Christmas Musings of a dissillusioned student

[djf99]

Hi,

First of all, Merry Christmas everyone.

This post is going to sound like a bit of a bitching session - sorry about that - but I really do want to get a few things off my chest, and I *really* do need some sensible advice.

I've bought a copy of the Sybex CCNA Certification Kit, which includes TL's CCNA study guide (6th edition). I've been using it exclusively as self study preparation for the CCNA 640-802. (I considered going with the book and real hardware - but decided with a sim - especially one designed to go with an appropriate course - it would be possible to set up more elaborate networks than I'd ever have enough hardware to do myself). I've read the book, done all the suggested exercises, worked through all the test questions at the end, the flash cards etc.

Probably somewhere between 50-100/hrs work. I've got a long background in IT, but negligible Cisco experience (I know, this sounds odd, but it's true) so I found the material presented in the book new (sort of), but not especially challenging. With one or two grey areas still, I'm getting 90% plus on bonus exams on the included CD - and where I'm falling down is mostly with the symantics of some questions and the odd picking the wrong answer when I "know". It's very hard for me to look at a 191.X.X.X address and "see" it as a public class B for example, my brain says "private class C" before I've reached the 1. Or in other words, I was at the point I was learning more about TL's writing style than Cisco networking.

I *thought* I was pretty near ready to sit the exams. I was a bit light on for "practical" work on the router sims, but I decided I'd try and assess myself based on the public samples of prep test-questions to see what extra prep work I should do, and to see what I should buy (if anything). Obviously no so good. 50-80% typically.

I haven't tried Transcender yet - I baulked at the price - more than sitting the Cisco exam.

As I'm sure most people here are aware, there are a lot of terrible exam prep products out on the net, and some of the sample questions were just obviously wrong, and certainly had wrong answers. But they also asked questions (typically 20-25%) about things that TL's book either doesn't cover at all, or does so in a purely cursory. I might be wrong, but does the book tell me if 100BaseFX has a max cable run of 400m or 412m?

So, tried the Brain Dump sites. I did a bit better, but was *much* slower as these "real" questions looked complicated - and sometimes were and sometimes weren't - but all took time to read fully and carefully. Plus I sort of did this open book (checked my answers as I went from source materials rather than the suggested answer), especially after reading here that submitting wrong answers suggested by the dump sites is a near guaranteed fail.

From this exercise, it's pretty clear I'm not ready to sit the CCNA exam. I suppose it could be possible that the sample questions provided were more difficult than average to help persuade me of the need to buy their products, but I don't think so. If the dump sites' questions are *real* questions

But ... more disturbingly, I don't think I could *ever* be ready to sit just by reading TL's book (with or without hands on experience). Granted, when I checked, most of the material is in there. Often just in the glossary - or at times there is one definition rigorous enough to meet the exam question requirements - with the same thing repeated several times elsewhere without enough detail to know how to answer a question directly - or worse - enough information to answer the question *incorrectly*.

For example, in my travels I've seen 2 questions along the same theme about OSPF Router IDs. The first had two obviously wrong options, but the last two choices of IP addresses on interfaces, the highest one was administratively down. Hmmm. My first thought was probably that the fact it was just administratively down wouldn't matter. Why force all those DR elections and stuff over the ID - which is hardly going to be any less unique because someone has shut down an interface? (OK, thinking about it more there *are* some very good reasons). BUZZ. The online Question flunked me. I decided that wasn't enough though, so I decided to get some more "hands-on" experience and fired up RouterSim, set up OSPF on a router with two interfaces, one up & one (with highest IP address) ad-down, & sure enough - it allocates the Router ID as the highest IP address of the known interfaces: the one that is shutdown. I quickly read over TL's book - and it seemed to confirm the RouterSim diagnosis: the question's suggested answer was wrong. (Turns out I missed a key word in the book: "active" IP addresses).

Still not convinced, I googled through the Cisco online docs. Most references refer to simply "highest IP address", but I found a pertinent FAQ entry about duplicate Router ID names, where the process (I'll post the link later - it's my edu machine) of Router ID selection explicitly referred to IP addresses on up/up interfaces.

And it was only at this point I realised that loopback addresses take precedence. I re-read TL's book, and yes - this is stated in one of the 3 explanations of this process IIRC - but most of the time TL's book just talked about the "highest IP" address (as do most references to this process on the Cisco docs site I might add). And IIRC in the book's examples, the loopback address always *was* the highest IP address IIRC, which only re-inforced my mis-conception.

My doubts about my understanding of this process have been enhanced by finding another sample exam test question where the choice is between loopback0 and loopback1, where loopback1 has the highest IP address. That test marked me right for selecting loopback1. Everything I've been able to find on the Cisco site suggests this is right too, as does the book. But ... the only place I've ever seen a concise, unambiguous and complete explanation of the precise OSPF Router ID allocation algorithm (there must be one on the Cisco site somewhere - but I haven't been able to find it - unless of course it varies between ISO/router versions) suggested that *only* loopback0 was checked. This might have been the best written explanation and most detailed, but as it was in the explanation to the suggested correct answer to this type of question, I can't really be sure if it's right or not.

But my main point is, the full algorithm to assign OSPF Router ID is clearly something Cisco now expects new CCNAs to know, yet a simple unambiguous expression of this seems to be elusive. I've spent hours on something that should take 3min, and I'm still not sure. I guess if I had dynamips or some real routers I'd be able to find out for sure, but I'd still be concerned about different behaviours with different versions.

Another example, TL's book tells me I'm expected to know the STP port states. I duly learned these for STP, only for one of the brain sump sites asked me "which of the following are valid RSTP port states?", which I'm pretty sure are not covered at all. And I think I'm expected to know the STP algorithm too, not just the port states - which I've yet to learn. It might be in the book and I just haven't recognised it because I don't know it yet, or not.

Clearly, TL's book alone is not enough to pass the CCNA (at least the 640-802). It simply does not cover all - or glosses over too many of - the things that are now expected of CCNAs.

Now, I know there is this disclaimer about "no book can be enough" repeated many times here and somewhere in the book itself. But quite frankly, that argument is a load of total BS. It's like saying: "no written manual can properly explain the workings of our products, you just have to get first hand experience them". If you were talking about hitting golf balls of tuning a car engine you might have a point, but the CCNA is a purely theoretical exam/certification, and the "no book can be enough" argument is simply no excuse for not appropriately presenting the material students are expected to be examined on. And on this subject, there is an all to obvious reason this disclaimer isn't on the front cover which - given the way the product was marketed to me - would have been the most appropriate spot.

Please don't get me wrong here, I thought the book was well written and easy to read, and I appreciate there is a balance between dryness and preciseness. I just suspect it no longer covers enough, in enough detail, to meet the needs to students sitting the current 640-802 exam.

Anyway, I'm at the point where I'm just not sure what to do.

.... to be continued