Subject: TCP/IP Exam

I know you usuall have only positive examples here, but here's my contribution anyway.I hate to admit it, but I bombed the TCP/IP Cert test. Hopefully I can help others, and help myself as well, by listing the problem areas I had.

For preparation, I did *not* have any Transcender material available. I did use my Microsoft supplied course material ("Internetworking with Microsoft TCP/IP on Microsoft Windows NT 4.0"), as well as the "Workstation Resource Kit" and "Running Microsoft Windows NT Server 4.0". For practice tests, I had two versions on the Microsoft PEP test, one current and one for NT Version 3.5-3.51. I had just completed the class for this portion of the program and had all my notes from class.

Comments on PEP test:

The older test was much more complex than the newer PEP test. Ultimately, it   was closer to the real thing, but woefully inadequate. Both tests emphasized   areas barely covered on the test I took. See below for details.

Comments on Class:

My instructor was terrific. None of my problem is his fault. I would not think   of attempting this class as a self paced study. I did learn many things about   subnetting that are just not covered in the Microsoft text books or are   glossed over that apply to real world operations.

The Official Microsoft test:

Nothing prepared me for the number scenarion questions presented! Iwas simply   overwhelmed by the magnitude of the problems as well as the sheer volume of   them. Easily a quarter of my test dealt with situation scenarios. Some were   "progressive" solutions and were obvious. Q1 was where nothing in the proposed   solution would work and the situation was progressively fixed untill all   conditions were met.

Forget all about NT... know Unix. The majority of my troubleshooting problems   revolved around a Unix box not behaving properly. You will need to know Unix   commands, not TCP/IP commands, for some of these questions. This is one area   my particular class was weak in. Except for printing scenarios, we concentrated   on NT and virtually ignored Unix to NT / NT to Unix problems.

Only three on my 59 questions had anything to do with how many subnets / hosts   are available. This is in direct contrast to my study materials where every   third question revolved around number of hosts / subnets were available.   One of my three test questions was asked to different ways: one question was   which of the four mask choices gave the greatest number of subnets. The next   time the question was which of the four masks gave the grestest number of   hosts. The third question was a growth secenario: your company has two subnets   now and expects to add six more in the near future and you need to select an   approriate mask that maximizes the number of hosts per subnet (255.255.240.0).

There were no questions relating to who was on your subnet given a network   address and subnet mask. There was, however, one question where you were   assigned a class B address, but your network address (164.100.19.0) indicated a   custom subnet was used. You were told you could not communicate with the rest   of the subnet. A dialog box from ipconfig /all showed the standard   class B mask (255.255.0.0). From the list of choices of why you could not communicate, you needed to select "wrong subnet mask".

There were no binary to decimal conversion questions nor decimal to binary conversion questions.

Know all trouble shooting in reverse: the study materials all dealt with   being able to ping by IP address, but not by name. The real test questions were   all ping by name, but not by IP. By this time I was so rattled from the   scenario questions, I was was worried about timeing, I froze and could not work   the problems backwards.

I don't know if my exam was a fluke or not, but no one else I know that's taken this exam has taken one with this format. Most everyone else's exams have followed the PEP tests pretty close. Hopefully, this will help someone else avoid my downfall.

Mitch