QID:BCM300 - Why?

[routermatrix]

Could someone possibly explain the purpose of QID:BCM300?

Better yet, if someone could 'un-stupid' me as to why the explanation makes any sense that would be very cool.  However, I would be astounded if anyone has actually 'done' this scenario and could demonstrate results aligning with the question, solution and explanation. 

Intuitively the solution seems incorrect, but I went ahead and 'did' this one in a lab (several times actually on both real gear and GNS3) to just to see the problem first hand.  Just take a look at the host on Vlan3 (10.1.3.2/24) - why in the world would RT1 accept a packet on its SVI that does not match the prefix for that (ingress)  interface (let alone a network address which does not exist locally at all)?  This is the result of ARP debugging on RT1 in the lab when the host on Vlan 3 tries to leave the subnet:

RT1#debug arp
ARP packet debugging is on
RT1#
Mar 18 08:16:25.557: IP ARP req filtered src 10.1.3.2 c421.0bb4.0000, dst 10.1.3.1 0000.0000.0000 wrong cable, interface Vlan3
RT1#
Mar 18 08:16:27.561: IP ARP req filtered src 10.1.3.2 c421.0bb4.0000, dst 10.1.3.1 0000.0000.0000 wrong cable, interface Vlan3
RT1#
Mar 18 08:16:29.549: IP ARP req filtered src 10.1.3.2 c421.0bb4.0000, dst 10.1.3.1 0000.0000.0000 wrong cable, interface Vlan3
RT1#


The output 'wrong cable' does not strike me as particularly encouraging.  Meanwhile, arp debugging on the router acting as host3 shows that it has no way to map its configured gateway IP to a link MAC address (dst 10.1.3.1 0000.0000.0000 FastEthernet0/0) - you can see the 2  second  'throttled' delay interval between retransmissions in the debugging timestamps:


host3#debug arp
ARP packet debugging is on
host3#ping 200.1.1.2 rep 1

Type escape sequence to abort.
Sending 1, 100-byte ICMP Echos to 200.1.1.2, timeout is 2 seconds:

.Mar 18 08:16:25.565: IP ARP: sent req src 10.1.3.2 c421.0bb4.0000,
                dst 10.1.3.1 0000.0000.0000 FastEthernet0/0
.Mar 18 08:16:26.565: IP ARP throttled out the ARP Request for 10.1.3.1
.Mar 18 08:16:27.565: IP ARP: sent req src 10.1.3.2 c421.0bb4.0000,
                dst 10.1.3.1 0000.0000.0000 FastEthernet0/0
.Mar 18 08:16:28.565: IP ARP throttled out the ARP Request for 10.1.3.1
.Mar 18 08:16:28.565: IP ARP throttled out the ARP Request for 10.1.3.1
.Mar 18 08:16:29.565: IP ARP: sent req src 10.1.3.2 c421.0bb4.0000,
                dst 10.1.3.1 0000.0000.0000 FastEthernet0/0
.Mar 18 08:16:30.565: IP ARP throttled out the ARP Request for 10.1.3.1.
Success rate is 0 percent (0/1)
host3#
.Mar 18 08:16:31.565: IP ARP: sent req src 10.1.3.2 c421.0bb4.0000,
                dst 10.1.3.1 0000.0000.0000 FastEthernet0/0
host3#



Suffice it to say, the host on vlan 2 also has a gateway mismatch as well.

Information relevant to the question is supplied below.  The url to the lab PDF documenting this scenario (qidbcm300.pdf), I think, demonstrates that the solution and explanation for this do not make sense.  Again, if someone could point out something I'm overlooking that would be ideal. 

That being said, I really serious in stating that I do not understand what this question is about. Support has been really helpful to me with previous questions but for some reason have not acknowledged this one (in spite of pointing it out several times) - either the problem I'm seeing and/or something I'm missing.

--------------------------

Question
2. (QID:BCM300) View the exhibit. The network administrator has configured router RT1 to provide interVLAN routing between the various subnets on the network. Based on the output provided by the show ip route command, what statement is true?


* Users will have full access to the Internet but will not be able to communicate with each other

* Because of the missing routing protocol, none of the VLANs can route traffic.

*  Users on all VLANs can communicate with each other and access the Internet.

*  Users can access all the resources on the network but cannot browse the Internet.


Stated solution
Correct Answer: Users will have full access to the Internet but will not be able to communicate with each other
Explanation:

If you compare the addressing information for VLAN 2 and VLAN 3, as shown on topology, it does not match the information displayed in the output of show ip route command.

C    10.1.1.0 is directly connected, Vlan 2
C    10.1.2.0 is directly connected, Vlan 3

Therefore, users will not be able to communicate with each other. However, using the static default route 0.0.0.0/0 1/0 via 200.1.1.2, users will have access to the Internet.



Topology 



Routing table on R1



Lab Document:
http://routermatrix.net/config-labfiles/...bcm300.pdf

[joshuad31]

wow.  Its obvious that you went to great lengths to work on this issue.  I doubt if any answer that is provided would really satisfy you but as to why this question is relevant I will say that if a host is configured with a static IP address it is conceivable that said host might be misconfigured.

When I look at this it seems to me that there would be no way possible for 10.1.3.2 to reach out to the internet simply because there should be no sub interface on the router in the same subnet as that host.  It doesn't have a gateway to go out.

I wouldn't do this with GNS simply because you really can not emulate a switch.  Also you really only need one switch not three.  I don't see what difference it would make if you had three switches since all the hosts are in different Vlans.

My advice:  Don't let this get to you... not all the questions on the test make sense but all the questions on this website have a fair chance of being on the test.  Let it go and focus on something much more productive and remember that these questions were produced by other people who were not perfect.

~J

[routermatrix]

Hey Man,

    Thanks for taking the time to respond to my post - I appreciate that. 

Allow me to suggest a slightly different way of looking at it, if I may.  I'm very satisfied  with the product.  I have and continue to recommend to others.  Support has been fantastic and I have found the questions not only to be very accurate overall in terms of reflecting exam content, but more often than not paired with a useful explanation.  I have passed the routing, qos and wan/security tests and allot if credit is certainly due to using H2P.

More important though,  these practice tests have help improve the overall level of performance coming out of my finger tips, so to speak.  I use questions like this one (300) frequently as the basis for lab exercises & many end up in document format.  'Doing it' is really where the  rubber hits the pavement for me personally; verify or disprove  what I 'think'  I know and generally for building chops.  Also, another really cool thing is serendipity that tends to result.  For instance, I discovered that the IOS DHCP client can functionally pull options such as GW, DNS and NTP server info as a result of doing this lab.  This is a novel example, but it ticked me none the less as I had the expectation it would be stripped down to just getting an IP, for whatever reason (I've seen this work with IP routing disabled and enabled). 

The paradigm for science courses  including a lab component (activity plus write-up) is very sensible, in my option.  A teacher once explained it to me as overlapping modalities;  something to the effect that people tend to remember 5% of what they hear, 14% of what hear & read,  40% of what is hear+read+write and 90%  percent of what they read, write, hear and say.  To this extent what I'm writing right now and everything that follows is gravy.

I appreciate the comment that 'not all the questions on the test make sense but all the questions on this website have a fair chance of being on the test'.  That sounds very reasonable and may in fact be so, generally speaking.  This I do no know.  Still, my inclination is that the reverse may in fact be more so case - ALL of the (h2p) questions having a chance to pop up on the test strikes me as 'the tail wagging the dog' (please excuse the pun).

However accurate the overall content of this or any testing program/service may be, its still a 'copy' of the original source (actual exam content) and therefore necessarily  prone to some degree of reproduction error.  That the real  exam questions may themselves be lifted from white papers or base on other sources further characterizes the point of degrading precision per iteration.

Along the same time, the source (in our case the actual exam) is inherently going to dictate debugging/errata /corrections and therefore be ahead in that regard.  Its really difficult for me to imagine that this question (and solutions, as seen in qid bcm 300) EVER really appeared on the 642-812 exam - the more logical probability to me seems that modeled question became distorted in translation (photographic memory references not withstanding) and has yet to be corrected on 'our' end.  But, assuming it was at one time on the exam, it would be amazing (I don't mean that in a pop culture 'like sooo totally' way either) to me personally if this one would not have been weeded out by now by.  Vendors such as Cisco put forth a significant effort creating exam content.  My impression from personal experience is that most actual cert questions (at least) attempt to measure something meaningful content wise beyond simply being tricky for its own sake.  However, to the best of my understanding, none of the solutions provided for question 300 are correct.

Well, I was able to locate a white paper from Cisco that looks like it could be the original source  - the topology image is essentially the same and IP addressing looks identical.  The routing table gives the same info except that it matches the host addresses (vlan2=10.1.2.0/24 and vlan3=10.1.3.0/24).  On this basis I'm now figuring question 300 may be aimed at examining inter-vlan troubleshooting procedures (final portion of the document).  However, I would not be surprised if something else more specific could be pointed out that I am simply not seeing ins terms of the questions intent.

CISCO DOCUMENT REFEREENCE:
http://www.cisco.com/en/US/tech/tk389/tk...e74e.shtml

When I took the routing exam, one of the 'Testlest' had a question that matched the EIGRP problem on h2p - interpreting output from the 'show ip eigrp topoology' command.  Same question.  Until the day before the exam I could have sworn that the solution for the final question on the H2P EIGRP testlet was incorrect.  After obsessing over it, it finally dawned on me that objective was to deduce the location of an indirectly connected domain peer (or something to effect).  It was a tricky question but also a really good one,  I think, in that it measures the ability to crunch EIGRP metrics given default K values.  In the process of doing a subsequent lab on this and further reading, I further discovered Cisco routers do no perform floating point calculations, which also may need to be accounted for when observing or calculating distance values. This might be a relatively trivial thing for a heavy hitter, but I'm on the outside looking in that regard.  In my opinion, this formula is something that I or anyone on this level simply ought to know (ie. 28160 default for a connected FE link).  As far question 300,  part of my intension here in airing this out (beyond test scoring, which is of course really important) is to keep checking back to see that I'm not really missing the boat on something larger conceptually that I ought to know/recognize. 

It really makes me wonder what others think when reading this question.  Do any of those solutions make sense to anyone?  I know its not feasible to drill into every possible detail of everything.  Nevertheless, this one trips me out.  Many people must have looked at this buy now - is it more practical to assume something is true as a matter of faith, even it does not seem make sense? Or is it that it that 'It may not actually work behind the wheel (of a switch) but that's the answers the exam is looking for'.  To me, this really cuts to the crux of 'academic' vs applied knowledge/skill,  memorizing solutions without comprehension,  'studying for the test', 'paper certifications', etc (yada diddly-de...).  If I click on a selection on h2p or the switching exam itself, I'd prefer to understand why a choice makes sense beyond route memorization if at all possible (least Karma reduce me to one of a Zillion^x monkeys with typewriters in a far off galaxy pounding out the complete volumes of Shakespeare Ad infinitum).

Ok, now that I've B'tchd on & on (keep adding them, if for some strange reason anyone has read this far ... on....), maybe something a little more fun but at the same time relevant.  First, I have had problems with GNS3 (esp in Windows), but, lack of  functionality to reproduce question 300 is not one of them.  The ESW modules that you can plop into a router run PVST+ and provide sufficient switching features (i.e Vlan SVIs as show in the routing table output) to reproduce the scenario.  For bandwidth reasons I put the lab PDF I created in a better location (the ADSL upload out of my closet is brutal).  I don't expect anyone to necessarily take interest at all, but it is there none the less and won't take five hours to load.  There is also a GNS3 zip & self extracting exe of the activity, an image with interface labels and a raw text file with configs.  The GNS3 topology is configured so the hosts get IP addresses though DHCP that match the addressing in the routing table output on RT1 (question 300 image).  Create vlans on R1 and pump them out via VTP and the host/routers should be able to reach each other and network 200.1.1.0/30 (maybe some interface teasing with sh/no sh also..).  If ones gateway box can be programmed for static or dynamic routes, internet connectivity can arranged: throw a cloud down, configure it with your host PC's LAN interface, plug RT2s (200.1.1.2) other interface into the cloud and address it  to match the local LAN.  So, its there if anyone has the time or inclination to take it for a spin.  I used 3700 images with the module in slot two; that could be adjusted accordingly.
Hear is the URL:

http://cid-05734cd1766b331e.skydrive.liv...ublic?uc=1

For whatever its worth, I now know literal translation of the phrase Ad nauseam :)

[joshuad31]

Wow you are intense.  Well as to what you said:
"However accurate the overall content of this or any testing program/service may be, its still a 'copy' of the original source (actual exam content) and therefore necessarily prone to some degree of reproduction error.  That the real exam questions may themselves be lifted from white papers or base on other sources further characterizes the point of degrading precision per iteration."

Actually the questions are obtained through Van Eck Phreaking.  Being a security expert I can say with 100% assurance that this is the way it is accomplished.
http://en.wikipedia.org/wiki/Van_Eck_phreaking

But what does that mean as to your question... well nothing probably.  I personally agree with you that this question doesn't seem to make logical sense but at the same time without obtaining the actual tapes from which this question was lifted off of its hard to know what the question really said, what the intent or purpose of the question was.  99% of how to pass questions are the "real" questions.   It seems like we found one that is somehow missing something.  But again I would rather have a flawed question on this site that could at least gear my mind to being prepared for whatever real question is out there.  At least I know that test questions are prone to have IP addressing mix-up issues and I would look for that as a first line of attack.

I hear that Cisco is going to be implementing some measures to better protect their intellectual property and that future tests may not be so easy to capture.  But until the time I am very grateful for the resources available.  I like you have learned a lot by analyzing the questions and their answers.

Joshua

[routermatrix]

Hey Joshua,
    Thanks once again for taking the time to respond to my post.  Thats a fascinating topic a larger sense that you bring up - hand no expectation that  moaning about a question (productively for myself and possibly others, hopefully ..) would lead into industrial espionage, so I'll just leave it at that. 
 
    I must honestly say though, that the most interesting about your last post, far and away, was not addressing what I pointed out about GNS3 having switching functionality.  That was easily the most valuable thing I brought up in that entire diatribe practically speaking (including the link to the Cisco site on intervlan routing).  Were our places reversed, I would imagine being enthused to get that piece of info.  And please excuse me if that sounds ticky tacky in any way, as its not my intent or interest AT ALL.  I Just strikes my curiosity as switching is the main thing on my mind these days. 

  The ESW modules are limited, but can do all that is required (trunking, access ports, PVST+ etc) to run useful scenarios, such as one that was the basis for this thread.  I'm open to correction if I'm misunderstanding the program.

http://www.blindhog.net/gns3-how-to-buil...ching-lab/

Good luck it (switching and all else),
Dave

[Renquest]

I have nothing to add or take away.  This was a pleasure to read.