Answers to Chapter 7 Review Questions

Answers to Chapter 7 Review Questions

1:

Label the port types (RP=Root Port, DP=Designated Port, NDP=non-Designated Port) and the STP states (F=Forwarding, B=Blocking) in Figure 7-30. The Bridge IDs are labeled. All links are Fast Ethernet. Assume that there is only a single VLAN and that the portvlanpri command has not been used.

A:

Figure A-6 provides the labels requested in Question 1 for Figure 7-30.

Figure A-6. Two Back-to-Back Catalysts with Crossed Links

Cat-B becomes the Root Bridge because it has the lower BID. Cat-A therefore needs to select a single Root Port. In the previous examples of back-to-back switches, the links did not cross and Port 1/1 became the Root Port because of the lower Port ID (0x8001).

In this case, the crossed links force you to think about the fact that it is the received Port ID that influences the Cat-A, not Cat-A's local Port ID values. Although Cat-A:Port-1/2 has the higher local value, it is receiving the lower value. As a result, Port-1/2 becomes the Root Port. Understanding this issue is critical to effectively use portvlanpri load balancing.

2:

When do bridges generate Configuration BPDUs?

A:

Bridges generate Configuration BPDUs in the following instances:

• Every Hello Time seconds on all ports of the Root Bridge (unless there is a Physical-Layer loop.

• When a non-Root Bridge receives a Configuration BPDU on its Root Port, it sends an updated version of this BPDU out every Designated Port.

• When a Designated Port hears a less attractive BPDU from a neighboring bridge.

3:

When do bridges generate Topology Change Notification BPDUs?

A:

Bridges generate Topology Change Notification BPDUs in the following instances:

• A bridge port is put into the Forwarding state and the bridge has at least one Designated Port.

• A port in the Forwarding or Learning states transitions to the Blocking state.

• A non-Root Bridge receives a TCN (from a downstream bridge) on a Designated Port.

4:

How many Spanning Tree domains are shown in Figure 7-31? Assume that all of the switches are using ISL trunks and PVST Spanning Tree.

A:

10+3+2+15=30.

Although the same numbers are used for all of the VLANs, the routers break the network into four Layer-2 pockets (Cat-1 through Cat-3, Cat-4, Cat-5 through Cat-6, and Cat-7 through Cat-10). The VLANs in each Layer-2 pocket then form a separate STP domain.

One of the tricks in this layout is to notice that Cat-5 and Cat-6 form a single Layer-2 domain containing two, not three VLANs. Because of the backdoor links between the two switches, the routers do not break this into separate Layer-2 pockets.

5:

When is the Root Bridge placement form of STP load balancing most effective? What command(s) are used to implement this approach?

A:

When traffic patterns are well defined and clearly understood. In hierarchical networks such as those adhering to the multilayer design model discussed in Chapters 14 and 15, Root Bridge placement is an extremely effective form of STP load balancing. Simply collocate the Root Bridge with the corresponding default gateway router for that VLAN (see Chapters 14 and 15 for information). For non-hierarchical, flat-earth networks, load balancing usually requires different VLANs to have server farms in different physical locations.

When placing Root Bridges, either the set spantree priority or the set spantree root commands can be used.

6:

When is the Port Priority form of STP load balancing useful? What command(s) are used to implement this approach? What makes this technique so confusing?

A:

This form of load balancing is rarely useful. It can only be used with back-to-back switches. It should only be used in early versions of code or when connecting to non-Cisco devices. The set spantr ee portvlanpri command is used to implement this feature. This technique can be very confusing because it requires that the set spantree portvlanpri command be entered on the upstream switch.

7:

When is the Bridge Priority form of STP load balancing useful? What command(s) are used to implement this approach? What makes this technique so confusing?

A:

The Bridge Priority form of STP load balancing can be useful if you are using pre-3.1 code and cannot use Root Bridge placement (because of traffic patterns) or portvlanpri (because the switches are not back-to-back). If you are using 3.1+ code, portvlancost is generally a better choice. The set spantree priority command is used to implement this approach. This technique can be confusing for several reasons:

• The Bridge Priority values must be adjusted on devices that are upstream of where the load balancing takes place.

• The Bridge Priority values must not be adjusted too low or your Root Bridge placement is disrupted.

• It can be difficult to remember why each Bridge Priority was set.

8:

When is the portvlancost form of load balancing useful? What is the full syntax of the portvlancost command? What is the one confusing aspect of this technique?

A:

The portvlancost form of load balancing is useful in almost all situations. It is the most flexible form of STP load balancing. The full syntax of the portvlancost command is:

 set spantree portvlancost mod_num/port_num [cost cost_value] [preferred_vlans] 

One confusing aspect to this command is that it only allows two cost values to be set for each port. One value is set with the portcost command and the other is set with the portvlancost command.

9:

What technology should be used in place of portvlanpri?

A:

EtherChannel.

10:

What are the components that the default value of Max Age is designed to account for? There is no need to specify the exact formula, just the major components captured in the formula.

A:

The default Max Age value of 20 seconds is designed to take two factors into account: End-to-end BPDU propagation delay and Message Age Overestimate.

11:

What are the components that the default value of Forwarding Delay is designed to account for? There is no need to specify the exact formula, just the major components captured in the formula.

A:

The default Forward Delay value of 15 seconds is designed to take four factors into account: End-to-End BPDU Propagation Delay, Message Age Overestimate, Maximum Transmission Halt Delay, and Maximum Frame Lifetime.

The last two factors (Maximum Transmission Halt Delay and Maximum Frame Lifetime) could be simplified into a single factor called "time for traffic to die out in the old topology."

12:

What are the main considerations when lowering the Hello Time from the default of two seconds to one second?

A:

Lowering the Hello Time value can allow you to improve convergence time by lowering Max Age or Forward Delay (you have to do this separately) but also doubles the load that STP places on your network. Notice that load here refers to both the load of Configuration BPDU traffic and, more importantly, Spanning Tree CPU load on the switches themselves.

13:

Where should PortFast be utilized? What does it change about the STP algorithm?

A:

In general, PortFast should only be used on end-station ports. It allows a port to immediately move into the Forwarding state when it initializes. Other than that, the processing is the same. When using redundant NICs that toggle link state, it can also be useful for links to servers.

14:

Where should UplinkFast be utilized? In addition to altering the local bridging table to reflect the new Root Port after a failover situation, what other issue must UplinkFast address?

A:

UplinkFast should only be utilized in leaf-node, wiring closet switches. After a failover, UplinkFast must generate dummy multicast packets to update bridging tables throughout the network in addition to updating its own bridging table.

15:

Where should BackboneFast be utilized?

A:

To work correctly, BackboneFast must be enabled on every switch in a given Layer 2 domain.

16:

Where is PVST+ useful?

A:

PVST+ is useful when you are trying to connect traditional PVST Catalyst devices with 801.Q switches that only support a single instance of the Spanning-Tree Protocol.

17:

Can MST regions be connected to PVST regions?

A:

MST and PVST regions cannot be connected through trunk links (MST switches only support 802.1Q trunks, and PVST switches only support ISL trunks). However, the two types of switches can be connected through access (non-trunk) links (although this is rarely useful).

18:

Can you disable STP on a per-port basis?

A:

STP cannot be disabled on a per-port basis on Layer 2 Catalyst equipment such as the 4000s, 5000s, and 6000s. In fact, some Layer 3 Catalyst switches (Sup III with NFFC) require that STP be disabled for the entire device (all VLANs).

19:

Why is it important to use a separate management VLAN?

A:

It is important to use a separate management VLAN to prevent CPU overload. If the CPU does overload as a result of excessive broadcast or multicast traffic, the Spanning Tree information can become out-of-date. When this occurs, it becomes possible that a bridging loop could open. If this loop forms in the management VLAN, remaining CPU resources are quickly and completely exhausted. This can spread throughout the network and create a network-wide outage.

20:

What happens if UplinkFast sends the fake multicast frames to the usual Cisco multicast address of 01-00-0C-CC-CC-CC?

A:

If UplinkFast sends the dummy frames to the usual Cisco multicast address of 01-00-0C-CC-CC- CC, older, non-UplinkFast-aware Cisco Layer-2 devices do not flood the frames. Therefore, this does not update bridging tables through the network.