These six products offer practice questions for the NT Server in the Enterprise exam. But as our expert reviewer shows, if you don't already know your stuff, you could be led astray.

Practice Makes Perfect— Or Does It?

These six products offer practice questions for the NT Server in the Enterprise exam. But as our expert reviewer shows, if you don't already know your stuff, you could be led astray.

The Windows NT Server in the Enterprise exam is dreaded among MCSE wannabes as the make-it-or-break-it test. Along with the Microsoft NT 4.0 TCP/IP exam, 70-068 produces the most “open-sweat-glands per second” in the testing center (as measured by the Performance Monitor object: ShakyMCSETester) than all the other exams in the premium track. This reputation is well deserved; the test is inclusive of all of the core knowledge one had to acquire in order to pass the NT Workstation and Server exams, and then some.

Reviewed

Enterprise complicates the test-taker’s life with the added nuances of trust relationships, cross domain administration and resource access, and even the basics of knowing how to set up IIS Virtual Servers. Although the occasional miscellaneous question such as the IIS one shouldn’t break you, there’s a core set of exam objectives (mostly related to multiple domain scenarios, services, and utilities) that, if not completely understood, will be your downfall. In this article I review six self-study products that purport to help you pass the Enterprise exam. My assignment: to assess how well the guides help you acquire and reinforce this core knowledge.

The products fall into two major categories. First, I review three books that are a combination of “concept introduction”/sample tests. The focus of these books is to introduce concepts to the reader and then immediately test that knowledge with a mini-exam. Then I look at three others that fall in the category of paper-based testing where the bulk of the content is dedicated to practice exams followed by answer explanations. These books are primarily used to supplement other knowledge-building books as a measure of preparedness for the real Enterprise exam. I’ve reviewed them for technical accuracy and for the structure and adequacy of question content.

MCSE The Core Exams in a Nutshell, By Michael Moncur
O’Reilly & Associates; www.oreilly.com
ISBN 1565923766, $24.95, 1998
51 pages of the 432-page book cover the Enterprise exam

I’ve never been a fan of cram books that provide the essentials of a test without really explaining the “why” of how things work. For that same reason, I flunked college algebra my first semester in college—I just couldn’t see the practical applications of math. My second semester I passed algebra, not because I suddenly became mathematically enlightened but because I learned to play the game—I memorized the formulas and set steps for resolving algebraic problems, and barely squeaked by.

If your goal is to memorize facts and figures without knowing the “why,” then this book is for you. Short on explanation, this book delivers what any respectable Internet “braindump” site would afford you at no cost.

Although technical mistakes were few, some were glaring. For example, when discussing the use of trust passwords for establishing a trust relationship, the author incorrectly asserts that the trust can’t be disconnected without the use of this password.

In another part of the book, when discussing WINS, the author asserts, “You do not need to configure a database to use the WINS Server. The server maintains a database of NETBIOS names and IP addresses automatically by receiving broadcasts from clients.”

I took this to mean that WINS clients don’t need to be configured to talk to the WINS server but will instead fill the WINS database automatically via broadcasts, something that just isn’t true. Or maybe the author meant something else.

What clinched the technical rating at the bottom of the scale, however, was the paragraph that discussed “Master Domains.” The author states that, “If you use an adequate number of domain controllers per domain, each master domain can support a maximum of 20,000 users. Thus, a network with three master domains could support 60,000 users.” I’m not sure if the author was speaking “practically” when giving these figures, but there’s nothing practical about the Enterprise exam. A master domain within the definition of the Single or Multiple Master Domain model can contain a maximum of 40,000 user accounts per Master Domain, something that you can bet is going to be a testing sticky point. For example, if a test question presented a scenario of the XYZ company having 60,000 employees and wanted to know the minimum number of Master Domains that would be needed, the answer would be two, not three as the book asserts. What’s even more confusing, however, is that in a later section labeled, “Highlighter’s Index,” (in essence, a braindump) the correct figures are given for the maximum number of users per Master Domain.

Many testable subjects don’t show up at all or are briefly touched upon. For example, very little information was given on optimizing the directory synchronization process (PDC-BDC), no information on how the directory replication service works (export-import computers) other than a blurb in a different part of the book, and no mention of the utilities needed to examine or interpret core memory dumps.

At the end of the Enterprise chapter you can take one practice exam consisting of 51 questions. These questions are typically one- or two-liners. The answer section provides the correct answer but rarely explains why this is the correct answer and only occasionally explains why the wrong answers don’t fit. Some of the questions are ambiguous and can trip up the novice test taker.

Sample Question

You are managing a network using a single master domain. The CENTRAL domain is the master. EAST and WEST domains contain resources, and are physically located across WAN links. Users in the EAST and WEST domains are complaining about slow logon times. You must choose a solution to improve logon times.

Required Result: Improve user logon time for EAST and WEST.
Optional Result: Reduce WAN traffic required for logons.
Optional Result: Improve user logon time for the CENTRAL domain.

Solution: Install two BDCs for the CENTRAL domain: one in the same physical location as the EAST domain and one in the location of the WEST domain.

  1. The solution meets the required result and both of the optional results.
  2. The solution meets the required result and only one of the optional results.
  3. The solution meets the required result only.
  4. The solution does not meet the required result.

My answer was A and the book’s was B. If we break down the question into its essential parts we come up with the main problem—logons are too slow for users who work in the EAST and WEST sites. In a single Master Domain model (an essential piece of information to answer this question) all of the user accounts reside in the Master Domain (CENTRAL) whereas the computers that EAST and WEST users physically log on at are in their respective domains. When an EAST user goes to log on, even though he or she may be physically sitting in the EAST location, the user’s account is in the CENTRAL domain database, and the logon will have to be validated by a CENTRAL domain controller. The problem lies in the fact that all of the CENTRAL domain controllers are physically sitting across the WAN link, and all logon validation traffic is subsequently also being transmitted across the WAN link—hence the slow logons.

By physically placing a CENTRAL BDC in the EAST and WEST physical locations, the logon validation by a CENTRAL domain controller can be done locally rather than through the WAN link. This may sound confusing, but essentially what we’re doing is physically relocating domain controllers without changing their current domain memberships.

Think of it the same way that U.S. voting is accomplished. Let’s say you’re a resident of the state of Colorado and happen to be on vacation in Hawaii, but you still want to vote in your home-state election—something that can be accomplished with an absentee ballot. You don’t have to travel all the way back home to vote. The means by which to vote in Colorado has been physically relocated to Hawaii so you can accomplish the desired task.

So, it’s obvious that the required and first optional results are met since logon traffic is no longer being sent across the WAN link. What isn’t so obvious is whether the second optional result has been met—improving user logon time for the CENTRAL domain. What’s really being asked here is, if you’re physically sitting at a computer in the CENTRAL site, will your logons be any faster? This is ambiguous and open to much discussion. One could argue that there was no problem to begin with for the CENTRAL-based users because they aren’t situated across the WAN link from the domain controllers (the book answer). Or one could argue that if the EAST and WEST users are enjoying faster logons due to the now local CENTRAL BDCs, this is placing less strain on the domain controllers that physically reside in the CENTRAL domain, which should improve logon response for CENTRAL-based users (my reasoning). Either way, you spend time having to read too little or too much into the question.

Although the real exam may have similar ambiguous questions, test preparation tools don’t need to introduce these. If your intention is to hone your test-taking skills, there are other tools out there for that purpose. If your intention is to acquire and retain knowledge (the reason most folks are reading this article) for exam preparedness, then you don’t need to be needlessly confused by poorly worded questions such as this one.

In summary, this book doesn’t accomplish its intended goal of being a “comprehensive study guide that covers the required exams for MCSE certification.” Let’s face it. Cramming is something you may get away with in college, but only if you have all the right facts and figures to jam into that cranium of yours. Core Exams in a Nutshell isn’t exactly all it’s cracked up to be.


MCSE NT Server 4.0 in the Enterprise Ace It!,
By Mark B. Cooper
IDG Books Worldwide; www.idgbooks.com 
ISBN 0-76453-265-0, $24.99, 1999, 466 pages

MCSE Ace It! is touted as a “supplement to other certification resources” and rightly so; it’s not comprehensive enough to serve as your primary preparation tool for passing the Enterprise exam.

I found the content to be primarily NT Workstation and NT Server exam-oriented, and not enough content was related to Enterprise concepts. Sadly, this seems to be the trend in these types of books—fill ‘em up with pages and pages of duplicated content from other books in the series by the same publisher. Although this makes for great review material, you can end up spending way too much time on information that, at this point in your MCSE certification track (having already passed NT Workstation and Server, I presume), you should already have down cold.

Case in point: The author dedicates the first six chapters to concepts like configuring disks, system configuration, managing users and groups, user profiles and system policies, managing printing, client-based server tools and services for Macintosh—all Workstation/Server concepts. Then finally in chapter seven the first Enterprise concept—Windows NT Directory Services—is introduced. Now, if only the rest of the book would continue Enterprise concepts, but alas, that’s not to be. Chapter eight through the end of the book is back to Workstation/Server—sharing and securing file systems, NetWare coexistence and migration, networking with TCP/IP, RAS, optimizing performance, and advanced troubleshooting.

Don’t get me wrong. I did find tidbits of information here and there that pertain to Enterprise concepts, but who has time to wade through 466 pages to try and glean this information?

In the preface, the author states, “I assume that you are familiar with NT Server 4.0 in the Enterprise Concepts, or that you have other study material to help you understand these concepts. With that in mind, I cover NT Server 4.0 in the Enterprise to the depth of the exam objectives, but not beyond that point.” I don’t think he ever gets to “that point” in this book.

What’s missing? A lot. There’s no information on optimizing the directory synchronization process (PDC-BDC), i.e. no Pulse, Pulse Concurrency, or Replication Governor explanations. IIS isn’t touched upon in terms of setting up and supporting IIS virtual servers, but IIS authentication is introduced. The biggest complaint I have is the NT Directory Services discussion, which falls into the pattern of many other books advising the reader to “rote” memorize trust relationships (arrows pointing to and from something) and choice of domain models without discussing these concepts in-depth. I’m of the opinion that 466 pages could be better spent on Enterprise-specific concepts. Go figure.

The beginning of each chapter begins with a short “Are you prepared?” exam, typically, three questions long, that tests your need to read the subsequent chapter. Since the greater part of the book deals with NT Workstation/Server concepts, you could be misled into thinking that you were prepared for the Enterprise exam if you were already familiar with these concepts. In fact, an MCP who had already taken the Workstation and Server exams would have no problem answering all the questions at the beginning of each chapter with the exception of Chapter 7, which covers NT Directory Services.

At the end of each chapter is a “Have you mastered?” mini-exam, typically, five questions long. This is an appropriate tool for measuring knowledge acquired from that chapter. Each question’s correct and sometimes incorrect answers are adequately explained.

At the end of the book is a 50-question practice exam followed by an answer chart and exam analysis section. This exam is full of ambiguous questions, just plain bad information, and questions that test you on concepts that the book itself fails to provide content on. My advice is to not use this exam.

Sidebars and Additional Info

You are configuring a Windows NT Server for a high-speed database application. You want to configure the hard disk to maximize disk access speed and to provide a large data volume to the application. You plan on configuring the database application to record transaction logs to a fault tolerant disk set for emergency purposes. What would be the best disk configuration for the application?

  1. Create a data volume using a stripe set with parity.
  2. Create a data volume using a stripe set.
  3. Create a data volume using a volume set.
  4. Create a data volume using a mirror set.

My hesitant answers were A and D; the book’s answer was C.

There are a couple of things wrong with this question. Number one, it’s not very clear how many hard drives are present. The question mentions “the hard disk” but then also mentions the need for a “fault tolerant disk set.” These are at odds with each other. NT supports software-based levels of RAID that are fault-tolerant disk configurations—mirroring and striping with parity. A mirror set requires two separate hard drives and stripe sets with parity at least three. The only fault tolerant methods are A and D and—despite the ambiguity of the question—would be the ones I would have chosen had this been a real exam question.

Now for the book’s explanation for their answer C: “The fast data volume that can be created in Windows NT Server is a volume set. Volume sets provide higher performance than stripe sets because there is no overhead associated with creating fault tolerance information or striping across multiple drives.” This is completely wrong! Volume sets offer no performance increase. They were designed simply to let a user associate more space on a single or multiple hard drives with a single drive letter than that space normally allowed by disk partitioning itself, or to extend the available space to an existing partition. This kind of bad information raises a warning flag as to the technical soundness of the book.

Having said all that, I do believe that the formatting of the book is well done. I liked the little “test tips” and “test traps” that show up throughout the book. The “pop quizzes” and post-chapter tests were all beneficial in reinforcing the concepts just learned. The technical accuracy is questionable though. With the right content and technical editing, these Ace It! books have potential.

Another Sample Question
You are the administrator of a Windows NT Server computer on your network. You want to document the complete TCP/IP configuration on the computer. What can you do to get a detailed view of the configuration?
  1. Use Network in the control panel to view the TCP/IP protocol configuration.
  2. Use the ipconfig /all command
  3. Use the Internet Service Manager to view your TCP/IP configuration
  4. Use the ping /all command.

My answer was A and B; the book’s was only B.

If you’re using the TCP/IP protocol on an NT machine and want to get detailed information on how the protocol is configured, then go to the source of the configuration—the Network applet in Control Panel. The IPCONFIG utility with the /ALL switch will give you detailed information as well.

Answer C is incorrect because the Internet Service Manager is the tool that you use to manage Internet Information Server (IIS) and doesn’t display IP protocol-specific configurations. Answer D is incorrect because Ping is used to verify connectivity. Although you can ping the loopback address 127.0.0.1 to verify that the TCP/IP protocol initialized properly, it won’t display the configuration.


MCSE Test Success: NT Server 4 in the Enterprise
By Lisa Donald, Sybex Network Press; www.sybex.com 
ISBN 0-78212-147-0, $24.99, 1998, 443 pages

The book begins with a discussion of NT Directory Services, an Enterprise concept—but like its competitors reviewed here, most of the content is focused on NT Workstation/Server concepts. It does cover Enterprise-specific concepts missed by the others, such as optimizing directory synchronization and IIS virtual servers, but falls short in other areas such as Memory Dump utilities and advanced directory replication concepts.

Content aside, what I liked about this book were the clear and concise presentation of topics within the chapters and the well-designed study questions and practice exams at the end of each chapter. The 352 study questions and 132 sample test questions scattered throughout the book are especially helpful in reinforcing concepts learned in each chapter. At the end of the book is a 55-question final exam.

The study questions were fill-in-the blank, true-false, and draw-in-relationships types, whereas the sample tests were traditional multiple choice. Although the practice exam questions were simple one- or two-liners, they accomplish the goal of concept reinforcement. The final exam at the end of the book was designed more as a measure of exam readiness, but it’s disappointing in that the book only offers an answer key without explaining the answers.

Sample Question

You have just received a network printer that you must configure. The printer will use the DLC protocol. What must you configure? Choose all that apply.

  1. The DLC protocol must be added to the printer’s print server.
  2. The DLC protocol must be installed on all network clients that will use the printer.
  3. The DLC address of the network printer must be specified in the printer’s configuration.
  4. The fully qualified domain name of the network printer must be specified in the printer’s configuration.
  5. The hardware address of the network printer must be specified in the printer’s configuration.

My answers were A, B, and E. The book’s answers were A and E.

Now here’s another question that’s open to interpretation. The central question is, “What must you configure?” and doesn’t specifically state what components are being referred to.

So I started off by analyzing the answers one by one. Answer A is true for any NT box that wants to print directly to a print device. In the Microsoft world, a print device is the term used to describe the physical printer, whereas the term printer refers to the software definition that describes the print device (location, print driver, paper trays, etc.). To print to a DLC print device, you need to define in software the where, when, and how of the print device. The computer that stores this printer definition is called the print server.

A DLC print device such as an HP JetDirect card is a hardware print server. To send a print job directly to it from a computer on the network, you need to have the DLC protocol installed on your machine, and you need to know the hardware address of the JetDirect card. So answer E is a given.

Answer B states that the DLC protocol must be installed on all network clients that want to use the printer. This is true if the desired result is for these clients to print directly to the DLC device. If you don’t want to go through the hassle of installing the DLC protocol on every client machine that needs access, you can accomplish the same result by installing the DLC protocol on one NT machine. On this one machine you should then configure the appropriate printer definition (which makes this box the print server), and then share this printer on the network. Any computer that can access the share can print by sending its job to the print server, and the print server will then forward the job using the DLC protocol to the JetDirect card. This is also known as setting up a print gateway.

Imagine that you’re at the finish line for the Tour de France. The winner, Lance Armstrong, comes sailing across, and there’s a line of people waiting to talk to him. Now imagine that there will be no onsite interpreter, so the only people who can speak with him are those who understand and speak English. That would exclude everybody else. This is like trying to print to a JetDirect card directly. You must speak the same language.

Now if he had an on-site interpreter, you could speak through the interpreter who understands both languages. This is like using a print server as the go-between your computer that doesn’t understand the DLC protocol and the JetDirect card that only understands the DLC protocol. Got that?

Answer C isn’t correct because a DLC address is the same as the MAC address and should be called that.

Answer D is incorrect because fully qualified domain names are a part of the TCP/IP protocol but have no use in the DLC world.

In summary, this book is well written with a lot of good reinforcement questions. Of the three books that fall into its category, I would recommend MCSE Test Success as the better of the three.

Another Sample Question
Your NT Server is configured as a BDC. The BDC supports logon authentication for 1,000 computers. The primary function of the server is logon authentication and providing clients with logon scripts. How should the server service be configured for maximum optimization?
  1. Balance
  2. Maximize Throughput for Domain Controllers
  3. Maximize Throughput for File Sharing
  4. Maximize Throughput for Network Applications

My answer was D, and the book’s was C.

First, you must understand where you need to go to configure these options. If you open up the Network applet in the Control Panel and click on the Services tab, you’ll notice the Server service in the list. The Server service is what allows your computer to create shared resources on the network. So whenever you share out a folder or printer, it’s the Server service that makes this possible.

Because the Server service is sharing out files and printers, this makes your box a File and Print Server. Typically, a computer that’s a File and Print Server likes to have memory set aside for the purpose of file caching. File caching lets the server retrieve initial data requested by a client from the hard drive and then place that data in memory, so that the next user requesting the same data can have the request quickly satisfied. This is accomplished by the server being able to retrieve the data quickly from memory (where it was cached) rather than having to read the hard drive again (slow).

When configuring the server service, you’re offered one of four settings. Your choice here adjusts the amount of memory set aside for file caching. The first three settings—Minimize Memory Used, Balance, and Maximize Throughput for File Sharing—are related to using the machine primarily for File and Print Serving functions. You choose between the three settings based on the number of users that you anticipate being concurrently connected to the computer at any given time, less than 10, between 11 and 63, and more than 64, respectively.

The last option, Maximize Throughput for Network Applications, is what you’d select if your computer were primarily running a client/server application like Microsoft SQL Server, SMS, or SNA Server. When you select this option, you’re taking away the memory assigned to file server caching to free it up for use by the client/server application.

So here’s the central question: Is a domain controller considered a file and print server or a client/server application? In this question it’s performing both roles. It’s functioning like a database server by looking up logon authentication requests in the SAM database and returning to the user a yes/no answer to logon. At the same time the users are retrieving login scripts (files) from the domain controller, and this is a file and print server function. So really, both C or D could be right; but I tend to think that the more taxing of the two for server performance would be the SAM database lookups and therefore the function that we want to optimize for.


Test Prep MCSE: Windows NT Server 4 Enterprise, Second Edition, By Walter J. Glenn and Bruce Ford
New Riders; www.newriders.com 
ISBN 0-73570-009-5, $19.99, 1999, 492 Pages

Test Prep MCSE can be placed in the same category as the Coriolis Exam Cram practice tests and the Microsoft Press MCSE Exam Readiness Review; they’re all basically compilations of exam questions. Depending on your experience level, they’re not to be used as the primary tool for preparing for the Enterprise exam but as a supplement to a good self-study manual.

The category into which these tools fall is interesting. The content is structured completely opposite the first category of books I’ve reviewed here. In the first category, the books were structured with knowledge content offered first, followed by a set of questions designed to test that knowledge, whereas the second category offers exam questions first, and then knowledge-based explanations designed to confirm/reinforce/instruct knowledge. Which approach will work better for you? I tend to believe that the first category is good for novice NT users who need to acquire a knowledge base; the second category is more appropriate for experienced NT users. Or, better yet to maximize your test success, use a combination of both methods regardless of your experience level.

The book is divided into 32 different sections, each section having an average of 10 practice questions. After answering the questions in each section, you can check your answers with an answer key, followed by a detailed analysis of the correct/incorrect answers. At the end of each section a “further review” page summarizes the topics that should be studied in more depth. The total number of section practice questions is 403.

At the end of the book are two 55-question practice exams. I found the section questions to be more straightforward and conducive to learning than the practice exam questions, which were ambiguous and elusive. As I mentioned earlier, if you want to beef up test-taking skills, get a specific book on how to accomplish that. When publishing a test preparation aid such as this book, authors should realize that complicated and tricky questions are needless and counterproductive to learning.

Sample Question

After a trust relationship has been established between two domains, why is it possible for users to access resources across domains?

  1. Establishing the trust relationship merges the user account databases from the two domains.
  2. The logon services from the trusted domain pass the resource access request to the trusting domain’s logon services, where the resource access permissions are authenticated.
  3. Duplicate logon IDs are created in the trusting domain’s account database when you assigned permissions.
  4. The trust relationship allows logon services from the trusting domain to access the trusted domain’s account database.

My answer was D, and the book’s was B.

Ah, the dreaded trust question. This strikes more fear in the heart of test takers than all of the summer’s supernatural movie plots combined. Really, once you understand the mechanics of trusts, they’re not that difficult.

Where do we start? Oh, yeah, the trusting and trusted terms. Folks tend to get these confused and all mixed up. Imagine that you’re driving through Mexico and you get stopped by la policía. The cop politely asks for your papers, and you present your California driver’s license. He looks at you suspiciously and then goes back to his car.

You can see him talking on the radio while straining to read your driver’s license, but you’re not sure what he’s saying. Visions of a dark, damp jail cell flash across your mind. He walks back to the car and hands back your license. “Adiós, Señor. Que tenga un buen día,” he says as he motions you to leave.

How did you get out of that one? Well, it turns out that Mexico has a trust relationship with the U.S. and has access to the U.S. driver’s license database. Even though Mexico didn’t issue your driver’s license, they could verify who you were by calling up the U.S. The same thing happens between domains. DOMA (Mexico) trusts DOMB (U.S.), so that DOMB users/groups (U.S. drivers) can get access to DOMA resources (driving Mexican roads). In order for DOMA (Mexico) to verify the identity of trusted users (U.S. drivers), it’s necessary for DOMA Domain Controllers (Mexico Police) to have access to DOMB’s (U.S.) database for verification purposes. Boy, Mexico sure is one trusting country!

Now, back to our question. Answer A is incorrect because there’s no merging of databases that goes on when a trust relationship is established. In fact, a trust relationship is nothing more than an agreement between two domains. The U.S. agreed that Mexico could have access to the U.S. driver’s license database, but it didn’t give them a copy of the database.

Answer B is incorrect because it describes the opposite of what happens in a verification of access to a resource. If B were true, that would be like saying that the U.S. somehow psychically viewed your predicament and initiated the call to the Mexican police to let them know you’re an OK person. It’s Mexico that had to initiate the call because it’s their resource that someone was speeding on.

Answer C is incorrect. Again, the trust is just an agreement. Mexico isn’t going to issue Mexican driver’s licenses automatically to everyone in the U.S.

Answer D is correct. It describes exactly your predicament. And don’t let it happen again.

Let’s be honest here. There’s not a technical book out there immune from either bad technical content or poor technical editing. Some are worse than others. In the fairy tale story of Snow White, all the magic mirror could do was say to the wicked stepmother, “You are the fairest of them all,” because compared to all the other bad apples in the kingdom, it was probably true at the time. It takes a Snow White, the picture of perfection to come around to change that. I’m still waiting for that to happen in the technical world. The MCSE Test Prep is worth the money for the section questions, but I wouldn’t spend too much time on the practice exams.

Another Sample Question
To better manage traffic on your network, you manually configure one of the Windows NT Servers to be the master browser. What should you know about manually configured master browsers?
  1. The Windows NT Server configured to be the master browser will remain the master browser until it is manually reconfigured.
  2. The Windows NT Server cannot be manually configured to act as a master browser.
  3. The Windows NT Server may lose its status as master browser if it is downed or disconnected from the network.
  4. The Windows NT Server must be a domain controller to be a master browser.

This question is a little tricky. Let’s see why. You can manually configure an NT Server to be a master browser by changing some Registry entries. Under HKEY_LOCAL_MACHINE/System/CurrentControlSet/Services/
Browser/Parameters a couple of values can be modified for this purpose:

  • IsDomainMaster=True/False—if set to True, this Registry entry forces this computer to become the master browser for its subnet.
  • MaintainServerList=Yes/No/Auto—if set to Yes, this Registry entry configures the computer to be a “browser,” either the master or a backup for its subnet.

So that’s how this mess got started, someone went in and hacked the Registry with one of these entries.

Answer A is technically correct in that if the Registry change were made manually, to undo the action you’d also need to change the entry manually back to its original value in the Registry.

Answer B is simply not a true statement; we just discussed how this operation can be accomplished through a Registry change. This particular answer brings up an important point: If a test question states that you did something, i.e. make a computer a master browser, promote a BDC to a PDC, take out the garbage on Monday, then the act occurred and was probably a valid and successful act. I don’t believe that you’ll ever see on a Microsoft exam a question that states you did X and then an answer that states that X can’t be done. You just did it.

Answer C is also a correct statement. If the master browser is down for a period of time, someone on the subnet is going to take notice and force an election so that a new master browser takes over. This could be a temporary situation in this case, however, because as soon as the old master browser comes back online, depending on which Registry value we changed, it could get its old job back. But regardless, the statement is true nonetheless.

If Bill Clinton (IsDomainMaster=True) went missing, it wouldn’t take long for someone to notice. Al Gore (MaintainServerList=Yes) would then rush in to take his place. However, if Bill showed up again, he’d get his job back, and Al would be back to playing second string.

Answer D is incorrect for the same reason stated for answer B being incorrect.

So there you have it, two potential answers. If you’re lucky enough to get this question and it asks you to choose two answers or all that apply, no sweat. If the question asks you to choose a single answer, then break out the handkerchief. A or C, which will it be? Well it all hinges on how you interpret answer C—temporary or permanent loss of master browser status. Flip a coin; you have a 50-50 chance.


MCSE Readiness Review Exam 70-068: Windows NT 4.0 Server in the Enterprise
By Dave Perkovich
Microsoft Press; http://mspress.microsoft.com 
ISBN 0-73560-539-4, $29.99, 1999, 421 Pages

Ring, ring, ring! Joe Smith, VP of “New Markets” at Microsoft Corp. picks up the phone. “Joe, Bill here. Get up to my office ASAP!” Click. Joe takes a big gulp, throws recent charts and spreadsheets into his briefcase, and runs across the Redmond campus. Exasperated, he says to the secretary, “Mr. Gates is expecting me.” She hurries him in.

“Joe, look at these headlines,” Bills says as he thrusts the Sunday paper toward his visitor. Scanning the front page, Joe sees the big news: “Technical Certification Hot, Hot, Hot! MCSE in Lead, Training Companies Cashing In.” Joe smiles and says, “That’s great news! So why am I here?” Bill glares at him, “Why aren’t WE making money off this?!” Joe stops in at the Microsoft Press building on the way back to his office.

You knew it was just a matter of time. Microsoft Press was already making a killing off the Microsoft self-study manuals, so why not test preparation material as well? That’s exactly what the MCSE Readiness Review series intends to do. Let’s take a look at how good a job Joe did.

Of the seven products I’m reviewing, only three came with a CD. On the CD that comes with this book, Microsoft bundles an electronic assessment test from Self Test Software. The exam is 60 questions long and a subset of the retail version of Self Test’s NT 4.0 Server in the Enterprise product.

The book recommends taking the electronic assessment exam first and using the results of that test to figure out your weak areas, so you can focus on pertinent sections of the book. However, this review is based on the book content and not Self Test’s software.

I like the format of this book. You’re presented with a multiple-choice question and on the next page an analysis, answer by answer (the same format as this review) of the question. This is effective for two reasons: 1) You can go through the book at your own pace rather than having to answer a lengthy practice exam followed by an equally lengthy review; 2) the answer-by-answer format is helpful for understanding where you may have gone wrong (or right) when answering the question. Again, this type of study aid isn’t sufficient in and of itself—it’s an assessment tool that needs to be supplemented with other study guides.

The book is made up of 32 separate sections, each focused on a separate test objective. The objectives are prefaced with a mini-summary that discusses the technical concepts you’ll be tested on when you take the real exam. And who should know better what’s coming on the exam than the exam creator? At the end of each objective is a “further reading” list referencing none other than Microsoft Press books. No big surprise here.

Each objective has an average of four questions per section with a book grand total of 148 questions. Most of the questions are short and simply worded. Many of the questions are scenario-based. If you’ve ever taken a Microsoft exam, you know what I’m talking about—those ugly little questions that are looking for a proposed solution to meet primary and optional results.

Be careful. I noticed on at least one question that even though I answered it correctly, the explanation for why you should choose that answer was wrong. Mirror, mirror, on the wall…

Sample Question

A network consists of a master domain named Corp, and resource domains named Research and Sales. The Corp domain is trusted by both the Research and Sales domains. No other trust relationships exist. Joan has a user account in the Research domain and the Corp domain.

Which shared resources in the Corp domain will Joan be able to access when she logs on to the Research domain?

  1. All resources to which the Corp\Power Users group has access.
  2. All resources to which the Corp\Domain Users group has access.
  3. All resources to which the Corp\Domain Guests group has access.
  4. All resources for which the Corp\Domain Admins group has access.
  5. No resources in the Corp domain are accessible by users in the Research domain.

The problem with this question is that it tells us that the network is using a Master Domain model, yet Joan has a user account in more than one domain. This scenario doesn’t abide by the general rule of any domain model, which is that you should create a single account for a user and not multiple accounts. Specifically, the single Master Domain model is an agreement between multiple company business units that all user accounts will be centrally maintained in one domain: the master. Resource domains are created and configured to trust the Master Domain in a one-way trust relationship. The Master Domain is trusted, and the Resource domains are trusting. This is typically drawn with the Master Domain as a circle on the top of a diagram and the Resource domains as separate circles below the master. An arrow points from the Resource domains toward the Master Domain.

So if Joan were in a true Master Domain model, she should have a single account in the Master Domain database that is Corp. Of course the question doesn’t follow the established rules for a Master Domain but instead deviates from the rules, at the expense of all novice test takers out there who are already confused enough with the whole trust relationship thing. But I digress…

So, the question is poorly structured to begin with, but let’s deal with it. It appears that Joan has two separate accounts, one in Corp and one in Research. If she’s physically sitting at a computer in Research she can log on either to the Research domain or through the trust to the Corp domain. If she were logged on to the Corp domain, then she’d be a member of Corp/Domain Users with all the permissions and the privileges assigned to that group. If she were logged on to the Research domain, then she’d be a member of the Research/Domain Users group with all associated privileges.

Answers A through D are incorrect. If Joan were logged on with a Research account, she wouldn’t be seen as a member of any Corp group.

Answer E would be correct because Corp doesn’t trust Research users to have access to its resources (and rightfully so in a Single Master Model).

In short, I can summarize the MCSE Readiness Review book as follows:

Scenario: You’re searching for a study guide that will help you pass the NT Server 4 in the Enterprise exam.

Required Result: The study guide must provide a number of practice questions and explanations for both the correct and incorrect answers. This study guide shouldn’t be used alone but as a supplement for another self-study guide. Also, the study guide must clearly define the testable objectives on the exam.

First Optional Result: The practice questions and the answer explanations must be 100 percent technically correct.

Second Optional Result: The format of the book should be easy to follow and allow you to answer questions at your own pace.

My answer: Meets the required and only one of the optional results.

Another Sample Question
You’re configuring a Windows NT member server to act as a router between two IP networks. What must you enable on the Windows NT server to allow it to route TCP/IP?
  1. RIP
  2. SAP
  3. DLC
  4. WINS
  5. BOOTP

Battle of the acronyms. Actually, don’t be surprised if on the real exam, these acronyms are actually spelled out, i.e. Routing Information Protocol, vs. RIP, Data Link Control vs. DLC, and so on.

Seems like a simple enough question, and it is, although it’s problematic nonetheless because none of the answers seems to fit. Let me explain.

Answer A—RIP is a dynamic routing protocol that dynamically updates routing tables. RIP routers broadcast the entire content of their routing tables to alert other adjacent routers about available or non-available paths to other networks. So this protocol has something to do with routing, but any MCP can tell you that what makes an NT box a router is by taking a multihomed computer and enabling IP forwarding under the TCPIP protocol properties. IP Forwarding makes the NT computer a static router, but a router nonetheless. RIP is optional.

Answer B—SAP is the Service Advertising Protocol and is used by a Novell Server to let Novell clients and other Novell Servers know that it’s present on the network. Think of it as Novell’s browsing service. This has nothing to do with routing the IP protocol.

Answer C—DLC is the Datalink Control Protocol and is used primarily as the transport protocol to get access to HP printers (JetDirect) and IBM mainframes. Again, this protocol has nothing to do with IP routing.

Answer D—WINS is the Windows Internet Naming Service and is used on NT networks to resolve user-friendly NETBIOS names to IP addresses. Sorry, still no routing functionality here.

Answer E—BOOTP is the Bootstrap Protocol and was designed originally to get diskless workstations up and running with an operating system. No association with IP routing.

Darn! This means the only viable answer is A, but for all the wrong reasons. Oh, well good thing there’s no penalty for guessing on the exams. The rule of thumb is always choose some answer on the exam, even though it may not make much sense—choose the best answer. While teaching MCSE classes I came up with a phrase that eventually ended up on a T-shirt that I wore to class: “Do you want to be right or do you want to be an MCSE?”


MCSE Core Four Practice Tests Exam Cram
By Ed Tittel, James Michael Stewart, and Gary Novosel
Certification Insider Press 
(The Coriolis Group); www.coriolis.com 
ISBN 1-57610-475-3, $23.99, 1999
61 pages of the 300-page book cover the Enterprise exam

I have to admit that the first thing I did before starting this particular review was to visit the Coriolis Web site to see if there were any errata sheets on this book. My experience with the series in the past leads me to believe that if you buy the book, you must download the extensive errata sheets. Since I didn’t find any for this particular book, I figured the editors had gone through it with a fine-tooth comb.

The book is actually a compilation of test questions for the Enterprise exam as well as the NT Workstation, NT Server, and Networking Essentials exams. The 61 pages covering the Enterprise portion consists of two 55-question practice exams and the associated answer analysis. A companion CD includes a browser-based adaptive exam for the Enterprise exam. You can take a 15-question test that simulates the adaptive testing process.

I diligently went through both paper-based practice exams and was impressed with the overall clarity of the questions. They were for the most part to the point, and the answers could be readily ascertained. However, when I checked my answers against the answer keys, I was shocked to see how many I’d missed (according to the book). I immediately went to the answer analysis section to see why I’d missed so many. I was appalled at the number of “wrong” answers and bad explanations that followed.

Sample Question

our company has two domains: Corp and Sales. Corp is the master domain and Sales is a resource domain. Corp and Sales each have a PDC, a BDC, member servers, and several Windows NT Workstations. You want to create a group called CorpBackup to be able to back up the PDCs, BDCs, member servers, and Windows NT Workstations in both domains.

Required Result: Members of the CorpBackup group must be able to back up both domain controllers.
Optional Desired Results: Members of the Corpbackup group must be able to back up all member servers. Member of the CorpBackup group must be able to back up all workstations.
Proposed Solution: Create a Global group called CorpBackup in the Corp domain and add this new global group to the global Domain Admins group on the Primary Domain Controller for each domain.

Which result does the proposed solution satisfy?

  1. The proposed solution produces the required result and both the optional desired results.
  2. The proposed solution produces the required result and only one of the optional desired results.
  3. The proposed solution produces the required result but does not produce either of the optional desired results.
  4. The proposed solution does not produce the required result.

My answer was D because you can’t take a Global group and make it a member of another Global group. The book’s answer was A—it produces the required and both of the optional results. I thought maybe there was a typo and that they meant to say “add the Corpbackup global group to the local Administrators group” (which would work); but when I checked the analysis, this is what it said: “By adding the Corpbackup global group of the trusted domain to the Domain Admins group of each domain, you effectively give that group the right to back up all computers in both domains. Remember, by default the Domain Admin’s global group is added to the Administrators local group and all Windows NT computers in the domain.” Their solution just won’t work.

Another Sample Question
You have a Windows NT Server with IIS installed, and you want to host three Web sites on this server. Each site will have its own registered domain name and IP address. Which of the following actions must you perform to accommodate this situation?
  1. Set up a DHCP Relay Agent.
  2. Install DNS and add an entry for each site in the DNS database.
  3. Assign each IP address to the network interface card of the server and associate each IP address with the appropriate Web directory.
  4. Create a hosts file with an entry for each domain name with the corresponding IP addresses.

You’re being asked to configure what are called IIS Virtual Servers or multiple Web sites. IIS can be configured to host multiple company or departmental Web sites by making them appear to be running on separate computers, while in reality they’re all running on the same computer. This is accomplished by binding multiple IP addresses to the same network card and through the Internet Service Manager tool used to configure IIS, associating each IP address with a Web site home directory. This is answer C and the first step in the process.

After you’ve configured the multiple virtual servers, these are accessible through an IP address. However, how many folks do you know who want to remember an IP address to connect to one of the Web sites? They’d rather use a user-friendly name such as www.xyz.com. This can be done as long as some service can resolve the name down to the IP address for the client. Both answers C and D provide name resolution services and are viable answers. If you’re hosting these virtual servers on your company intranet, either method would work. If you’re hosting the virtual servers for the benefit of Internet users to access, then you absolutely need DNS.

If you’re hosting the XYZ Web site, and you’re advertising your Web site as www.xyz.com, that’s what Internet users will type into their browser for access. The user-friendly name must be resolved to the proper IP address for the XYZ Web site. If DNS is properly configured, then the DNS server can perform the name lookup for the user.

Setting up a DNS server to support name lookup on the Internet requires that you register your domain name with the Internic, i.e. XYZ.COM. You must then set up a DNS server (or use your ISP’s) to maintain the entry that maps www.xyz.com to its IP address (to be authoritative for the XYZ.COM domain). The Internic won’t put www.xyz.com in their DNS server database as the book contends, but will refer other DNS servers your way so that the resolution can be made through your DNS server.

For the small number of questions in the Enterprise section (a total of 110), I found too many inexcusable mistakes and too much bad technical information being given out for this book to be a practical preparation guide—at least not until the errata sheet is released. As currently published, this book can be detrimental for the novice NT user who’ll accept the incorrect explanations at face value. It’s a shame; the questions themselves are nicely worded. My advice is to put this one on the shelf until the content is re-edited and the content corrected.