Skip to main content

Rapid Spanning Tree Protocol

You may think why we may need RSTP. Go back my previous blog post where I wrote how STP work and its port state. Now we are going to talk about RSTP. Lets learn about it and you may know why we need it.

Recall my previous theory, as we already know STP is created very long time ago and STP have below problem to converge the link.

1. Listening : 15s of listening for BPDUs. Switch sends/receives BPDUs on this state
2. Learning : 15s of Learning MAC Address, populate switch CAM table
3. Forwarding : Port is Forwarding Traffic
4. Blocking : Switch will wait up to 20s before moving a block port into listening phase, because this the time where switch wait if the primary link may came live.

STP port state take minimum 30s to maximum 50s to link up. This is the port process which have to pass every switch when boots up or converge time.

STP downtime is the biggest problem for todays Network.

STP have problems but also we have solution !

Port Fast :

Port fast disable Spanning Tree. When you type this command on an interface it will display warning message. Like this



Because it won't really detect the node mention in Warning message. Just enable the command and unplug the cable then plug back in use show command it’s immediately transition into forwarding state because now we have the port fast enable.

You can enable port fast all end node connected ports, that allow immediate active.
Switch(config)#int range fa0/2-24
swirch(config)#spanning-tree portfast
****output omitted****
On the other hand we have problem with uplink link. We can't enable portfast on the uplink ports, as you can see in warning message. 50s downtime cause big problem to our network. So there has to be other solution for this problem. Yes you already heard about it, RSTP. Rapid Spanning tree is very fast and very good. RSTP give us faster convergence, yes what we need. Think about VoIP, dropping one packet isn't tolerance for it. Other transfer like file transfer can be retransmitted. But voice can't be retransmitted. Real time transmitting must need, we don't have loss in the link.

Some of you still using uplink fast or backbone fast, that wouldn't going to work if you have different switch other than cisco. Those are cisco proprietary. So RSTP standard have availability to participate all vender switches. Which speed things up. RSTP is very good, you can see in your network.

 RSTP with its port state 


Discarding : Its replacement of blocking state of STP, we study before. It’s mean, taking down layer 2 connectivity not sending out any traffic.
Learning: Its mean exactly STP's learning state, it is the process to learn MAC address is on that port, without it switch just act like hub, its forward every packet to every port.
Forwarding: It’s also like STP port state, forwarding means not blocking any traffic.

Port roles

Root port : The way reaching to root bridge.
Designated port : All root bridge port are designated port. Other switch have lower priority or lower mac address elected as designated port. Other switch should be block the traffic.
Alternate port : Instead blocking state in STP, where all port should be re-elected and take 50s to downtime of network. RSTP remember that port with alternate port to reach Root Bridge. If root port failure, its immediately turn it into forwarding state. So it’s doesn't have to wait 50s.
Edge port : Port fast is edge port. Where non switch are connected or host. We can use port past command to tell switch it is edge port.

Why   RSTP is better

a. RSTP doesnt forget its port unlike STP which re-elected when primery link is down.

b. Its dosn't use extra time of STP like listening BPDUs, bonus blockig penalty,

c. If change into trunk port, that will flood to each network with TC packets, where TC is topology changes. Those are part of the BPDU's.

Comments

Popular posts from this blog

MongoDB Failed to start lsb an object/document-oriented database

After long gap I'm going to post new blog. Due to the work load and research I have little time.  MongoDB could not started after installation of 3.6 version, However downgrade to the 3.4, still the Mongodb couldn't start and show below error message. Search don't help to resolved the issue. After so much hit and trail nothing work. One Chinese forum posted this command and its works for me.  cd /var/lib sudo rm -rf ./mongodb sudo mkdir mongodb sudo chown -R mongodb mongodb/ sudo serivce mongodb restart

Cisco ASA in GNS3

Here is another tutorial running Cisco ASAv on GNS3 using Qemu. For Configure GNS3 In My case I have used ASAv952-204.qcow2 Go  to the GNS3>Edit>Preferences>QEMU>Qemu VMs>New Follow on screen procedure.  Don't forget to enable kvm and memory allocation.  Now you can see I can run ASAv 9.5.2 Blank Password.

IOS XR GNS3 QEMU

This time lets have tutorial on Cisco XR 9k series router image running on the GNS3. Please don't ask for the XR image. Your are smart enough to get it. My system configuration: Ubuntu 16.04 GNS3 1.4 RAM 8Gig i7 processor Used XR Image iosxrv-k9-demo-6.0.1.qcow2 This image is VIRL extracted image. You need to convert this image into QEMU image, follow this link I strongly recommend you to run it on the Linux system. Now you have converted image, then go to the GNS3>Edit>Preferences>QEMU>Qemu VMs>New then follow the onscreen procedure. Setting for QEMU XR Image. RAM:- 4Gig CPU:- 1 Adapters at lest 4.  -enable-kvm Here you can see I can run the XR on my system. Interface is up and system is already booted.  I have run 3 XR router  here is my system RAM CPU usages. RP/0/0/CPU0:XR3(config)#int gi0/0/0/0 RP/0/0/CPU0:XR3(config-if)#ip add 192.168.13.2 255.255.255.252 RP/0/0/CPU0:XR3(config-if)#co