Alright, so I’ll go straight forward with the STP variations or enhancements. It is notable that 50 seconds convergence time of STP is very slow nowadays. A 50 second delay in service may cause a heartbreak, headache and a bad day, right? IEEE made a standard or improvement of STP and it is called RSTP. Anyways, I forgot to tell you that STP is also an IEEE standard 802.1d
RSTP / 802.1w
Rapid Spanning Tree Protocol only has three port states :
Discarding- combination of disable, blocking and listening
Learning- STP fucntion of learning ; learns MAC address but does not forward frames
Forwarding- STP operational function; learns MAC address and forward frames
As a summary, RSTP provides faster convergence because it has simpler process or less steps to go from discarding to forwarding. The image below would be a great help for us to understand STP as well.
Regarding with the convergence time of RSTP, I am not with the authority to declare how fast RSTP is because my research says its either 10 or 19 secs but probably it is faster than STP. So with regard to the port states, RSTP just combine some of the states and that becomes faster. However, RSTP does not end with that. RSTP has amendments or addition of port roles. Lemme clarify that there is a difference with port states and port roles.
Port States are blocking, listening learning, forwarding and disabled while port roles are designated, root port and alternate block port. In RSTP, designated and root ports are maintained while alternate block port is split into back-up and alternate port.
Alternate port receives more BPDU message within ANOTHER bridge. As the name implies, it provides an alternate path to the root bridge and can replace root port if the link fails. Figure shows that BPDU message from B to another bridge which is A. Remember that is the port which is blocked.
Back-up Port receives more message on the same bridge as you can see in the figure, the port is only blocked on the same bridge. This port does not guarantee path to the root bridge but provides redundant in the event link fails. So which is better, it always depend on the situation but alternate port wins.
PER VLAN SPANNING TREE
Per VLAN spanning tree is a Cisco proprietary spanning tree protocol. How it works? The goal is to balance the load to each switch. Here’s the thing, cisco switches are very expensive yet effective. Can you still remember the hierarchy for switches? There are core, distributed and access switches. We always wanted a core switch for the company but always be aware that not all companies can afford Cisco core switches. Core switches are considered as backbone switch and has capability of to handle more traffic compared to other switches. We always assign core switch as root bridge but what if we don’t have core switch? A solution for this is the so called PVST or per VLAN SPANNING TREE.
From the name itself,”per VLAN”, it means to say that if we want just to use access switches and we would like to consider the capability of that switch to handle the traffic. We will just distribute the load to all access switches by letting the access switches be the root bridge per assigned VLAN. Let say SW1 will serve as root bridge if we will send the traffic to VLAN 10. SW2 will serve as the root bridge if we will send the traffic to VLAN 20. SW3 will serve as the root bridge if we will send the traffic to VLAN 30. In that way, access switches will not be overwhelmed.
MST ( Multiple Spanning Tree)
PVST is great!, we can balance the load of traffic to be assigned to switches. It actually has enhancement which is RPVST+ or Rapid PVST+ that has convergence of one second. The main drawback of PVST is that what if there are 300 vlans need to be assigned in a network? It may sound “ok” but come to realize that it will require CPU’s power and memory for the switch to calculate the root bridge, ports, etc.?
In MST, they set an instance to the group of VLAN.For example we will set VLANs 1-30 to instance 1 and so on. Therefore, we can have just 10 instance with 30 vlans each and it will have less CPU utilization. In order for switches to be part of the same instance, we require switches to have the following:
3. VLAN # associated
Let’s go with the configuration to easily understand it:
SwitchA(config)#spanning-tree mode mst
SwitchA(config)#spanning-tree mst configuration
SwitchA(config-mst)#instance 1 vlan 2,8,22,6
SwitchA(config-mst)#instance 2 vlan 13,23,44,55
SwitchB(config)#spanning-tree mode mst
SwitchB(config)#spanning-tree mst configuration
SwitchB(config-mst)#instance 1 vlan 2,8,22,6
SwitchB(config-mst)#instance 2 vlan 13,23,44,55
SwitchC(config)#spanning-tree mode mst
SwitchC(config)#spanning-tree mst configuration
SwitchC(config-mst)#instance 1 vlan 2,8,22,6
SwitchC(config-mst)#instance 2 vlan 13,23,44,55
You would see that switches share the same configuration name, revision# and vlans included in instance.
The necessity to configure STP?
STP is enabled by default. So why we have to configure STP?
Always remember that the basis to elect a root bridge is the MAC address of the switch. The thing here is that the older the switch, the lower its mac address. And since we endorse core switch as a root bridge, and access switch is manufactured earlier then if we just let STP runs in a network without configuring STP, it may happen that access switch will be elected as root bridge and that will not maximize the potential of the core switch.
*****************************END OF BLOG*******************************