STP PROCESS

[cosimodagostino]

Hello everyone, without complicating our life with RSTP or for VLAN spt etc. etc., let's assume that we have 4 switches A - B - C - D connected to each other with an active and an alternative path. A is the root bridge; On D we have a designated port that connects to B and therefore operational and a port that connects to C in a blocking or alternating state. Let's assume I need to add an "E" switch; this sends the pdu bridges and goes into a block state. If in my network topology everything is working correctly and there is no need for malfunctions and therefore there is no need for stp to redo the calculations, but I want to add a switch, after how long does stp proceed to redo the calculations to also include switch "E" in the topology?

[balaji.bandi]

Can put small diagram how these connected. If you make a Loop between D and B, it calculate the best path and rest will be blocked as per STP process.

If you using rapid-pvst its quicker, again this in technical.

STP root bridge election to be configured based on the topology.

https://www.ciscopress.com/articles/article.asp?p=2832407&seqNum=6

[Martin L · Accepted solution]

RSTP is similar to original STP but it is much more faster!  RSTP uses synchronization process where a proposal-agreement "handshake" between closest neighbors happens.  This is very fast, 1 seconds or less!  There is a TC BPDU sent over to all neighboring switches over on nonedge ports so that they can flush its MAC tables.  TC happens only when nonedge port moves to Forwarding state (port goes up).

If u add an "E" switch to D, First they do sync handshake; Then, D will do sync process with its neighbors. Sync is so fast that I had hard time to spot ports going to blocking state (called Discarding in RSTP).  I guess it depends on type of TC BPDU (superior or inferior).  If your new switch agrees that A is the root sw, there is no "significant" topo change; no need to inform sw A about new sw E, except to flush its MACs, which I have hard time to see in the lab as well.

If you have A - B - C - D topo and add new Superior sw E to D that will become new Root. First sw E sends Proposal to all neighbors on its nonedge ports. Sw D gets it; Now D must do sync with E, so D puts port to C into Discarding state before sending Agreement back to E (so that there are no loops).  Once Sw E and D are done (agreed), the link between D and E is now Forwarding immediately.  Now sync process is performed between D and C.  Sw D sends proposal to C who sees new Root sw; C moves its nonedge ports into Discarding state (to avoid possibility of loop forming) and sends agreement back to D. and so on.... this is called sync "wave" of proposal-agreement handshakes.

The process of sync is fast; likely 1 second or less in simple topology with modern switch models.

In case of adding "old" legacy STP switch E to D, that link will wait 2x15 seconds before Forwarding (Listen and Learn MACs before Forwarding them).

Regards, ML
**Please Rate All Helpful Responses ** 

[paul driver]

Hello
in rstp it’s almost immediately just long as it takes for d to send a proposal to e and e to sync its non edge ports and to agree on that proposal

if d doesn’t receive an agreement to it’s sent proposal then the ports will fallback to the old stp timers of stp (listening learning forwarding) as in rstp these timers are initially not used as such transition much faster.

please review here