I’ve seen a lot of discussion recently about the CCIE with regards to how “hard” the test really is. There is no denying that the exam is of a very high difficulty level. The discussion revolves around application. It has been said that one of the reasons the CCIE lab exam is so difficult is because it doesn’t test the candidate on “real world” network designs. According to these folks, the CCIE lab tests you on things that you would never see yourself doing in reality outside a lab environment solely for the purpose of seeing how well you can follow directions. There is some merit to this, as the overview for the CCIE clearly states that this is not a “best practices” examination of networking theory. It’s a practical implementation test with a given set of parameters and instructions. There was also a story told in one of my bootcamps with Narbik Kocharians about a student taking a mock lab that took two hours to finish the first section because he spent all his time doing it the “right” way and ensuring there couldn’t be any possible problems down the road. He thought like an engineer working on a production network instead of a CCIE candidate. Those clues tend to lend credence to the idea that the CCIE is hard because you are doing things you might not do otherwise.
The more I thought about this, the more I realized the CCIE lab exam is a lot like another type of test that almost every one of us has taken at some point in our lives – a spelling bee. The time honored tradition of rounding up a group of students and giving them strange words out of the dictionary to see how well they can disassemble them and regurgitate them back in serialized order. When you think about it, there’s a lot in common with the granddaddy of networking exams. Both are practical, in that multiple choice isn’t allowed (curiously, the first round of the Scripps National Spelling Bee uses a multiple choice format sometimes, similar to the CCIE written qualification exam). Both exams don’t give any points for partial credit. Transposing two letters of a word gets you the same number of points as forgetting to enable mls qos on a switch in the lab (zero). Both exams give you all the answers up front. For the CCIE, it’s all there in the documentation. In the spelling bee, you usually get a word list of some kind, either the Spell It! book or Webster’s Third International Dictionary. In both cases, the amount of documentation that must be sorted through is rather large. Both tests tend to introduce a large amount of performance anxiety. And finally, both tests tend to focus on things you wouldn’t normally see for the sake of testing the candidate’s abilities.
Think about this for a moment. The winning words for the last three National Spelling Bee winners were (in order) cymotrichous, Stromuhr, and Laodicean. I can’t even pronounce those words, let alone use them in general conversation. There are even differences in the vocabulary I use in my blog posts versus the words I use in conversation. Does it make the above words any less valid if the only appear in a dictionary? No, it doesn’t. Yes, many of the constructs in the CCIE lab are presented in such a way as to test the candidate’s grasp on applying concepts. Yes, the lab is crafted in such a way as to eliminate several obvious choices that make life easier. Just like a spelling bee doesn’t give you access to the dictionary. Yes, there is a time crunch in the lab. Just like a spelling bee doesn’t give you three hours to think about how to spell the word. You only have 2.5 minutes to spell the word from the time it’s first pronounced. Overall, both the spelling bee and the CCIE lab exam take specific examples that demonstrate advanced concepts and give the test takers a short amount of time to produce results. It shouldn’t matter that I may never configure multi-router redistribution or RIP neighbor relationships across RSPAN VLANs. The point is that these examples are designed to test my knowledge of a subject, just like cymotrichous is designed to test my spelling ability a lot better than dog or cat.
There’s no denying the CCIE is a hard exam. The question of real word application versus crafted lab scenarios is a semantic one at best. While many feel that making the exam reflect scenarios that you might encounter in your job every day would be more appropriate, I feel that having it test a broad subject matter with intricate questions is a better application. I’d much rather be looking at a problem and think to myself, “Hey! I’ve seen this in the lab before!” That way, I feel more comfortable having seen it work in a controlled environment before. At the end of the day, making the CCIE lab a “real world” test is as bad an idea as making the National Spelling Bee only test over words used in everyday conversation. The test would soon become a very rigid and insular example of the mythical “real world” that would either need to be updated every six months to stay current or it wouldn’t be updated frequently enough and eventually become what people are accusing it of today, namely being a “bad” example of the real world. I think it’s better to stretch our horizons and spend a little time thinking outside the box for solutions that may not apply in every day life but force us to think about our methods and processes. Whether that involves routing protocol configuration or challenging the “I before E except after C” rule, the end result is the same. People question more and dig deeper rather than just accepting someone’s idea of what reality looks like. And, after all, we know that in our world, I and E really come after two Cs.
Not having taken the lab, but having talked to many who have, it seems to me that there is something to be said for testing one’s ability to deal with strange and poor practices, since that’s really the way many networks get cobbled together. We rarely walk into a greenfield deployment. Usually it’s a network that already exists and was not built with current (or sometimes any) best practices. In some ways, the unusual configurations in the lab do reflect the real world, just in a highly concentrated way.
You’re absolutely spot on with this statement…
My observation is that the lab (the R&S at least) uses contrived situations to test the candidate’s knowledge of the underlying protocol theory. They are trying to put you into situations you haven’t seen before in order to see whether you can find solutions to bizarre problems based on 1) the theory of how things should work, and 2) your mastery of troubleshooting techniques. This is why having a really strong background in protocol theory and an exhaustive knowledge of show and debug commands is a big advantage.
My experience in the real world has been that both of those things (theoretical knowledge and a mastery of show and debug) is extremely valuable.
You cannot be more wrong. That is why you can have an experienced Security guy with 25 years in the market and he would fail the CCNA security… sucks