Impressions/Generalizations/Misguided Observations:

• Supposedly the failure rate for first-time Essentials test-takers is above fifty percent, according to rumor. That should make you feel better, huh?

• Test is tough. I studied the SSK (MS Press Self-Study Kit, Second Edition) book cover to cover, then went back and created a summary document of it totaling over a hundred pages (yes, I am anal). This took two months and most of my free weekends, which in the springtime was killing me, homey. Then, I did a well-known set of practice exams, scoring 100% on every one. The one thing I didn’t do the first time was go over brain dumps, which I now realize is critical. They help to prepare you for the 6-10 questions you’ll get on the exam which aren’t covered in the SSK or practice exams. In my case this would have made the difference between passing and failing the first exam.

• The second go-through of the actual exam was easier than the first, though many questions were the same.

• MS deliberately muddies the water on several questions, using vague or misleading language, making the exam that much more difficult. I tell ya, I love those guys.

• Not much room for error what with the high passing grade. MS is undoubtably trying to weed out the driftwood (whine, whine).

• Time was not a problem; it took me around forty minutes the second time, a little over an hour the first.

• The "well known" practice exams are good, but just because you are literally perfect at them, as I was, doesn’t mean you’ll pass, which was a big misconception on my part. And, although they do help, it’s unfortunate that the issuing company feels compelled to charge such an outrageous fee for them (once again, whine, whine). If they were reasonably priced, people would be much more apt to buy them legally instead of them being passed around from student to student at training centers like a frat party bong. And yes, as a finance major graduate at college, I’m aware of the whole price vs. demand curve thing, but I stand by my opinion!

• I detest the fact that MS doesn’t tell you which ones you missed. All they tell you is the vague "areas" (standards/terminology, planning, implementation, or troubleshooting) you missed questions on - how very helpful. C’mon guys, we’re not asking for the answers, just the questions we missed so we can research the answers ourselves!

General Hints

• BRAIN DUMPS, BRAIN DUMPS, BRAIN DUMPS - they made the difference in passing or failing in my case, because the practice exams and SSK just don’t cover everything on the test. Read every dump at coverware.com if you have to, it’s time well spent.

• Relax. What’s a $100,. anyway? If you think about it, you shouldn’t be nervous about it. You can always take it again, almost immediately (I retook it four days later).

• Mark the difficult quesitons and go back to them. If you get hit with hard ones at the beginning, not only will you spend too much time on them (in your mind anyway) and feel rushed for the rest of the exam, but you’ll also be demoralized, which can ruin your concentration.

• There’s an excellent site that has a 100-question+ practice exam, http://frontpage.idsonline.com/sallard/. Only, the author, who is to be otherwise commended for compiling this helpful list of questions, seems to have a moral problem with publishing the answers. Most people will (or at least should) use the answers as a gauge to what they should go back and concentrate on. There’s enough info in the book, practice tests, and brain dumps to digest without having to go through and hunt down answers to other questions, in my opinion. But, thanks to the author, his questions were helpful.

• More practice exams (four sets of 29, I believe) are available at http://www.visualmcp.com/ne/exam/welcome.asp?set_num=1. These are quite similar to some of the actual questions like the ones above and the practice exams but are different in that it’s interactive; you’re graded at the end. Nice service, though I didn’t take them all.

Test-Specific Hints

• If in doubt regarding a WAN connectivity question, choose ATM as your answer UNLESS it also lists "bandwidth as needed" or "connection at any speed" as answers - then it’s FRAME RELAY. ATM is usually fixed speed (155 Mbps), though it can go higher, according to the SSK.

• If in doubt regarding a network troubleshooting questions, choose Protocol Analyzer as your answer UNLESS they say over a long period of time, in which case they’re looking for Performance Monitor logging as the answer.

• MANY questions on Connectivity Devices (routers, gateways, and bridges especially), so review the SSK, brain dumps, and practice exams thoroughly.

• Not one question on wireless communication (infrared, etc. ad nauseum). Blow these off. Hopefully some MS geek won’t respond to this statement by adding a bunch of questions of these.

• Know your IRQs! Don’t bother with the Base I/O Port addresses.

• Know your cabling types, in particular that UTP Cat. 5 supports 100 Mbps.

• Know your Troubleshooting Tools. Though the book has perhaps only ten pages on these, there are an inordinate number of test questions on this topic.

Key Areas That The Book Doesn’t Cover (KNOW THESE)

• NDIS/ODI - a standard which allows network adaptor cards (NACs) to work with multiple protocols. Another way to put this that the NDIS/ODI-compliant cards are protocol-independent, or that cards complying with NDIS/ODI free the NAC developers from having to understand APIs in the MAC sublayer of the Data Link Layer. THERE WERE TWO QUESTIONS PERTAINING TO THIS ON THE TEST.

• Frame Types - you may have the incorrect frame type setting on your NAC if: 1) your computer initializes correctly (meaning IRQ and Base I/O Port settings are fine), 2) your network is running NWLINK IPX, and 3) nobody else on network is having problems. BOTH TESTS HAD AT LEAST ONE QUESTION REGARDING THIS

• Transceiver Settings - if you face a question similar to the one above, but NWLINK IPX is not included in the question AND frame type is not a possible answer, choose "incorrect transceiver settings" as your answer.

• Protocol Mismatch - if one type of computer can’t access network (e.g. Unix machine) but others types (e.g. NT, DOS) can, then you have a protocol mismatch occurring between the non-connecting computer and server. FIRST TEST, ONE QUESTION

• Token Ring vs. Ethernet Frame Sizes - if it takes longer to copy a file from a token ring segment to an Ethernet one than vice versa, it’s because the TR frame is larger and must be broken down into several "chunks" to fit into the physically smaller Ethernet frame. ONE QUESTION ON SECOND TEST. Note: There has been some disagreement on web sites regarding the correct answer to this, but the vast majority of contributors listed the above answer as the correct one.

• VOMs - resistance reading for properly-working terminator should be 50 ohms and infinity for both the cable and tee connector. A reading of zero on any one of these means it’s bad. For instance, a reading of zero impedance on a cable between the jacket and core means it’s got a short in it (signals will jump from core to jacket). BOTH TESTS HAD AT LEAST ONE QUESTION ON THIS. Note: Research this further, I have never actually used one of these bad boys.

• TCP/IP Window Size - reducing this will not reduce network utilization at all. BOTH TESTS, ONE QUESTION.

• Packet Size Reduction vs. Network Utilization - ditto to above.

• Protocol Mismatch - if one type of computer can’t access network (e.g. Unix machine) but others types (e.g. NT, DOS) can, then you have a protocol mismatch occurring between the non-connecting computer and server. FIRST TEST, ONE QUESTION

• Performance Statistics - when CPU usage is low (less than say 10%), Page Faults per sec. is low (fewer than say ten or so) but Disk Access is up around 25% or so, then your problem is slow read/write performance. How to fix? Implement disk striping with parity. BOTH TESTS, ONE QUESTION

• RG-58 /U cabling is not compatible with the 802.3 10Base2 Ethernet specification. Translation: if you try to use this with "regular" Ethernet coax, RG-58 A/U, you’ll have problems, so replace the /U cable. BOTH TESTS, ONE QUESTION

• Mesh Topology - if you see several computers with links between each and every one (some contributors liken it to the Star of David) you have a "mesh" network. They are the most fault tolerant of all configurations because of all the redundancy but they are difficult to reconfigure once set up (i.e. adding another computer). FIRST TEST, ONE QUESTION

• IPX/SPX - next best choice for small peer-to-peer setup if NetBEUI not listed as possible
• answer. SECOND TEST, ONE QUESTION.

• V.35 Modem Standard - know this formula: Connection speed equals modem speed divided by eight, multiplied by 100, then multiplied by number of modems. So, for eight 28.8 modems, you’d have 28.8/8 x 100 x 8, or 2880 Kbps, or 2.88 Mbps. SECOND TEST, ONE QUESTION.

• Subnet Mask - an example of weird wording: they define this as the "number sequence that specifies how IP addresses are interpreted." BOTH TESTS, ONE QUESTION

• Hexadecimal Numbering Scheme - although using hexadecimal numbers (0-9, A-F) will work as NetBIOS computer names, it’s kinda silly when you can use more descriptive alphabetical names (e.g. A1445 vs. AppServer). But, you could do this if you felt so inclined. MS just throws this in there as a confusion factor (God bless ‘em). Just remember that you’re limited to fifteen characters; the sixteenth character is reserved. FIRST TEST, TWO "SUPPOSE THAT..." (SCENARIO-BASED) TYPE QUESTIONS.

• UTP Category 3 - the minimum UTP that complies with 802.3 Ethernet 10BaseT standard. I THINK THIS WAS ON AT LEAST ONE TEST.

Miscellaneous

• There was a question regarding the proper type of cabling to use in a situation where there was a meter of ceiling space and requirement for 100 Mbps, I believe. The only reason I could think of as to why they’d mention the ceiling space is if they were trying to infer that the ceiling was being used as a return plenum (i.e. one big open return duct to mechanical unit, not uncommon in commercial or office sites). In that case, you’d have to use plenum-grade cabling (just like SSK says), but they didn’t offer that choice. They just said UTP, STP, Fiber Optic, or 10Base5 (I think). I guessed UTP, but it may have been wrong (but of course I’ll never know, now will I). I’ve seen this question in other brain dumps, but I didn’t get a consensus as to the answer.

• There was another question which asked for the likely solution to the problem where your computer fails to boot up after you install a NAC. The computer has an existing SCSI adaptor and hard drive. I guessed "UMB conflict" as the answer, but it may have been "IRQ conflict" instead. UMB was listed as the answer in one of the practice exams, however, so once again I don’t know if I even got this wrong.

Anyway, hope this helps. If I had read this document before I had taken the test the first time, I would have passed it then. So, good luck to you. Many thanks to Brian Shafrik, creator of one of the best MCSE brain dump I’ve seen, at www.coveware.com.