STP Vs. RSTP – Answer 3
Today’s question and answers actually exercise more STP logic than RSTP, with the RSTP materials basically reviewing terms and features. Look back to the original question first before looking at this answer. As usual, the answer post gives the answers, and the reasons why each answer is either wrong or right.
Related posts:
General Advice
Before listing the answers, let me summarize a few key points that can let you quickly rule out some of the answers as incorrect.
Ruling Out Port Roles that do NOT Apply to STP
One key difference between RSTP and STP relates to two port roles added by RSTP: the alternate port and backup port role. STP simply does not include these port roles. So, knowing this fact, any of the answers that list “STP” as the protocol, and either alternate port or backup port as the port role, cannot possibly be true.
Root switches Cannot have Root Ports
Both STP and RSTP use the same logic of choosing one switch to be the root switch. Then, both STP and RSTP have each non-root switch determine its port that’s part of the best path back towards the root switch. That port is that switch’s root port.
As a side effect of the above rules, one switch – the root switch – does not have a root port. So, any question that identifies the root switch cannot then claim a root port exists on that switch!
Correct and Incorrect Answers
The answers, plain and simple. Ask questions if you have them.
(Note: All three RSTP answers happen to mention port roles that RSTP has in common with STP. All the logic shown with each answer applies to both STP and RSTP.)
RSTP, SW1, G0/2, Designated Port – Correct. SW1 is the root switch, so it will advertise cost 0 Hellos onto each of its links. The root switch will therefore always win the election to be the designated port on every one of the root switch’s ports in that spanning tree.
RSTP, SW2, G0/1, Root Port – Correct. The obvious possibility for SW2’s root port is SW2’s G0/2 port, which connects directly to the root switch (SW1). However, SW2 may pick its G0/1 port as its root port. For example, if all links are running at gigabit speeds, with all defaults, SW1’s root cost over G0/2 will be 4, and its root cost out G0/1 will be 8, so SW2 picks G0/2 as its root port. A configuration change to make SW2’s G0/2 port have a cost of 9 makes that path have a greater root cost, so SW2 picks its G0/1 port.
RSTP, SW3, G0/2, Designated Port – Correct. SW1 becomes the root switch, so SW3 (G0/2) will compete with SW2 (G0/1) to become the designated port. With all defaults, they must use tiebreakers, but any configuration change that gives SW3 a better root cost than SW2 will result in SW3’s G0/2 becoming a designated port.
STP, SW1, G0/1, Alternate Port – Incorrect. The Alternate port concept is an RSTP concept, and does not apply to STP.
STP, SW2, G0/2, Backup Port – Incorrect. The Backup port concept is an RSTP concept, and does not apply to STP.
STP, SW3, G0/1, Alternate Port – Incorrect. The Alternate port concept is an RSTP concept, and does not apply to STP.
Figure 1
Hello Wendell, I have the same answers that you in questions 1, 4, 5 and 6. I used another assumptions in questions 2 and 3, so, I obtained differents results than yours and I wanna know if I’m correct. In both questions, I assumed all default STP costs, without configuring new ones.
2) Incorrect. With the default costs, it is not possible that G0/1 is RP because it is not the least cost path to the root, unless G0/2 (the RP) fails.
3) Incorrect. Given that SW2 G0/2 is RP and therefore SW2 G0/1 is DP (if SW2’S BID< SW3's BID, because there's a tie in root cost=19), the SW3's G0/2 can't be DP too, so in RSTP, it must be ALT. If SW3 was STP, G0/2 would be in BLK state.
Thanks in advance and thanks for this way to teach us, giving always your best.
There’s a mistake in my answer, in the 2nd item the root cost that ties is 4, not 19.
For the 2nd item, I think this sentence should be related to SW2 instead of SW1: “with all defaults, SW1’s root cost over G0/2 will be 4, and its root cost out G0/1 will be 8”, am I right?. Thank you.
Made same 2 mistakes than you Marcelo, and unsure why we’re wrong… 2) I still don’t get how G0/1 becomes RP… , and 3) Though that would be in bloking state as well.
Hey Ben,
Turns out the early comments in this thread are years old – impossible to tell, but ignoring your #1 in your list of 3 for that reason…
Your #2 – Per the explanation above, configuring SW2’s G0/2 port with a cost of 9 or greater, with default costs on other ports in the topology, gives SW2 a lower root cost using its G0/1 port… making G0/1 its root port.
And #3 – not getting your point on that one. Feel free to follow up…
Wendell
Hey Wendell, just a quick question on rstp behavior.
When RSTP converges, it selects root port, then designated THEN alternate and backup roles if additional ports beyond the 2 root and designated receive bpdus for the same spanning tree instance?
It goes in order because alternate ports end up discarding and that would break RSTP if it selected alternates first, in this case if everything was default, Switch 2 g0/1 could be an alternate port.
Taylor,
Yes.
🙂
I think I followed you, but yes, a non-root switch begins by determining which switch is too, then by selecting a root port to reach that switch, before doing anything else, including choosing an alternate to the root port. I think was the crux of your argument, but I must admit, I didn’t quite follow it all. But if that connects and confirms the dots for you, great!
Wendell
It was more so, how does STP decide what to make other ports (that are not root ports). does it choose designated ports first, before thinking about any alternate ports? Or does it choose alternate ports first then designated? But I got it now! Designated is necessary, so it is selected first, because it is forwarding. Anything else (alternate or backup) would be discarding and that could break STP. To choose alternate ports first, could result in no designated/forwarding ports, and only a root and alternate port, so only leftover ports (after root and designated) are considered for alternate and backup roles.
Sometimes I have a hard time wording my questions 🙂 Thanks!
Hello Wendell,
Please I need some clarification on your explanation to this answer:
“RSTP, SW2, G0/1, Root Port – Correct. The obvious possibility for SW2’s root port is SW2’s G0/2 port, which connects directly to the root switch (SW1). However, SW2 may pick its G0/1 port as its root port. For example, if all links are running at gigabit speeds, with all defaults, SW1’s root cost over G0/2 will be 4, and its root cost out G0/1 will be 8, so SW2 picks G0/2 as its root port. A configuration change to make SW2’s G0/2 port have a cost of 9 makes that path have a greater root cost, so SW2 picks its G0/1 port.”
I picked out this answer as correct before reading your explanation. What confuses me about your example here is the underlined portion. Why would SW1 being the root have a cost of 4 & 8 through its G0/2 & G0/1 interfaces respectively? My understanding is that as the the root switch cost is always 0 and root will advertise the cost of 0 to any link connected to it. How did the cost of 4 & 8 arise? Or did you mean to say SW2’s cost? Because that’s the only scenario I could come up with that makes the cost possible.
Hi Tewa,
You can think about how STP works from a human perspective, looking at a network diagram. You can also think about the process the switches use. The explanation here is probably using some of both, which makes it a bit confusing. So, focusing on what the switches do…
SW1 is the root switch. It sends Hellos, listing itself as the root switch, cost 0.
When SW2 receives a Hello on any inbound port, SW2 does not consider it’s root cost to be equal to the value in the received message.
For instance, receiving the cost 0 Hello directly from SW1, on its G0/2 port, SW2 adds its G0/2 port cost to the cost in the received hello. So, using its G0/2 port, SW2 has a root cost of 4. Then, say SW3’s root cost is also 4, through its root port of G0/1. So SW3 forwards a Hello to SW2, listing cost 4. When SW2 receives the message, it adds the cost in the Hello (4) to SW2’s G0/1 cost (4) to get 8. So, SW2 chooses between the two options: cost 4 and cost 8. The cost 4 path is best, so the port in that path (G0/2) is its root port.
But if those interface STP costs were different due to configuration, you might get different answers.
Hope this helps,
Wendell
PS See CCNA 200-301 OCG Vol 1 Chapter 10 section “How Spanning Tree Works” for more detail.
Thanks Wendell for the explanation. I now understand the point but I think you got the port numbers mixed up.
or instance, receiving the cost 0 Hello directly from SW1, on its G0/1 port (I think this should be G0/2 as per the figure?), SW2 adds its G0/1 (Should be g0/2?) port cost to the cost in the received hello. So, using its G0/1 port (G0/2?), SW2 has a root cost of 4. Then, say SW3’s root cost is also 4, through its foot port of G0/1. So SW3 forwards a Hello to SW2, listing cost 4. When SW2 receives the message, it adds the cost in the Hello (4) to SW2’s G0/3 cost (4) (there is not g0/3 in the figure, should be G0/1?) to get 8. So, SW2 chooses between the two options: cost 4 and cost 8. The cost 4 path is best, so the port in that path (G0/1) (you mean g0/2?) is its root port.
Tewa,
You were indeed correct. Thanks for the thorough review and fixes! I have edited the above comment. Let me know if I missed something.