IEOC - Internetwork Expert's Online Community

Sat for the R&S the other day in San Jose - here is my feedback/experience:

Regadring the lab itself, I agree with adriano's last post stating that the lab was around a level 7 difficulty. After going through almost all labs in workbook 2, I was able to move quickly through with less than 2 hours to spare. Double-checked my work, fixed a couple issues and time was up. The only trouble I ran into were the questions regarding technologies/features that were not covered in the Class on demand / Workbook Vol I&II. I feel my biggest mistake was not reading through the configuration guides to fill in the gaps. A buddy of mine who passed on his first shot printed out most of the configuration guides and read through them cover-to-cover!

Leaving the Lab, the thing that has worked for me was to ask myself, "If I were to fail, would I know why?" I felt somewhat confident in answering "No" - which is good.I've found that in previous attempts (as well as attempts in other IE's) that if you're learning a couple new basic things while going through the lab or you can answer "Yes" to the previous question, another attempt might be in order...

Results came in around 4 hour mark - most other report around 3 hours.

Results = FAIL

I went through the results and they didn't seem right - I actually had good marks - all except for Core Knowledge section...THE SHORT ANSWER SECTION....  0%. How could that be? Not sure - those questions were pretty easy and I checked my answers on CCO after I submitted them. Now, I figured that what happened was the proctor went through my lab and once he discovered I failed - he marked the rest of the test randomly. This does happen - especially on the Voice exam - depending on who is grading (I've seen this a couple times personally and heard from others).

The next day, I sat down to calculate this out. Figure the total lab is about 79 points and short answer 21 points - where if you fail short answer, you fail exam. Based on that, you can figure out total points lost based on percentages and add 21 points - assuming you remember how many questions there were in a section and approximate point values. Lucky for me, I did remember the values. I did the math, double checked it, verfied it with a friend (who is good with math),  and discovered that, without the short answer, I scored an 84....pass...This tells me that the proctor who graded my exam actually did mark me 0% on Core Knowledge and failed me based on that.
I started to think about the short answer questions I had gotten ... did I really mess them up? I verified again on CCO and found that I may have misinterpreted one question, but still gave an answer that would strongly suggest I know what I'm talking about....at least in my opinion.

After thinking about the short answer, here is my feedback:
The questions are easy, but make no mistake about it - you can fail the exam in a heartbeat based on them. Everyone across the forums is for the addition of questions - as am I, but there is a major problem with them:

I'll use Anthony's latest sample questions, because they're almost exactly what you can expect.

Example 1:

-Which multicast PIM method uses a push and pull approach?-

A CCIE should know this answer. This should be drilled into him/her. This is something you would know after going through lab after lab. It is critical to know this otherwise you may not be able to interpret a question on the lab itself. This is the sort of question that should be on the exam because someone who studies should know this like the back of their hand. This is multicast 101 - If you answered the question incorrectly, you should not have your CCIE.

Example 2:

-Which bits of the ToS byte are used for DSCP markings?-

If I asked a CCIE this question and he didn't know the answer off the top of his head - SO WHAT?!? (!) This question is not something that you drill on. This has nothing to do with configuration (nor design). What does this tell you if someone can answer this question? They know how to calculate DSCP markings? CCIE is a configuration exam (design is thrown out the window) which should test configuration experience, not book smarts.
A better question?
-What method of marking allows you to assign a value of af31?-
That could be asked on the CCIE exam (mark blah blah with value of af31). If you've configured enough QoS - you should know the answer to that!

My point is, these questions need to be designed around what the typical CCIE should know - months after he has passed. Imagine you study for months and months, get to the lab and get asked "which bit is set in the header of the hello packet that would cause the router to blah blah). Yeah - you read over the answer ages ago - but you forgot it because YOU DIDN'T NEED TO KNOW IT FOR THE TEST.

I suggest - make the proctor take the short answer too. He's a CCIE - he should know this stuff. You think the proctor remembers what multicast address protocol X uses to communicate off the top of his head after passing so long ago? I don't think so.

I've rescheduled for my next attempt already ... re-reads take about a month - at this point, it's not worth fighting over.

Please note: If I had received a PASS, I would still have put in the time to make the same comments as above.

Scott,

This is how I calculated my score:

Based on percentages, figure out how many points lost assuming you know how many questions there were per section.

% = Points awarded / Total Points. Points lost = Total Points - Points Awarded.

Since I knew the number of questions per section, I played with the first formula (testing different total points) until I received the percentage on my score report. My total points lost = 16. 79 total points - 16 points lost = 63. 63 remaining points + 21 short answer points = 84.

Anthony - according to the proctor, the lab and short answer scores are not independent. Also, according to my score report, it shows they are not independant.

My point is not so much the scoring - it's the purpose of these questions. According to the Certification Online Support, the reasoning behind the open questions are as follows:

"Including several randomly-selected short answer questions on the lab will allow Cisco to increase the validity of CCIE and ensure only qualified candidates achieve certification."

This is great, but the questions asked do not reflect the goal. A CCIE should not be expected to know "What is the size of the 802.1Q tag?" off the top of his head two years after passing. That question is no different than asking random questions about default timers, ect. He should know:

-"Which trunking encapsulation encapsulates the entire frame?" or "Which encapsulation tags all packets by default?"

Short answers are a good idea, but need to be fixed. Again, even if I received a PASS, I would still have posted the same feedback.

noairbag:

Regadring the lab itself, I agree with adriano's last post stating that the lab was around a level 7 difficulty.

I've been wondering about this for awhile... Many folks have said that the lab ranks about an IE 7. But looking through vol II, there are exactly three labs that are a 7 (labs 6, 12 and 18). Seems to me that vol II is a bit top heavy with most of the labs being an 8 or more (12 labs). On one hand, I can see the benfit to learning more complex labs to better understand the technology and to develop confidence. But on the other hand, wouldn't it be better for those interested in passing the exam to have more labs that are equal in difficulty to the actual lab?

Thoughts?

Dear noairbag

It is realy wondering if they ask about the Size of the protocol headers. actually it scares me too.

were your short answer question like that ? i mean they realy ask you about the protocol size or default timers?

I just took this test last week.  I'm sorry to disagree with Scott, but some of the questions have no place in a lab environment.  They look like they came out of a test pool for the CCIE written.  Some were fair, and could be part of a CCIE lab exam, and others would be known only through straight memorization.  BTW, I did not pass.  I did extremely well on the lab itself, even scoring 100% in three of the core sections (My failing lab report showed me).   Unfortunately, I failed because I scored a 0% in Core Technologies.  Percentage-wise, it is hard to believe I would not have passed a lab that didn't have these questions, but that is neither here nor there.  So it is possible to do well on the lab but fail because of the questions.   Also, the max amount of time you could really spend on the questions I had was 15 minutes.  There was NOTHING open ended about them.  You either know them or you don't.

I am now going back to the CCIE written material covered in the lab blueprint before my next test.   It takes away from the lab portion, but at 21 points of the overall score (the lab exam added up to 79), you cannot afford to overlook them.   On the other hand, it may be simpler to pass the lab if you know your written material very well.

noairbag:

Example 1:

-Which multicast PIM method uses a push and pull approach?-

A CCIE should know this answer. ... If you answered the question incorrectly, you should not have your CCIE.

I agree 100%.

noairbag:

Example 2:

-Which bits of the ToS byte are used for DSCP markings?-

If I asked a CCIE this question and he didn't know the answer off the top of his head - SO WHAT?!? (!) This question is not something that you drill on. This has nothing to do with configuration (nor design). What does this tell you if someone can answer this question?

I think this question is just as the first... if you can't answer this, you don't deserve to be CCIE certified. I don't care if the CCIE is a configuration exam, you should still know your basic theory.

Just my 2 cents.

Anyway, sorry you didn't pass, and I wish you the best of luck on your re-take!

I would say the real lab i had was around 8 -9.No less than 60% on just Switching/IGP , was tricky enough to had me spent more time in those sections. So yeah i think its a good idea to go through the harder labs and finish them with good timing.

IPv6Freely:

noairbag:

Example 2:

-Which bits of the ToS byte are used for DSCP markings?-

If I asked a CCIE this question and he didn't know the answer off the top of his head - SO WHAT?!? (!) This question is not something that you drill on. This has nothing to do with configuration (nor design). What does this tell you if someone can answer this question?

I think this question is just as the first... if you can't answer this, you don't deserve to be CCIE certified. I don't care if the CCIE is a configuration exam, you should still know your basic theory.

So what's the right answer? Is it....

the 6 most significant bits

the 6 left most bits

bits 1-6

bits 0-5

11111100

In my opinion, this kind of question should be presented as a multiple choice because any of these answers could be right depending on what the tester used as reference materials. Yet, we don't know what the grader is using to determine the correct answer. It's not an open-ended question.

I'm pretty sure the first two and the last one would all be accepted as correct, as they are black and white answers. The other two are too open for interperetation by the proctor.

Either way, my point was you should know which bits they are... I wasn't talking about the grading method. I was replying to the thread starter who seemed to no think knowing the answer to such a question was important.

IPv6Freely, you're reaching - I don't think knowing which bits they are applies to theory. I think the idea of ToS and DSCP - how they are used, when to use them, difference between DSCP and IP Prec, ect. - is part of theory. Counting bits and bytes tells you I memorized something - it doesn't show you I understand something. If I can explain DSCP and IP Prec using bits - then I show I understand something. If you can't explain difference between IP Prec and DSCP, then no CCIE for you!! But number of bits/bytes...come on....
Thanks for wishing me luck on my next attempt, though. I wish you luch as well.

trmas explained it correctly - the questions are more black and white than open ended. Let's put it this way, if all you studied for the lab involved all of internetwork expert's products (class on demand/ Workbooks I->III), you would not be prepared for the "open ended" questions. The answers to my short answer questions were on CCO and in the CCIE Written book. Until my next attempt, I will be reviewing my CCIE written book......for the lab.....

IPv6Freely:

Either way, my point was you should know which bits they are... I wasn't talking about the grading method. I was replying to the thread starter who seemed to no think knowing the answer to such a question was important.

If the proctor is Israeli, he might think that the most significant bits start on the right side....ok, I jest.

To your actual point...I'm with noairbag on this. I don't think this question is appropriate. I think the scenarios in the CCIE lab and the lab questions should directly relate to situations or issues that one might see in the real world. Knowing which specific bits are used for a particular encoding is not really relevent because it's not something you can tune anyway -- it's just the implementation according to an RFC. How exactly does knowing which DSCP bits help understand how to tune and troubleshoot QoS?

Perhaps a better question would be "Describe what types of traffic are most appropriately assigned Low Drop Preference?" That's open ended. Now, knowing how much overhead a particular technology adds is relevent because the designer has to make a choice (between 802.1q or ISL, for instance). But again, that's not an open ended question.

The reality is that most people who gravitate toward the proctor role didn't do to well working with customers (as SEs or in TAC). So you are going to get questions that don't quite fit in with the actual objectives of the program; which, ultimately, is to generate sales revenue for Cisco.