Answers to a Root Question with No Direct Link to the Root
This post answers the Spanning Tree question linked here. If you want to noodle the question a little longer, check out the answer to a similar question that showed three switches in a triangle. To solve this question, just use the same concepts shown with that switch triangle. Then start ruling in the correct answers, and ruling out the incorrect answers. Details and answers below the fold!
Answers:
A, B, E
Figure 2: Four Switches, S4 as the Root
Ruling in the Right Answers
First, you can (and probably should) repeat the triangle analysis from the previous question. Why? In this question, a triangle exists between root (S4), S1, and S2. The choices made by both S1 and S2 then determine the Hello details they send to S3, and the current question is about S3’s choices. So, the same logic seen in the earlier question, with a single switch triangle, can be used to analyze S1’s and S2’s choices of root port in this question, which in turn affects S3’s choice of root port.
So, repeating that same analysis described in the previous question’s answer, first think about the S1, S2, S4 triangle. From that, you can determine what settings matter to S1’s and S2’s choice of root port, which in turn impacts S3’s choice.
S1’s logic includes S1’s port cost on ports F0/4 and F0/2, as part of two possible root paths
S2’s logic includes S2’s port cost on ports F0/4 and F0/1, as part of two possible root paths
To choose the root, all three switch’s bridge ID’s (BIDs) are considered
From this analysis, we can rule in answers A, B and E. If you want more background on whys and wherefores, check out the answer to the previous STP question, while spells out more of those details.
Rule Out the Rest!
If you understood the analysis from the previous question, you could actually rule out the rest of the incorrect answers as incorrect, without thinking about switch S3 at all. Why? First, two of the incorrect answers mention the port priority, and as pointed out in the answers to the previous question, the port priority is examined as a tiebreaker – but only in cases when a switch has multiple parallel links between two switches. In this question, no such parallel links exist. As a result, the tiebreaker of lowest-neighbor-Bridge-ID will break any ties first, before needing to consider port priority. So you can rule out both answers that mention port priority.
The other incorrect answer mentions the root switch (S4) and the port cost on one the root switch’s interfaces. As discussed in the answer to the previous question, the cost of the root’s ports does not impact the other switches’ calculations of root cost. For those same reasons, this answer can be ruled out as incorrect.
Hi. I have a question about the involvement of F0/1 on SW2 in choosing the root port for S3. I thought that the first step in choosing a root port for a non-root switch was to find the route with the smallest cost to the root switch adding all the outgoing ports costs. For example if we are talking about S3 there would be three posibilities:
SW3 F0/2 -> SW2 F0/4 -> S4
SW3 F0/1 -> SW1 F0/4 -> S4
SW3 F0/1 -> SW1 F0/2 -> SW2 F0/4 -> S4
And if there was a tie, the tiebreaker would be the lowest-neighbor-Bridge-ID. So I’m not seeing SW2 F0/1’s involvement anywhere in the decision of choosing S3’s root port. I’m sure that I am probably wrong but I would be glad if you could tell me why. Great questions, by the way :).
Hi Victor,
Glad you like these questions! They’re so good, I have to think hard about them when I get questions like yours. 🙂
Anyway, you missed on combination: S3 – S2 – S1 – S4. That path uses outgoing port F0/1 on S2 (which leads to S1).
Wendell
Yes, I see it now. Thank you so much for answering.
Hi Wendell,
I got the answers but tackled them slightly differently. Instead of looking at them like triangles I instead mapped out all the various paths to the root from S3, and then added up the default interface costs. Similar to how you tackled the STP puzzle.
From there I scanned the interfaces in the questions to see if they were used in any of the various paths and what the impact could be if I changed the default setting.
e.g.
S3 – S2 – S4 = 19 + 19 = 38
S3 – S1 – S4 = 19 + 19 = 38
S3 – S1 – S2 – S4 = 19 + 19 + 19 = 57
S3 – S2 – S1 – S4 = 19 + 19 + 19 19 = 57
Was this a correct way to answer this?
Hello Wendell,
In the book, page 232 Table 9-10 you have made a typo regarding RSTP state for port administratively disabled state you put discarding, while it should say disabled.
best regards
Barnabas
This is unrelated with the post, but when electing a RP of a switch, what happens if there are three devices, SW1, hub, and SW2 connected colinearly and there are two lines between SW1 and hub, 1 line between jub and SW2, and SW2’s BID is better than SW1, and finally SW1’s two interfaces connected to a hub is working with the same cost? SW2 is root, SW1’s two interface’s root cost ties, sender’s BID ties, sender’s port priority ties, and sender’s internal port numbers in the STP BPDU messages tie.
So which interface is RP?
Hi Karma,
Well, the practical answer is that it doesn’t matter, as it’s a somewhat ridiculous design in any network built in the previous 20 years. I’ve never tested it, so I can’t say with confidence. Anyone who cared about STP design and predicting the root port wouldn’t build that design. But I get the idea as a curiosity.
If I had to guess, I’d say it’s based on SW1’s lowest port priority between the two ports, and if that’s a tie (eg, using defaults), then it’s SW1’s lowest internal port number comparing the two ports. On Cisco switches, the internal numbering tracks to the the outwardly-visible numbers, eg g0/1/0 is lower/better than G0/1/1.
The rules for the more normal tiebreak cases (still unusual designs today) are in chapter 9 of Vol 1, by the way.
What happens to MAC address table entry’s inactivity timer, if the ‘destination address’ of a frame matches that of the entry? Is the timer reset or not
Matching a destination address has no effect on the timer for MAC table entries. Only the source MAC of an incoming frame (which is used for learning the MAC address) resets the timer to zero.
Also, can a switch be simultamepisly in more than one different STP (or RSTP) roles or states?
First, a switch port has a port role and port state, rather than the switch itself. That said, within the spanning tree for a VLAN, one switch interface has one port role and one port state. EG, a port can be the switch’s root port, or a designated port, or an alternate port, but not two or more at the same time. Same deal, can’t be forwarding and blocking state concurrently.