powmia's CCDE marathon

hoogen82

Good luck to you..

Essendon

I'm sure this is in the bag. This dude's different I tell ya!

snadam

good luck!

Jobene

Viel Glück:)

Best of Luck

instant000

And the result?

carterw65

Yea, results please!

powmia

Successfully Failed!

Essendon - It's a different kind of test, I tell ya

Jobene - Vielen Dank!


The Test

---

First off, I've never been f'd with so much by an exam. It was the most fun exam I've taken. There was not a single "I've never seen/heard of that before" or "oh crap, I'm not familiar enough with that protocol" moment for me. There was only a single question out of 120 (that's an average, it's public knowledge that each scenario is 25-35 questions), that I did not know the technical reason for an answer. I understood everything else very well. Unfortunately, unlike the CCIE, knowing everything just isn't enough (sounds messed up, right?). It's hard to describe the experience well enough, without breaking the NDA, but I'll give it a shot.

I'm going to chalk this failure up to the fact that I have now seen that the only way to prepare for this exam, is to take the exam. There are a few people that have passed this on the first try, more power to them... those are few and far between. I had been told a couple of times to just go take the exam, at the early stages of my study. I thought they were crazy. Why take the exam if I'm not prepared? I've also seen multiple recommendations from those
that have passed that the best thing to do is get a CCIE, get to the point at which you are ready for the CCDE, then brute force the exam (take it and take it until you pass). I now fully understand both of these recommendations. It appears to be pretty common, as well. There were 7 of us total yesterday. I confirmed with a few, but my impression was that everyone already had a CCIE. Only a few of us were there for the first time. This was, at
least, the second or third attempt for most of the guys. I saw one that passed (not his first time), I'm not sure if anyone else did.

For the most part, I was aware of my mistakes while I was taking the exam.

Mistake #1: Time management

I had no idea how the time would go in this exam. There aren't any vendors that provide a CCDE workbook with mock labs and an accurate time to completion that should be aimed for. I finished my first section (two scenarios) in about 2 hours and 45 minutes. It seemed like I was going too slow during the scenarios, which caused me to speed up a bit. Then, the scenarios went MUCH faster towards the end of them, and not because I was going faster. Most likely, because you move from analyze (most time consuming) to design (time consuming) to implement (goes very quick). If I had a better feeling of how this would go, I would have taken my sweet time and probably would have thought my decisions through quite a bit more. Instead, I got a lovely message saying that I had to wait 2 hours and 15 minutes before resuming the exam after lunch. I finished my second section (last two scenarios), in about 3 hours... same thing.

Mistake #2: Being a Rookie

I definitely made some CCDE Practical Exam Rookie mistakes. I let them mess with me too much. Here's an example:

I'm consulting for Company XXX. Long story short, their overall architecture has come down to a decision between Technology 1 and Technology 2.

Immediately, in my mind, I see technology 1 as the best solution for their network. I then get a series of documentation and go through a series of questions about Technology 2. The questions are pointing out the benefits and drawbacks of both technologies. These are subtly, but obviously leading me in a direction that emphasis the benefits of Technology 2 and the drawbacks of Technology 1. After 30 minutes of being convinced that Company XXX's engineers and management believe Technology 2 is better for their very specific circumstance, and that I should believe so too.... I get the golden question. Which one do you recommend, 1 or 2? I have an internal dilemma and break. I select Technology 2. Then, I get a series of follow on questions and customer chosen solutions that blatantly point out that I am a moron for choosing Technology 2. Lovely. I also get the follow on questions wrong, because they have no correct answer.. they are for the wrong solution in the first place.

Being a rookie was my biggest downfall. I will attribute the bulk of my failure to this. There were multiple instances in which I went down a rabbit hole and should have known better.

Mistake #3: Being an engineer

This one wasn't so much of a problem for me, but did affect me a bit. I thought too much about some things. I had a few instances, in which I was thinking what would be best for the network, as opposed to what would be best for the customer.


Two Complaints

---

1: For $1500, you would think that a room full of CCIEs that are having their minds bent for 8 hours would be provided with coffee. I'm pretty pissed about that. I left a comment for Cisco on my first question after lunch, regarding the coffee situation.

2: I had one scenario that I felt was written poorly. It felt like a beta. There was a long series of questions about doing something, in which none of the options were even remotely close to how things should have been done (I know, that's part of the exam). But, pretty much, they were looking at a bidirectional problem in only a single direction. I had a hard time choosing the best answers, since I didn't feel like any of them solved the problem. Of course, it is very possible that I just didn't understand what was going on, but I was pretty confident about this.

At one point, I hit a question that got me pretty ticked. During the exam, you have what are called branching questions. If you chose a certain solution, you will go down a branch that follows up with how to implement that. Sometimes, if you chose a wrong solution, you will get a question with no correct answer. Well, I got a question with no right answer. I did not do anything remotely close to what would have branched me to this question, so I have to believe that this was just another part of the exam. Basically, I had options for how to implement a specific concept. That concept, in any way shape or form, was impossible to perform. Not because it was theoretically or technically impossible, but because the documentation that I was given explicitly laid out information regarding the network that made this concept a no-go. I am 100% sure of that. Though, this was towards the end of the day... and like I said.. no coffee ... so 99% sure of that.


The Way Ahead

---

I'm not too worried about retaking the exam. It took some getting used to the way they ask their questions and the designs (the mindset that the test writers were in) that they are looking for. For the first two scenarios, my designs were pretty awful. It just took some getting used to. After lunch, my diagramming was pretty much spot on and my choices were much better.

I really don't feel the need to study any further. I actually think I have much more knowledge about the operation and interaction of protocols, as well as when and why they are used, than this exam calls for. I will simply go about my business. Working my day job and throwing in a few SRNDs here and there until my next exam date.

My next attempt is scheduled for 22 November in London. It's pretty annoying that you have to wait 3 months between attempts, but it is what it is. They can't generate enough content for the exam to offer it more frequently.

After I passed this exam (how it was supposed to play out), my intention was to get the CCIE SP material from INE. Combining everything I've gone through; work, my CCIE R&S, and the CCDE, I feel that the CCIE SP will be a piece of cake. I don't really feel the need to get another CCIE. At this point, however, I think I would be a fool not to just knock it out. If I were single, I could probably just lab for a few weeks to get my speed up and be good to go. Far from single, so that won't be happening. I am about to order the Nexus 7+ SP from INE. That way, I can walk around with my ear buds in and watch the videos while I'm doing whatever else. Then I'll have the workbook on there to go through at my leisure. I said I would do this after I passed the CCDE, but I'm done studying for the CCDE. I might as well do something productive in the meantime (which will also help a bit for the DE). At some point, I need to get my JNCIP out of the way too.


Final Thought

---

Awesome exam. I love a good challenge, and this one takes the cake. It was a lot of fun. I'm sure it could have been a lot worse for me, but I wasn't stressing one bit when I went in there. I was excited to give it a shot and see what all the hype is about. I'm excited to do it again. I did just want to get this over with, it would have been convenient to pass on the first attempt. Yet, I feel that I did get it over with. Now I know where I stand; It's just a matter of time. Cisco might as well put a number on hold for me.

Zorodzai

@powmia - nice writeup. Am sure you will get it next time round, you really sound like you know your stuff

carterw65

Good attitude brother. I think the most important part is you understand what the deal is. You can go back and kick it in the ____.

Thanks for the write up too.

flashdumper

powmia,
how can I reach you?

powmia

Flash, you just did

I see your name, and have to ask. Are you looking for a ****?

flashdumper

powmia,
Haha, I can tell u the story how that nick was made like 15years ago, and I didn't know the english well...
I was working with cell-phones, and there were some software, one of them named as flasher, another one was dumper. So here is my nick came from

Some other ppl and myself in the group are preparing for CCDE and taking our attempts in Nov,22. One of us failed his 1st attempt today.
I think we have the same goals and can discuss this stuff. I just read ur post and thought that you would be interested in that.

So my answer is NO, and I'm asking you again, Is there any way to speak privately with you?

powmia

Well, I guess I don't get an opportunity to provide the good folks here with some entertainment then

I'll PM you my gmail if you change your settings to allow it.

flashdumper

Not sure If i'm able to PM or get one as I'm not a common person on that forum, but u can reach me at the same prefix on gmaildotcom.
same nickname in skype.

flashdumper

powmia wrote: »

Well, I guess I don't get an opportunity to provide the good folks here with some entertainment then

I'll PM you my gmail if you change your settings to allow it.

OK, I see u didn't contact me. I've enabled the privacy setting so u can PM me.

Pseudo_node

Since you are now just casually studying until the next attempt, hopefully you will still update the thread with interesting tidbits and design thoughts.

powmia

It won't be as frequent, but I'll make an effort to keep this thread going. I may have to put a header on some topics that says ### Not Relevant to CCDE ###

such as...

### Not Relevant to CCDE ###

I ordered my Nexus 7+ SP from INE for the SPv3.

Has anyone used INE's SPv3 workbook, then gone on to take the practical? If so, how accurate (in regards to length and difficulty) are the full labs in the workbook? Looking at them, I could probably finish most of them in 1-2 hours. I was expecting to see something that looked more like IPExpert's R&S workbook, in which all of the higher difficulty labs legitimately take 6-9 hours to complete.

### Back to the relevance ###

I'm going to sit down today and write down / draw out everything I can remember from my CCDE attempt before I start forgetting. I am confident that I know exactly where I made my mistakes. Maybe looking at them down the road will help me recall the approach I need to take in my next attempt.

powmia

I think this would go well on the back of a CCDE t-shirt.

Mrock4

That's awesome!

Pseudo_node

Agree...that is awesome!! Also dont forget the old design adage: fast, reliable and cheap--you usually can only have two of these items in a design, but generally not all three!

powmia

Pseudo_node wrote: »

Agree...that is awesome!! Also dont forget the old design adage: fast, reliable and cheap--you usually can only have two of these items in a design, but generally not all three!

I thought MPLS-TE was perfect, and everyone uses it

powmia

flashdumper

Ye, still this interlink is a single point of failure...

powmia

Full routes from two providers. It would take a failure of the interconnect in conjunction with the failure of one provider to send a route to a particular destination. That's not a single point of failure.

Correct me if I'm wrong

flashdumper

powmia wrote: »

Full routes from two providers. It would take a failure of the interconnect in conjunction with the failure of one provider to send a route to a particular destination. That's not a single point of failure.

Correct me if I'm wrong

Ye, I ment that. It's a natural thing to lose some destination prefixes on ISP side due to different reasons and it doesn't depend on you.
So, a single point of failure from Customer point of view.

P.S: Get in the skype, we have to discuss a lot of stuff

powmia

lol, I'm working on it! I can't skype from work and the home life is what it is I did at least install skype on my laptop last night... progress!

In the real world, yes, if those edge routers were in close proximity, I would run a second line (on different line cards if possible). In a situation like the diagrams above... I would be hesitant to change my distro switches from receiving only a default to joining my iBGP and getting full routes.

What are the possible failure scenarios here, though? A better way to look at it, would be; where might you be sending traffic to? Three scenarios:

1. You're sending to a network that is single-homed to (or only reachable through) one of your providers.
2. You're sending to a network that is dual-homed to (or reachable through) both of your providers
3. You're sending to a network that is upstream of both of your providers.

Failure scenario for case 1: doesn't matter if you send traffic out the other link, you aren't reaching them either way.

Failure scenario for case 2: the destination lost their connection to one of your providers. I would assume that this provider would either be peering with the other provider, or have transit. In this case, you can still send traffic out either link. The only thing that would change as far as NLRI would be an increase in AS_PATH from the provider with the dropped customer.

Failure scenario for case 3: irrelevant, I would say that this couldn't be considered a single point of failure.

The most likely traffic affecting situation would not be that you lost an individual route, but that you lost all routes (ie, connection to one provider dropped). In this case, you aren't worried about having redundancy between the two edge routers, because you probably aren't concerned about constantly having to send traffic across that link to send to individual destinations.

Technically, the best solution in some cases could be to just not use a null0 route, but to point the route to the provider's interface. That way the advertisement of the default route can be conditional, and one edge router (and the rest of your network) could rely on the OSPF default from the other... no iBGP required. Thoughts?

flashdumper

What an effort you've made!! Installed a program! Champagne to that speedy fellow!

In the real life, i'd rather keep that null route and iBGP with interlink.
Coz you still need last two, in case if some routes are gone from one of the ISPs u can get out through another one.

The exam is really real life oriented.

In skype it isn't necessary to speak, we can also chat there

powmia

Can't send packets from the Skype application out our boundary at work. Patience

Yeah, the exam is based on real life, as in you get customers that do dumb things. I saw a number of things on the exam that I would never see in a working production environment.

flashdumper

oouch... mr, CCDE candidate can't use skype at work! Your employee shd be rly rough and tough!

Ye, a number of things u'll never have in ur production and that's the thing! u shd analyze that, having the proper background and mindset.

powmia

powmia wrote: »

Can't send packets from the Skype application out our boundary at work. Patience

Yeah, the exam is based on real life, as in you get customers that do dumb things. I saw a number of things on the exam that I would never see in a working production environment.

To add to this.. and to agree w/ flashdumper. This is exactly what the CCDE is all about. Anyone can memorize best practices. Throughout your career; sure, you'll see poor designs, but you'll also correlate the proper ones with the documented best practices. Eventually, it's easy to get the idea that "this is how it's done, because it's a best practice..." without understanding why. The exam is all about understanding not only why something is done that way, but why it isn't done other ways. You need to understand the other methods of designing a technology, what their pros and cons are, and how you can modify them accordingly to any situation... basically, you need to understand the technology, not just pull out the cookie cutter.

Going back to the iBGP example above. The exam is based on real life, and in real life.. those edge routers might be 1000km apart and you have a customer that tells you they don't want a direct circuit between them, for whatever reason (usually a dumb reason, and that's what you'll see in the exam).