Handling MPLS task on the Lab


Wanted to get your feedback on how one should be handling MPLS tasks and "full connectivity" requirements.

What I find is that the labs are layed out in a particular order, ie.-Switching then Routing, then BGP/MPLS, then services. This makes sense as this is how on should build this. 

When we get get to the MPLS section in our labs, when configuring PE to PE connectivity, we may lose connectivity to those subnets which form the MPLS core. Now, to me, this makes sense as in the real world we do not see any routes/information that creates the SP's MPLS backbone. We only see the routes from our other sites; anything that is imported and is redistributed between BGP and our IGP.

On the exam, how are we to handle this? Are we to assume full reachability/connectivity, even though there is the posibility of lose of reachability due to the technolgies being implemented? Or are we to accept the lose of those subnets, at the sacrafice of meeting the requirement? Or are the labs meant to take this into account and would not put a student in this awkward scenario?

Or am I just overthinking and worrying too much [*-)]



  • You shouldn't expect to have to offer full reachability between different VRFs.  INE labs put a certain terror into candidates that full connectivity is required in every single scenario, but this isn't necessarily the case with the actual lab exam.  I found the actual lab exam to be far more straight-forward than some of the INE mock labs.


    Still, it's actually a good question, and the best thing you can do is to always always always ask the proctor for their opinion.

  • Thanks Kish. I was thinking, well rather hoping, that this would be the case.

    I just think its just pure evil that the scenarios presented may have you doing some filtering from BB routers if your IGP peering, then have you in a later task configure BGP to those routers, only to later on enable MPLS, thus undoing the earlier requirements.

    It used to be that you should be on the lookout for these traps, but now it seems like this is what is suppose to happen!

    I'd like to take your word on the actual lab being a bit more straight foward, but judging by some of the feedback by others on this forum and friends who have taken the exam, makes one second guess themself.

    I suppose I will go back to my lab and cram some more and hopefully will be able to ask the proctor the "right" questions!


  • I found the requirements on the actual lab quite straightforward, the requirements for reachability were explained, so I did not have to assume anything. If you are unclear just ask the proctor.

    I highly recommend you read over the exam before starting, this way you can avoid any situations where you have to undo some configuration because of a later task requirement.

    If your lab has any MPLS make sure you understand how this relates to your IGP config, it is very likely that the MPLS config will come after IGP config, so make sure you are aware of the relationship between your IGP and MPLS. If you rush ahead and do all your IGP config before understanding your MPLS requirements then you may get in to a situation where you have to start changing your previous config, and the last thing you want to do in the lab is start panicking in this situation and making mistakes. To avoid these sorts of traps just take a few minutes to read over the exam and then start working through it with a solid plan, stay nice and calm and you will find that the exam is not too difficult, assuming you know all the technologies :)



  • You shouldn't expect to have to offer full reachability between different VRFs.  INE labs put a certain terror into candidates that full connectivity is required in every single scenario, but this isn't necessarily the case with the actual lab exam.  I found the actual lab exam to be far more straight-forward than some of the INE mock labs.


    Still, it's actually a good question, and the best thing you can do is to always always always ask the proctor for their opinion.




    I agree with you, I feel personally that INE mock labs really, almost misrepresent the v4 experience at least from the config section, they still seem to hang onto the v3, 8.5 hour all configuration mind set.


    I used INE for my one and only v3 attempt (and still do for v4), but back then their mock labs really got me in shape for the 8.5 hour all configuration end-to-end enterprise style network v3 tested for, and I aced it..... stupid Flanders O.E.Q. Section....


    INE's mock labs for v4 really need to almost drop the focus on full full full reachablity... oh and route redistribution, and they have made me good at it, but oh man.... if there was a 'CCIE Route Redistribution' track, it would be an over kill for it.


    Just my two cents....

  • In INE's defense, they want you to be prepared for anything.  Redistribution is very much a part of the v4, though it may not be as large and intense as v3, but I wouldn't know that since I never studied for or attempted the v3.  I actually feel like the INE mock labs prepped me pretty well for the v4.  It seemed to accurate reflect the layout and format, and the technologies on my exam were a subset of the technologies covered by INE.  I didn't see anything on the exam that surprised me.


    Just to clarify, Mario, I'm not saying that the lab exam is simple.  When Jason and I say that the requirements are straightforward, we mean that there's little ambiguity.  And the nice part about the exam is that there's a proctor available for clarifications.  I asked the proctor quite a few questions, mostly just to confirm what I already believed to be true.  The only real ambiguity I saw was in one of my diagrams, which the proctor quickly clarified for me.  Nearly everything else was me just hitting the panic button.


    Cisco isn't out to trick you.  The lab exam is an amalgamation of all the technologies covered on the blueprint.  If you're comfortable with the technologies, or at least where to reference them on the DocCD, then it largely comes down to whether you can make everything work in the time allotted.  Don't downplay time management, it's one of the major tripping points for candidates.



  • I hear that the lab isn't simple, if it was I wouldn't be taking it a 3rd time! [:)]

    I really do not fear any of the technologies on the exam, as to everyone's point if you've studied hard, it won't matter what they throw at you.

    Also, I do realize that this is a test, and the goal of a test taker is to employ whatever testing strategies to pass. If you study hard, you will be rewarded.

    Now I've highlighted that I do not fear most of the technologies, as I do understand that each protocol and technology has its limitations, thus the whole point of attaining a CCIE.

    My concern is that I am told I must have full reachability, introduce a technology like MPLS where it would hide some subnets, and as a result cannot reach those subnets and BAM! A $1400 donation to Cisco! But i digress...

    As far as a testing strategy, it used to be a rush to 80 points. Its my understanding that this format constitutes passing every section, sections being troubleshooting and configuration. So would passing a section suggest 80% correct? If so, would foregoing a section in the configuration, like MPLS, be wise? This would put the test taker in a situation where everything else must be 100%, but assuming the point value for that section is lets say only ~8% of the overall exam, do we still have a chance to pass? I.E.- I have 92% of the configuration correct, with the loss of points due to a section being incomplete?

  • Sorry Mario, I didn't realize the context of your question. I assume you took v3? Everything I've heard about v3 is that it was very strict regarding full reachability requirements.  The v4 isn't nearly on the scale, and again my exam was very straightforward with what should be reachable and what should not.


    Yes, you must pass each section of the exam individually, each at 80%.  For troubleshooting, you'll be given some number of tickets, probably 10 or 11, and you need to get all but 2 correct.  Configuration then requires 80% as well, and sure, you can skip MPLS or any non-core section if you can get your points elsewhere. But then it's a game of numbers, and whether you can get through the rest of the exam without losing too many points.


    If a task is core, however, skipping it could fail you. For example, if you're given an MPLS requirement that affects full reachability, you'd be best not to skip it.

  • My apologies for the confusion.

    Do the same rules apply for version 4 around not using static routes and no new subnets apply?

    Again, this is in reference around the MPLS section, where I would imagine the possibility of such a scenario where full reachability is a must,so could one use something like a static route to leak a subnet? Assuming that something like this would be permitted.

    Just speaking in generalities, and trying to better understand the new format. I will take your word for it that the exam is more direct and clear on their requirements rather than spend time on interpreting what they want you to accomplish.

  • I can't really vouch for much since I've only taken the exam once, and only the v4 exam. I advise taking 1 minute to read through the exact rules of the lab, and use the proctor for any confusion.  To my knowledge, there's no inherent requirement to attain full reachability, so just do as the lab instructs.


    The rules on my exam were exactly what I expected based on the INE lab exams.

Sign In or Register to comment.