synchronization.
• The following diagram illustrates the handshake sequence between A and the Root bridge:

Figure 1b: Proposal/Agreement Between Switch A and Root Bridge

1. In the figure above, when the new link is created between the Root and Switch A, both ports are initially put on blocking state. 
   • They both start in designate blocking status and both sends BPDUs with the proposal bit.
   • Port p0 of the Root bridge sends a superior BPDU.
2. Port  p1 of Switch A immediately knows that it is the new root port when it receives the superior information. Switch A starts a sync to verify that all of its ports are in-sync* with this new information.
   • Switch A proceeds to block the other non-edge designate ports (p2 and p3).
3. Switch A sends a BPDU back to the Root bridge - this time with the agreement bit set.
   • Switch A then tells the root bridge that it can go ahead and put its port in forwarding status.

*A port is in sync if it meets either of these criteria:

• The port is in blocking state, which means discarding in a stable topology.
• The port is an edge port. 

Step 2

• After the links between the Root and Switch A has converged, a new negotiation sequence is started on the links between Switch A’s neighbors – B and C.
• As you can see on the diagram on Figure 2, instead of blocking on the segment above Switch A, the network now blocks below Switch A.  

Figure 2: RSTP Convergence Step 2

• At this stage, the same proposal/agreement sequence discussed above happens between Switch A and Switch B.
  • Because there is no other non-edge designated ports (assume that the nodes on bottom of Switch A are edge devices) on Switch B, it has no ports to block in order to authorize Switch A to go to the forwarding state.
  • It immediately transitions to forwarding state.
• Switch A and Switch C initiate the same sync operation.
  • This time, when Switch C receives the proposal bit from Switch A along with the superior BPDU.
  • Switch C blocks its non-edge designated port (link between C and D)
  • Switch C then sends agreement BPDU to A and tells it to start forwarding.

Step 3

• The end of convergence process ends here, where the port on D stays blocking.
• The time it takes for the whole operation to complete happens within the speed of one BPDU transmission.

Figure 3: Final Topology

In Summary…

• RSTP convergence starts of with the switches sending proposals BPDUs.
• The recipient of the proposal makes sure it is synchronized by putting all its non-edge designated ports in blocking, effectively isolating itself from the rest of the topology.
• These blocked ports also send proposal messages to their nearest neighbors and those neighbors perform the synchronization operation themselves.
• This action creates a sort of moving “wave” of switches synchronizing all the way down the tree.
• Each “wave” independently makes decisions whether to forward or not, depending on the agreement message received from the neighboring switch. This way, the transition happens rapidly as opposed to waiting for the BPDU message to propagate to the whole topology before making a decision which port will ultimately by blocking and forwarding.

Topology Change

• With 802.1D, when a topology change is detected:
  • The switch that detects the change sends a Topology Change Notification to the root.
  • The Root sends a Configuration BPDU with the TC flag bit set towards the rest of the topology.
  • The switches down the tree ages out their CAM table (MAC address table) in 15 seconds (forward-delay time) to quickly delete old information.
• In 802.1w, topology change is only detected when a non-edge port transitions to the Forwarding State.
  • A port moving to blocking does not generate a TC BPDU.
• If a topology change is detected:
  1. A switch sends out BPDUs with their TC bit set out to all non-edge designated ports.
     • This happens for the duration of the TC While timer, which is 2 times the Hello interval.
  2. MAC addresses associated with the non-edge designated ports are flushed from the CAM table, thereby forcing the bridges to re-learn and re-populate its CAM database.
     • The port that received the TC message does not have to clear its learned MAC address.
  3. This process happens for every switch that receives the TC message. Each receiving switch starts the TC While timer and must also send out TC messages out their non-edge designated ports.
  4. While the TC While timer is active, the switches send BPDUs even on the root port. 
• This process floods the TCN to the entire much more quickly because the initiator immediately sends the TC information to the network instead of waiting for the Root bridge to take care of the notification. 

802.1D Compatibility

• In the event a designated port sends a proposal message and does not receive an agreement reply, it reverts to using the 802.1D rules of convergence – it mus wait the forward delay time before forwarding.
• 802.1D cannot understand 802.1w.
• If an RSTP switch receives an STP BPDU, the RSTP switch changes its mode to 802.1D on the port where it receives the STP BPDU. If the STP switch is removed from the topology, an administrator must re-configure the port manually back to RSTP – that is because it doesn’t know whether the STP switch is still present or not.

This entry is not an authoritative guide. These are merely notes and rehash of the primary text materials and resources that I use. For a thorough guide of the BCMSN course, consider purchasing Building Cisco Multilayer Switched Networks (BCMSN) (Authorized Self-Study Guide) (4th Edition)by Richard Froom, Balaji Sivasubramanian, and Erum Frahim and CCNP BCMSN Official Exam Certification Guide (4th Edition) by Dave Hucaby ; as well as following the links on the reference section of this entry.

Rapid Spanning Tree Protocol (802.1w)

• Rapid Spanning Tree Protocol or RSTP or IEEE 802.1w was introduced by the Institue of Electrical and Electornics Engineers in 1998 .
  • The original STP 802.1D was introduced in 1985.
  • 802.1D – 2004 incorporates RSTP and obsoletes the original STP.
• RSTP selects one switch as the root port of an active spanning tree-connected topology and assigns port roles to individual ports on the switch, depending on whether the ports are part of the active topology.
• Adds port roles: Alternate and Backup roles.
• New port state: Discarding in addition to Learning and Forwarding

* UplinkFast, BackboneFast, and Portfast are Cisco proprietary enhancements to 802.1D

• Terminologies and parameters found in 802.1D remains primarily the same for RSTP.
• 802.1w is capable of reverting to 802.1D for interoperation with legacy bridges on a per-port basis.
  • Downgrading to 802.1D cancels the benefits of 802.1w for that paritcular segment.
• Per VLAN version of RSTP is called RPVST+ equivalent to 802.1D version PVST+
• With RSTP, if a swtich, switch port, or LAN experiences some kind of failure, rapid connectivity is possible.
  • A new root port and designated port of the connecting bridge transitions to forwarding through an explicit handshake protocol between them.
  • RSTP allows switch port configuration so that the ports transition to forwarding directly when the switch re-initializes.

RSTP Port States

• There are only 3 RSTP port states:
  • Discarding - This is a combination of  801.2D blocking, listening, and disabled states
  • Learning
  • Forwarding
• Port states are defined according to what the port does with incoming frames  - if incoming frames are dropped or ignored, outgoing frames are as well.
• RSTP decouples the role of a port from the state of a port.
  • 802.1D STP mixes the state of a port, whether blocking or forwarding traffic, with the role it plays in the active topology.
• RSTP considers there to be no difference between a port in blocking and listening states (ie both discard frames and no MAC addresses are learned).

RSTP Port Roles

• Root Port
  • The closest port (measured in “path cost”) to the root bridge.
  • The STA elects a single root bridge for the whole bridged network, per-VLAN, or STP instance.
  • The root bridge sends BPDUs that are better than the ones that any other bridge sends.
  • The root bridge is the only bridge that does not have a root port.
• Designated Port
  • For bridges in the same segment, the designated port is the port on the bridge, in that LAN segment, that is sending the best BPDU.
• Alternate Port
  • Becomes the root port if the active root port fails.
  • Blocked from receiving root BPDUs from another switch. An alternate port has to receive BPDUs from a different bridge than itself. See figure below.
• Backup Port
  • Becomes the designated port if the active designated port fails.
  • Blocked from receiving root BPDUs from the designated port for a shared LAN segment from the same bridge on which the port is located. See figure below.
• Disabled Port
  • Has no role within the operation of spanning tree.

Figure 1: Alternate Port vs Backup Port

 BDPU Format: 802.1D vs. 802.1w

• Recall that 802.1D BPDU (1byte) only used 2 bits in the Type field (either the most significant bit is set or the least significant bit)
• RSTP uses all 8 bits.

Figure 2: 802.1D vs 802.1w Frame Comparison

• With 802.1D, a non-root bridge generates a BPDU only when it receives one on its root port.
  • Here, there is more frequent relaying of BPDUs compared to the self-generated ones.
• With 802.1w, even if a bridge does not recieve a BPDU from a root it still send a BPDU every 2 seconds – this is self-generated.
  • This constant transmission of BPDUs act as keep-alive mechanism.
  • If three BDPUs in a row (three consecutive hello times) are not received, the bridge will assume that connectivity is lost.
  • This allows the aging of protocol information (max age) to occur much faster thereby detecting failure much faster.
• A mechanism similar to backbone fast allows RSTP to accept inferior BPDUs.

Rapid Transition to Forwarding State

• The major selling point of the 802.1w is its rapid transition.
• With 802.1D, even after the designated port has been established, it still waits the forwarding timers before the port transitions to forwarding. That’s 15 sec listening and 15 sec learning.
• RSTP actively confirms that a port can safely tranisition to the forwarding state without having to rely on any timer configuration.
• The type of port is used by RSTP to base its forwarding decision when a switch detects a failure:
  • Edge Ports
    • These are ports directly connected to end stations and are typically unable to form bridging loops.
    • Directly transitions to forwarding state bypassing the listening and learning stages.
    • They are equivalent to PortFast feature.
    • However, unlike PortFast, an edge port that receives a BPDU immediately loses edge port status and becomes a normal spanning tree port. 
  • Link Type (point-to-point)
    • RSTP can only achieve rapid transition to the forwarding state on edge ports and point-to-point links.
    • Derived from the duplex mode of a port:
      • A port in full-duplex is point-to-point.
      • A port in half-duplex is assumed to be on shared medium, such as a hub.
    • Switched networks today operate mostly in full-duplex and thus are treated as point-to-point links by RSTP. This makes then candidates for rapid transition to the forwarding state.

References:

1. Understanding Rapid Spanning Tree Protocol – Cisco Systems, Inc.

This entry is not an authoritative guide. These are merely notes and rehash of the primary text materials and resources that I use. For a thorough guide of the BCMSN course, consider purchasing Building Cisco Multilayer Switched Networks (BCMSN) (Authorized Self-Study Guide) (4th Edition)by Richard Froom, Balaji Sivasubramanian, and Erum Frahim and CCNP BCMSN Official Exam Certification Guide (4th Edition) by Dave Hucaby ; as well as following the links on the reference section of this entry.

The past week has been pretty steady for me as far as studies go. My elder son got sick with 103 ~ 105 fever for pretty much all of last week. Becuase of that, I was forced to stay home instead of driving to LA to see my mom. It gave me the chance to study a little bit more than usual. I also got some quality lab time with my Catalyst 3550 switches; working on some spanning tree and etherchannel stuff. This week I built ten more cross-over Ethernet cables to add to the 5 I already have at home. So that should allow me to practice more bundling multiple EtherChannel links.

It also helped that I didn’t exercise/work out as consistently as I should’ve been. My son needed constant attention so I couldn’t really devote a un-interrupted work out time. So this week I’m restarting week 1 of my P90X program. So I didn’t get to work out much but I did watch as much basketball as there was on. For the unfamiliar, it was the championship series between Los Angeles Lakers and Orlando Magic. Of course everyone knew it was the Lakers’ championship to take. The games were merely formalities.   So congratulations to the Lakers and the city of L.A. for a much awaited championship banner. Seven years without a championship is just too long. At least for L.A.

Anyway, this week I plan to continue my studies and labs. I want to finish up on Spanning-tree (I still have RSTP and MST to go over) then move on to multilayer switching next week.

Layer 2 EtherChannels

• When configuring Layer 2 channels, use the channel-group interface configuration command on the interface port or ports.
  • This command creates the port-channel logical interface.
• You cannot put a Layer 2 interface into a manually created port-channel interface.

Configure PAgP EtherChannel

When configuring EtherChannel use the following general template:

Switch(config)# interface type mod/num
Switch(config-if)# channel-protocol pagp
Switch(config-if)# channel-group number mode {on | [auto | desirable] [non-silent]}

• Lab testing showed that if ommited, verification defaults to PAgP negotiation protocol.
• Older switch models, such as Catalyst 2950, offer only PAgP option therefore the channel-protocol command is not even available.
• Channel group number can be from 1 to 64.
• Each interface in the EtherChannel bundle must be assigned to the same channel group number.
• PAgP defaults to silent mode with the Auto and Desirable mode, by default. It means it will not transmit PAgP packets.

Figure 1: This is the example topology used for this practice lab configuration

SW4#conf t
Enter configuration commands, one per line.  End with CNTL/Z.
SW4(config)#int fa0/1
SW4(config-if)#switchport
SW4(config-if)#channel-group 1 mode desirable
Creating a port-channel interface Port-channel 1

 

!

SW4(config-if)#int fa0/2
SW4(config-if)#switchport
SW4(config-if)#channel-group 1 mode desirable

!

SW4(config-if)#int fa0/3
SW4(config-if)#switchport
SW4(config-if)#channel-group 1 mode desirable

!

SW4(config-if)#int fa0/4
SW4(config-if)#switchport
SW4(config-if)#channel-group 1 mode desirable

The configuration of SW5 follows the same basic steps as with SW4. Only this time we’re using int range to configure a range of interfaces for EtherChannel with one swing.

SW5#conf t
Enter configuration commands, one per line.  End with CNTL/Z.
SW5(config)#int range fa0/1 – 4
SW5(config-if-range)#switchport
SW5(config-if-range)#channel-group 1 mode desirable
Creating a port-channel interface Port-channel 1
SW5(config-if-range)#

Verification

• SUmeans port channel 1 is a Layer 2 EtherChannel and is in use.
• The P flags on the ports indicate that they are active in the channel.
• Although assigned to PO1 channel group, port Fa 0/48 has a D flag because the physical connection on the other end does not exist.

Configure LACP EtherChannel

The following template configures LACP, inlcuding some parameters:

Switch(config)# lacp system-priority priority
Switch(config)# interface type mod/num
Switch(config-if)# channel-protocol lacp
Switch(config-if)# channel-group number mode {on | passive | active}
Switch(config-if)# lacp port-priority priority 

• The system priority defaults to 32, 768.
  • Ranges from 1 to 65, 535.
  • A lower system priority allows a switch to become the main decision maker about the EtherChannel.
  • If left to the default value, the lower MAC address wins.
• The lacp port-priority command allows you to configure an interface to be active port or on standby.
  • You can configure more interfaces than the maximum eight the system allows to be active.
  • By configuring ports with low port-priority, you ensure that they are active on the channel. With a higher priority they are reserved as standby.
  • For example, you may configure 4 ports with a lower priority to make it active in the channel. Then you may configure another 4 with higher priority value than the fist four - which puts them on standby. If one of the lower priority ports goes down, one of the standby ports become active.
  • The default priority is 32, 768. If all ports are left in default value, the lower port numbers (in interface number order) are used to select the active ports.
  • LACP port priority value can range from 1 to 65, 535.

The following is the SW4 configuration example (refer to the same topology diagram on figure 1):

SW4#conf t
Enter configuration commands, one per line.  End with CNTL/Z.
SW4(config)#lacp system-priority 150
SW4(config)#int rang fa0/1 – 4
SW4(config-if-range)#channel-protocol lacp
SW4(config-if-range)#channel-group 1 mode active
Creating a port-channel interface Port-channel 1

If a port priority were to be configured, it would look something like this:

SW4(config-if-range)#lacp port-priority 150

Similarly, configure SW5 as follows:

SW5#conf t
Enter configuration commands, one per line.  End with CNTL/Z.
SW5(config)#int range fa0/1 – 4
SW5(config-if-range)#channel-protocol lacp
SW5(config-if-range)#channel-group 1 mode active
Creating a port-channel interface Port-channel 1

Verify

• All ports Fa0/1 – 4 are active in the channel group
• The negotiation protocol used is LACP. 

Configure Layer 3 EtherChannel

• When configuring Layer 3 EtherChannels, you must manually create the por-channel logical interface first, using the interface port-channel global config command.
• After that , you can put the logical interface intot the channel group by using the channel-group interface configuration command.
• Use the no switchport interface command to put the interface in Layer 3 mode. 
• If you decide to move an IP address from a physical interface to an EtherChannel, delete the address from the physical interface first, then configura it on the port-channel.

SW5 config t
SW5(config)# interface port-channel 2
SW5(config-if)# no switchport
SW5(config-if)# ip address 10.10.10.1 255.255.255.0
SW5(config-if)# no shut

!

SW5(config)#int ra fa0/1 – 4
SW5(config-if)#no switchport
SW5(config-if)#channel-prot lacp
SW5(config-if)#channel-group 2 mode active

References:

1. Configuring EtherChannels – Catalyst 3550 Software Configuration Guide 12.1(13)EA1

This entry is not an authoritative guide. These are merely notes and rehash of the primary text materials and resources that I use. For a thorough guide of the BCMSN course, consider purchasing Building Cisco Multilayer Switched Networks (BCMSN) (Authorized Self-Study Guide) (4th Edition)by Richard Froom, Balaji Sivasubramanian, and Erum Frahim and CCNP BCMSN Official Exam Certification Guide (4th Edition) by Dave Hucaby ; as well as following the links on the reference section of this entry.

A nice little box…

Cisco EtherChannel Technology

• The Cisco EtherChannel technology is Cisco’s method of aggregating or bundling parallel links to scale bandwidth.
  • Port Trunking or Link Aggregation is the standard term for this technology. It is not to be confused with Cisco’s use of the word “trunking” which refers to the network link used to carry multiple VLANs between switches.
• It is standards-based built upon the IEEE 802.3-compliant Ethernet mechanisms for full-duplex autonegotiation and autosensing.
• It allows you to bundle up to eight individual links to form one logical link:
  • 8 Fast Ethernet (full duplex) = 1600 Mbps
  • 8 Gigabit Ethernet (full duplex) = 16 Gbps
  • 8 10-Gigabit Ethernet (full duplex) = 160 Gbps
• Whenever parallel links between switches are present, the possibility of bridging loops increase. However, with EtherChannel, this problem is avoided by bundling these parallel links into a single logical link.
• Although an EtherChannel link is seen as a single logical link, the speed of the link is not always the sum of all the physical links’ speed.
  • For example, if we have 8 FastEthernet physcial links bundled together (1600 Mbps at full-duplex), the logical link doesn’t operate at that speed.
  • Although it is capable of carrying a total throughput of 1600 Mbps, traffic is not equally distributed to all the links.
  • Instead, traffic is distributed through each individual FastEthernet links within the EtherChannel. And each link operates at its inherent speed, that is 200 Mbps at full-duplex.
  • If one link within that bundle is used more than the other, the link will carry more load than the other links.
  • There is a way to balance the load more evenly among all other links (discussed at a later post).
• When bundling ports with EtherChannel:
  • All bundled ports must generally belong to the same VLAN.
  • If used as a trunk, bundled ports must be in trunking mode, have the same native VLAN, and pass the same set of VLANs.
  • Each port should have the same speed and duplex settings.
  • And each bundled port must be configured with identical spanning-tree settings.

EtherChannel Negotiation Protocols

I. Port Aggregation Protocol (PAgP)

• Cisco-proprietary.
• PAgP manages the automatic EtherChannel configuration and negotiation between switches.
  • Checks to make sure configuration is consistent on both ends of the link.
  • Manages link failures and addition of links.
  • Helps in network reliability by preventing STP loops or packet loss due to misconfiguration.
• EtherChannel formed only on ports that are configured for either identical static VLANs or trunking.
• If one of the bundled ports’ parameters are changed (ie VLAN, speed, duplex setting) PAgP also dynamically changes the parameters on the EtherChannel.

PAgP Modes

Options

The following options are available for the Auto and Desirable modes:

• Silent
  • Default option/keyword used for Auto or Desirable mode.
  • Allows ports to be added to an EtherChannel even if the other end of the link is silent and never transmits PAgP packets
  • Indicates that the switch does not expect PAgP frames from the partner device to prevent the switch from reporting the link to the STP as down.
  • This could be used to form an EtherChannel with a device such as a file server or network analyzer that does not participate in PAgP.
• Non-silent
  • Used with Auto or Desirable mode to indicate that the switch expects PAgP frames from the switch on the other end of the link.
  • This requires each port to receive PAgP packets before they can be added to the EtherChannel.
  • If a unidirectional link is detected (due to absence of PAgP packets received) the port status is reported to STP as down.

Mode Compatibilities

• On mode is compatible only with another port in the On mode.
• Auto mode is compatible only with Desirable mode.
• Desirable mode is compatible with Auto or Desirable.

II. Link Aggregation Control Protocol

• This is an industry standard based on IEEE 802.3ad (also known as IEEE 802.3 Clause 43, “Link Aggregation”)
• Just like PAgP, switches exchange LACP packets to learn parameters.
• On ports configured to use LACP, the protocol tries to configure the maximum number of compatibel ports in an EtherChannel, up to the maximum allowed by the hardware (eigth ports).
• There are different parameters that are used in LACP:
  • LACP System Priority
    • The lowest system priority is the one that makes decisions about what ports are actively participating in the EtherChannel at a given time.
    • The 2-byte system priority combined with the 6-byte switch MAC address form the system ID, which is used during negotiation with other switches.
    • This can be manually configured or automatically assigned by the swtich.
  • LACP Port Priority
    • LACP uses the port priority to decide which prots should be put in standy mode when there is a hardware limitation that prevents all compatible ports from aggregating.
      • For example, a set of 16 links defined for EtherChannel. Because only 8 are able to be bundled, the switch selects 8 ports with the lowest priorities and set as active. The other 8 goes on standby.
    • The 2-byte port priority + 2-byte port number makes up the port identifier.
    • The value can be determined by the switch or through manual configuration
  • LACP Administrative Key
    • Defines the ability of a port to aggregated with other ports, determined by:
      • Port’s physical characteristics such as: data rate, duplex settings, point-to-point or shared medium.
      • Manually configured restrictions.
    • LACP autmatically configures an administrative key value equal to teh channel group ID number on each port configured to use LACP.
    • It can also be manually configured.

LACP Modes

Resources:

1. Cisco EtherChannel Technology – Technology Whitepaper – Cisco Systems, Inc.
2. Understanding PAgP EtherChannel Configuration – Catalyst 6500 Configuration Guide 12.1E
3. Understanding LACP EtherChannel Configuration – Catalyst 6500 Configuration Guide 12.1E
4. IEEE 802.3 Standard - Click link –> Select User Type —> Accept/Begin Download. Go to “Clause 43, Link Aggregation”

This entry is not an authoritative guide. These are merely notes and rehash of the primary text materials and resources that I use. For a thorough guide of the BCMSN course, consider purchasing Building Cisco Multilayer Switched Networks (BCMSN) (Authorized Self-Study Guide) (4th Edition)by Richard Froom, Balaji Sivasubramanian, and Erum Frahim and CCNP BCMSN Official Exam Certification Guide (4th Edition) by Dave Hucaby ; as well as following the links on the reference section of this entry.

Still on the move. However, it is looking more likely that I will not meet my goal of taking the BCMSN exam in mid-July. I’m fine with that. In the past month I have come to see that for the last year and a half, I have been devoting way too much time on my study pursuit – starting from CCNA to my current status. There have been many things that occurred in the past year that allowed me to realize that I really need to slow it down one more notch. In a way, forces of nature made me come to this. Starting from the birth of my second child, then my mom becoming seriously ill, and now a heightened consciousness towards keeping my health in check.

I’m not stopping though. I just need to make sure that I continue to find that balance. Making sure that I continue spending that extra hour or two after work towards family time. In previous times, that was usually spent on reading, blogging, or labbing.

With my mom’s illness, I needed to make sure that I make that hour-long trip 3 times a week to spend 2-3 hours with my mom in the hospital. This has really taken a lot of time and patience on my part, but it is something that I gotta do. Seeing my mom in her present condition takes a lot of strength out of me and sucks out some of my mojo (lack of a better term). But this is also a time that tells me that I have to dig deeper inside and learn to fight through adversity. At least for my mom, if not for myself.

Lastly, I have taken on another challenge. Some of you who follow me on twitter know that I have been doing P90X. It’s an exercise program designed for coolness. So yeah, I’m trying to gain back coolness I sort of used to have in my earlier days. I call them my glory days – when I used to have a four-pack abs; or used to be able to almost dunk a basketball on an 8′ rim in my neighbors driveway; or pop a wheelie on my bmx bike for a cool 2 1/2 seconds. Coolness… here I come.

Next few  posts will be blog notes relating to EtherChannel technology.